Search results for: Fourier neural operator

Authors: Mahshid Arabi

Abstract:

With the rapid expansion of digital technologies and the internet, information security has become a critical priority for organizations and individuals. The widespread use of digital tools such as smartphones and internet networks facilitates the storage of vast amounts of data, but simultaneously, vulnerabilities and security threats have significantly increased. The aim of this study is to examine and analyze modern methods of information security management and to develop a comprehensive model to counteract threats and information misuse. This study employs a mixed-methods approach, including both qualitative and quantitative analyses. Initially, a systematic review of previous articles and research in the field of information security was conducted. Then, using the Delphi method, interviews with 30 information security experts were conducted to gather their insights on security challenges and solutions. Based on the results of these interviews, a comprehensive model for information security management was developed. The proposed model includes advanced encryption techniques, machine learning-based intrusion detection systems, and network security protocols. AES and RSA encryption algorithms were used for data protection, and machine learning models such as Random Forest and Neural Networks were utilized for intrusion detection. Statistical analyses were performed using SPSS software. To evaluate the effectiveness of the proposed model, T-Test and ANOVA statistical tests were employed, and results were measured using accuracy, sensitivity, and specificity indicators of the models. Additionally, multiple regression analysis was conducted to examine the impact of various variables on information security. The findings of this study indicate that the comprehensive proposed model reduced cyber-attacks by an average of 85%. Statistical analysis showed that the combined use of encryption techniques and intrusion detection systems significantly improves information security. Based on the obtained results, it is recommended that organizations continuously update their information security systems and use a combination of multiple security methods to protect their data. Additionally, educating employees and raising public awareness about information security can serve as an effective tool in reducing security risks. This research demonstrates that effective and up-to-date information security management requires a comprehensive and coordinated approach, including the development and implementation of advanced techniques and continuous training of human resources.

Keywords: data protection, digital technologies, information security, modern management

Procedia PDF Downloads 29

Authors: Shengnan Chen, Shuhua Wang

Abstract:

Successful production of hydrocarbon from unconventional tight oil reserves has changed the energy landscape in North America. The oil contained within these reservoirs typically will not flow to the wellbore at economic rates without assistance from advanced horizontal well and multi-stage hydraulic fracturing. Efficient and economic development of these reserves is a priority of society, government, and industry, especially under the current low oil prices. Meanwhile, society needs technological and process innovations to enhance oil recovery while concurrently reducing environmental impacts. Recently, big data analysis and artificial intelligence become very popular, developing data-driven insights for better designs and decisions in various engineering disciplines. However, the application of data mining in petroleum engineering is still in its infancy. The objective of this research aims to apply intelligent data analysis and data-driven models to exploit unconventional oil reserves both efficiently and economically. More specifically, a comprehensive database including the reservoir geological data, reservoir geophysical data, well completion data and production data for thousands of wells is firstly established to discover the valuable insights and knowledge related to tight oil reserves development. Several data analysis methods are introduced to analysis such a huge dataset. For example, K-means clustering is used to partition all observations into clusters; principle component analysis is applied to emphasize the variation and bring out strong patterns in the dataset, making the big data easy to explore and visualize; exploratory factor analysis (EFA) is used to identify the complex interrelationships between well completion data and well production data. Different data mining techniques, such as artificial neural network, fuzzy logic, and machine learning technique are then summarized, and appropriate ones are selected to analyze the database based on the prediction accuracy, model robustness, and reproducibility. Advanced knowledge and patterned are finally recognized and integrated into a modified self-adaptive differential evolution optimization workflow to enhance the oil recovery and maximize the net present value (NPV) of the unconventional oil resources. This research will advance the knowledge in the development of unconventional oil reserves and bridge the gap between the big data and performance optimizations in these formations. The newly developed data-driven optimization workflow is a powerful approach to guide field operation, which leads to better designs, higher oil recovery and economic return of future wells in the unconventional oil reserves.

Keywords: big data, artificial intelligence, enhance oil recovery, unconventional oil reserves

Procedia PDF Downloads 283

Authors: Asma Omri, Iheb Sifaoui, Sofiane Sayahi, Hichem Besbes

Abstract:

Future vehicle systems demand advanced capabilities, notably in-cabin life detection and driver monitoring systems, with a particular emphasis on drowsiness detection. To meet these requirements, several techniques employ artificial intelligence methods based on real-time vital sign measurements. In parallel, Frequency-Modulated Continuous-Wave (FMCW) radar technology has garnered considerable attention in the domains of healthcare and biomedical engineering for non-invasive vital sign monitoring. FMCW radar offers a multitude of advantages, including its non-intrusive nature, continuous monitoring capacity, and its ability to penetrate through clothing. In this paper, we propose a system utilizing the AWR6843AOP radar from Texas Instruments (TI) to extract precise vital sign information. The radar allows us to estimate Ballistocardiogram (BCG) signals, which capture the mechanical movements of the body, particularly the ballistic forces generated by heartbeats and respiration. These signals are rich sources of information about the cardiac cycle, rendering them suitable for heart rate estimation. The process begins with real-time subject positioning, followed by clutter removal, computation of Doppler phase differences, and the use of various filtering methods to accurately capture subtle physiological movements. To address the challenges associated with FMCW radar-based vital sign monitoring, including motion artifacts due to subjects' movement or radar micro-vibrations, Long Short-Term Memory (LSTM) networks are implemented. LSTM's adaptability to different heart rate patterns and ability to handle real-time data make it suitable for continuous monitoring applications. Several crucial steps were taken, including feature extraction (involving amplitude, time intervals, and signal morphology), sequence modeling, heart rate estimation through the analysis of detected cardiac cycles and their temporal relationships, and performance evaluation using metrics such as Root Mean Square Error (RMSE) and correlation with reference heart rate measurements. For dataset construction and LSTM training, a comprehensive data collection system was established, integrating the AWR6843AOP radar, a Heart Rate Belt, and a smart watch for ground truth measurements. Rigorous synchronization of these devices ensured data accuracy. Twenty participants engaged in various scenarios, encompassing indoor and real-world conditions within a moving vehicle equipped with the radar system. Static and dynamic subject’s conditions were considered. The heart rate estimation through LSTM outperforms traditional signal processing techniques that rely on filtering, Fast Fourier Transform (FFT), and thresholding. It delivers an average accuracy of approximately 91% with an RMSE of 1.01 beat per minute (bpm). In conclusion, this paper underscores the promising potential of FMCW radar technology integrated with artificial intelligence algorithms in the context of automotive applications. This innovation not only enhances road safety but also paves the way for its integration into the automotive ecosystem to improve driver well-being and overall vehicular safety.

Keywords: ballistocardiogram, FMCW Radar, vital sign monitoring, LSTM

Procedia PDF Downloads 72

Authors: Jeremy Jenrette, Z. Y. C. Liu, Pranav Chimote, Edward Fox, Trevor Hastie, Francesco Ferretti

Abstract:

Suitable shark conservation depends on well-informed population assessments. Direct methods such as scientific surveys and fisheries monitoring are adequate for defining population statuses, but species-specific indices of abundance and distribution coming from these sources are rare for most shark species. We can rapidly fill these information gaps by boosting media-based remote monitoring efforts with machine learning and automation. We created a database of shark images by sourcing 24,546 images covering 219 species of sharks from the web application spark pulse and the social network Instagram. We used object detection to extract shark features and inflate this database to 53,345 images. We packaged object-detection and image classification models into a Shark Detector bundle. We developed the Shark Detector to recognize and classify sharks from videos and images using transfer learning and convolutional neural networks (CNNs). We applied these models to common data-generation approaches of sharks: boosting training datasets, processing baited remote camera footage and online videos, and data-mining Instagram. We examined the accuracy of each model and tested genus and species prediction correctness as a result of training data quantity. The Shark Detector located sharks in baited remote footage and YouTube videos with an average accuracy of 89\%, and classified located subjects to the species level with 69\% accuracy (n =\ eight species). The Shark Detector sorted heterogeneous datasets of images sourced from Instagram with 91\% accuracy and classified species with 70\% accuracy (n =\ 17 species). Data-mining Instagram can inflate training datasets and increase the Shark Detector’s accuracy as well as facilitate archiving of historical and novel shark observations. Base accuracy of genus prediction was 68\% across 25 genera. The average base accuracy of species prediction within each genus class was 85\%. The Shark Detector can classify 45 species. All data-generation methods were processed without manual interaction. As media-based remote monitoring strives to dominate methods for observing sharks in nature, we developed an open-source Shark Detector to facilitate common identification applications. Prediction accuracy of the software pipeline increases as more images are added to the training dataset. We provide public access to the software on our GitHub page.

Keywords: classification, data mining, Instagram, remote monitoring, sharks

Procedia PDF Downloads 121

Authors: Xiangtuo Chen, Paul-Henry Cournéde

Abstract:

Crop yield prediction is a paramount issue in agriculture. The main idea of this paper is to find out efficient way to predict the yield of corn based meteorological records. The prediction models used in this paper can be classified into model-driven approaches and data-driven approaches, according to the different modeling methodologies. The model-driven approaches are based on crop mechanistic modeling. They describe crop growth in interaction with their environment as dynamical systems. But the calibration process of the dynamic system comes up with much difficulty, because it turns out to be a multidimensional non-convex optimization problem. An original contribution of this paper is to propose a statistical methodology, Multi-Scenarios Parameters Estimation (MSPE), for the parametrization of potentially complex mechanistic models from a new type of datasets (climatic data, final yield in many situations). It is tested with CORNFLO, a crop model for maize growth. On the other hand, the data-driven approach for yield prediction is free of the complex biophysical process. But it has some strict requirements about the dataset. A second contribution of the paper is the comparison of these model-driven methods with classical data-driven methods. For this purpose, we consider two classes of regression methods, methods derived from linear regression (Ridge and Lasso Regression, Principal Components Regression or Partial Least Squares Regression) and machine learning methods (Random Forest, k-Nearest Neighbor, Artificial Neural Network and SVM regression). The dataset consists of 720 records of corn yield at county scale provided by the United States Department of Agriculture (USDA) and the associated climatic data. A 5-folds cross-validation process and two accuracy metrics: root mean square error of prediction(RMSEP), mean absolute error of prediction(MAEP) were used to evaluate the crop prediction capacity. The results show that among the data-driven approaches, Random Forest is the most robust and generally achieves the best prediction error (MAEP 4.27%). It also outperforms our model-driven approach (MAEP 6.11%). However, the method to calibrate the mechanistic model from dataset easy to access offers several side-perspectives. The mechanistic model can potentially help to underline the stresses suffered by the crop or to identify the biological parameters of interest for breeding purposes. For this reason, an interesting perspective is to combine these two types of approaches.

Keywords: crop yield prediction, crop model, sensitivity analysis, paramater estimation, particle swarm optimization, random forest

Procedia PDF Downloads 231

Authors: Pascale Brasseur, Binu Gurung, Nicholas Halfpenny, James Eaton

Abstract:

Development of minimally invasive treatments in recent years provides a potential alternative to invasive surgical interventions which are of limited value to patients with spinal metastases due to short life expectancy. A systematic review was conducted to explore the efficacy and safety of radiofrequency ablation (RFA), a minimally invasive treatment in patients with spinal metastases. EMBASE, Medline and CENTRAL were searched from database inception to March 2017 for randomised controlled trials (RCTs) and non-randomised studies. Conference proceedings for ASCO and ESMO published in 2015 and 2016 were also searched. Fourteen studies were included: three prospective interventional studies, four prospective case series and seven retrospective case series. No RCTs or studies comparing RFA with another treatment were identified. RFA was followed by cement augmentation in all patients in seven studies and some patients (40-96%) in the remaining seven studies. Efficacy was assessed as pain relief in 13/14 studies with the use of a numerical rating scale (NRS) or a visual analogue scale (VAS) at various time points. Ten of the 13 studies reported a significant decrease in pain outcome, post-RFA compared to baseline. NRS scores improved significantly at 1 week (5.9 to 3.5, p < 0.0001; 8 to 4.3, p < 0.02 and 8 to 3.9, p < 0.0001) and this improvement was maintained at 1 month post-RFA compared to baseline (5.9 to 2.6, p < 0.0001; 8 to 2.9, p < 0.0003; 8 to 2.9, p < 0.0001). Similarly, VAS scores decreased significantly at 1 week (7.5 to 2.7, p=0.00005; 7.51 to 1.73, p < 0.0001; 7.82 to 2.82, p < 0.001) and this pattern was maintained at 1 month post-RFA compared to baseline (7.51 to 2.25, p < 0.0001; 7.82 to 3.3; p < 0.001). A significant pain relief was achieved regardless of whether patients had cement augmentation in two studies assessing the impact of RFA with or without cement augmentation on VAS pain scores. In these two studies, a significant decrease in pain scores was reported for patients receiving RFA alone and RFA+cement at 1 week (4.3 to 1.7. p=0.0004 and 6.6 to 1.7, p=0.003 respectively) and 15-36 months (7.9 to 4, p=0.008 and 7.6 to 3.5, p=0.005 respectively) after therapy. Few minor complications were reported and these included neural damage, radicular pain, vertebroplasty leakage and lower limb pain/numbness. In conclusion, the efficacy and safety of RFA were consistently positive between prospective and retrospective studies with reductions in pain and few procedural complications. However, the lack of control groups in the identified studies indicates the possibility of selection bias inherent in single arm studies. Controlled trials exploring efficacy and safety of RFA in patients with spinal metastases are warranted to provide robust evidence. The identified studies provide an initial foundation for such future trials.

Keywords: pain relief, radiofrequency ablation, spinal metastases, systematic review

Procedia PDF Downloads 173

Authors: Stefan Papastefanou

Abstract:

Artificial intelligence (AI) is an interdisciplinary field of computer science with the aim of creating intelligent machine behavior. Early approaches to AI have been configured to operate in very constrained environments where the behavior of the AI system was previously determined by formal rules. Knowledge was presented as a set of rules that allowed the AI system to determine the results for specific problems; as a structure of if-else rules that could be traversed to find a solution to a particular problem or question. However, such rule-based systems typically have not been able to generalize beyond the knowledge provided. All over the world and especially in IT-heavy industries such as the United States, the European Union, Singapore, and China, machine learning has developed to be an immense asset, and its applications are becoming more and more significant. It has to be examined how such products of machine learning models can and should be protected by IP law and for the purpose of this paper patent law specifically, since it is the IP law regime closest to technical inventions and computing methods in technical applications. Genetic breeding models are currently less popular than recursive neural network method and deep learning, but this approach can be more easily described by referring to the evolution of natural organisms, and with increasing computational power; the genetic breeding method as a subset of the evolutionary algorithms models is expected to be regaining popularity. The research method focuses on patentability (according to the world’s most significant patent law regimes such as China, Singapore, the European Union, and the United States) of AI inventions and machine learning. Questions of the technical nature of the problem to be solved, the inventive step as such, and the question of the state of the art and the associated obviousness of the solution arise in the current patenting processes. Most importantly, and the key focus of this paper is the problem of patenting inventions that themselves are developed through machine learning. The inventor of a patent application must be a natural person or a group of persons according to the current legal situation in most patent law regimes. In order to be considered an 'inventor', a person must actually have developed part of the inventive concept. The mere application of machine learning or an AI algorithm to a particular problem should not be construed as the algorithm that contributes to a part of the inventive concept. However, when machine learning or the AI algorithm has contributed to a part of the inventive concept, there is currently a lack of clarity regarding the ownership of artificially created inventions. Since not only all European patent law regimes but also the Chinese and Singaporean patent law approaches include identical terms, this paper ultimately offers a comparative analysis of the most relevant patent law regimes.

Keywords: algorithms, inventor, genetic breeding models, machine learning, patentability

Procedia PDF Downloads 108

Authors: E. Nakamachi, K. Matsumoto, K. Yamamoto, Y. Morita, H. Sakamoto

Abstract:

In recently, electromagnetic and mechanical stimulations have been recognized as the effective extracellular environment stimulation technique to enhance the defected peripheral nerve tissue regeneration. In this study, we developed a new hybrid bioreactor by adopting 50 Hz uniform alternative current (AC) magnetic stimulation and 4% strain mechanical stimulation. The guide tube for nerve regeneration is mesh structured tube made of biodegradable polymer, such as polylatic acid (PLA). However, when neural damage is large, there is a possibility that peripheral nerve undergoes necrosis. So it is quite important to accelerate the nerve tissue regeneration by achieving enhancement of nerve axonal extension rate. Therefore, we try to design and fabricate the system that can simultaneously load the uniform AC magnetic field stimulation and the stretch stimulation to cells for enhancement of nerve axonal extension. Next, we evaluated systems performance and the effectiveness of each stimulation for rat adrenal pheochromocytoma cells (PC12). First, we designed and fabricated the uniform AC magnetic field system and the stretch stimulation system. For the AC magnetic stimulation system, we focused on the use of pole piece structure to carry out in-situ microscopic observation. We designed an optimum pole piece structure using the magnetic field finite element analyses and the response surface methodology. We fabricated the uniform AC magnetic field stimulation system as a bio-reactor by adopting analytically determined design specifications. We measured magnetic flux density that is generated by the uniform AC magnetic field stimulation system. We confirmed that measurement values show good agreement with analytical results, where the uniform magnetic field was observed. Second, we fabricated the cyclic stretch stimulation device under the conditions of particular strains, where the chamber was made of polyoxymethylene (POM). We measured strains in the PC12 cell culture region to confirm the uniform strain. We found slightly different values from the target strain. Finally, we concluded that these differences were allowable in this mechanical stimulation system. We evaluated the effectiveness of each stimulation to enhance the nerve axonal extension using PC12. We confirmed that the average axonal extension length of PC12 under the uniform AC magnetic stimulation was increased by 16 % at 96 h in our bio-reactor. We could not confirm that the axonal extension enhancement under the stretch stimulation condition, where we found the exfoliating of cells. Further, the hybrid stimulation enhanced the axonal extension. Because the magnetic stimulation inhibits the exfoliating of cells. Finally, we concluded that the enhancement of PC12 axonal extension is due to the magnetic stimulation rather than the mechanical stimulation. Finally, we confirmed that the effectiveness of the uniform AC magnetic field stimulation for the nerve axonal extension using PC12 cells.

Keywords: nerve cell PC12, axonal extension, nerve regeneration, electromagnetic-mechanical stimulation, bioreactor

Procedia PDF Downloads 264

Authors: A. M. Byrne, E. Bonfiglio, C. Rigby, N. Edelstyn

Abstract:

Introduction: Neuromarketing employs numerous methodologies when investigating products and advertisement effectiveness. Electroencephalography (EEG), a non-invasive measure of electrical activity from the brain, is commonly used in neuromarketing. EEG data can be considered using time-frequency (TF) analysis, where changes in the frequency of brainwaves are calculated to infer participant’s mental states, or event-related potential (ERP) analysis, where changes in amplitude are observed in direct response to a stimulus. This presentation discusses the findings of a systematic review of EEG measures in neuromarketing. A systematic review summarises evidence on a research question, using explicit measures to identify, select, and critically appraise relevant research papers. Thissystematic review identifies which EEG measures are the most robust predictor of customer preference and purchase intention. Methods: Search terms identified174 papers that used EEG in combination with marketing-related stimuli. Publications were excluded if they were written in a language other than English or were not published as journal articles (e.g., book chapters). The review investigated which TF effect (e.g., theta-band power) and ERP component (e.g., N400) most consistently reflected preference and purchase intention. Machine-learning prediction was also investigated, along with the use of EEG combined with physiological measures such as eye-tracking. Results: Frontal alpha asymmetry was the most reliable TF signal, where an increase in activity over the left side of the frontal lobe indexed a positive response to marketing stimuli, while an increase in activity over the right side indexed a negative response. The late positive potential, a positive amplitude increase around 600 ms after stimulus presentation, was the most reliable ERP component, reflecting the conscious emotional evaluation of marketing stimuli. However, each measure showed mixed results when related to preference and purchase behaviour. Predictive accuracy was greatly improved through machine-learning algorithms such as deep neural networks, especially when combined with eye-tracking or facial expression analyses. Discussion: This systematic review provides a novel catalogue of the most effective use of each EEG measure commonly used in neuromarketing. Exciting findings to emerge are the identification of the frontal alpha asymmetry and late positive potential as markers of preferential responses to marketing stimuli. Predictive accuracy using machine-learning algorithms achieved predictive accuracies as high as 97%, and future research should therefore focus on machine-learning prediction when using EEG measures in neuromarketing.

Keywords: EEG, ERP, neuromarketing, machine-learning, systematic review, time-frequency

Procedia PDF Downloads 111

Authors: Yuta Nakano, Kozo Kajiwara, Atsushi Hori, Takeshi Fujita

Abstract:

In association with the wave of electric vehicles (EV), low energy consumption systems have become more and more important. One of the key technologies to realize low energy consumption is a dynamic vision sensor (DVS), or we can call it an event sensor, neuromorphic vision sensor and so on. This sensor has several features, such as high temporal resolution, which can achieve 1 Mframe/s, and a high dynamic range (120 DB). However, the point that can contribute to low energy consumption the most is its sparsity; to be more specific, this sensor only captures the pixels that have intensity change. In other words, there is no signal in the area that does not have any intensity change. That is to say, this sensor is more energy efficient than conventional sensors such as RGB cameras because we can remove redundant data. On the other side of the advantages, it is difficult to handle the data because the data format is completely different from RGB image; for example, acquired signals are asynchronous and sparse, and each signal is composed of x-y coordinate, polarity (two values: +1 or -1) and time stamp, it does not include intensity such as RGB values. Therefore, as we cannot use existing algorithms straightforwardly, we have to design a new processing algorithm to cope with DVS data. In order to solve difficulties caused by data format differences, most of the prior arts make a frame data and feed it to deep learning such as Convolutional Neural Networks (CNN) for object detection and recognition purposes. However, even though we can feed the data, it is still difficult to achieve good performance due to a lack of intensity information. Although polarity is often used as intensity instead of RGB pixel value, it is apparent that polarity information is not rich enough. Considering this context, we proposed to use the timestamp information as a data representation that is fed to deep learning. Concretely, at first, we also make frame data divided by a certain time period, then give intensity value in response to the timestamp in each frame; for example, a high value is given on a recent signal. We expected that this data representation could capture the features, especially of moving objects, because timestamp represents the movement direction and speed. By using this proposal method, we made our own dataset by DVS fixed on a parked car to develop an application for a surveillance system that can detect persons around the car. We think DVS is one of the ideal sensors for surveillance purposes because this sensor can run for a long time with low energy consumption in a NOT dynamic situation. For comparison purposes, we reproduced state of the art method as a benchmark, which makes frames the same as us and feeds polarity information to CNN. Then, we measured the object detection performances of the benchmark and ours on the same dataset. As a result, our method achieved a maximum of 7 points greater than the benchmark in the F1 score.

Keywords: event camera, dynamic vision sensor, deep learning, data representation, object recognition, low energy consumption

Procedia PDF Downloads 97

Authors: L. Diogo, A. Legrand, J. Nguyen-Cao, J. Rogeau, S. Bornhofen

Abstract:

Lenia is a system of cellular automata with continuous states, space and time, which surprises not only with the emergence of interesting life-like structures but also with its beauty. This paper reports ongoing research on a GPU implementation of Lenia using CUDA C++ and OpenGL Interoperability. We demonstrate how CUDA as a low-level GPU programming paradigm allows optimizing performance and memory usage of the Lenia algorithm. A comparative analysis through experimental runs with existing implementations shows that the CUDA implementation outperforms the others by one order of magnitude or more. Cellular automata hold significant interest due to their ability to model complex phenomena in systems with simple rules and structures. They allow exploring emergent behavior such as self-organization and adaptation, and find applications in various fields, including computer science, physics, biology, and sociology. Unlike classic cellular automata which rely on discrete cells and values, Lenia generalizes the concept of cellular automata to continuous space, time and states, thus providing additional fluidity and richness in emerging phenomena. In the current literature, there are many implementations of Lenia utilizing various programming languages and visualization libraries. However, each implementation also presents certain drawbacks, which serve as motivation for further research and development. In particular, speed is a critical factor when studying Lenia, for several reasons. Rapid simulation allows researchers to observe the emergence of patterns and behaviors in more configurations, on bigger grids and over longer periods without annoying waiting times. Thereby, they enable the exploration and discovery of new species within the Lenia ecosystem more efficiently. Moreover, faster simulations are beneficial when we include additional time-consuming algorithms such as computer vision or machine learning to evolve and optimize specific Lenia configurations. We developed a Lenia implementation for GPU using the C++ and CUDA programming languages, and CUDA/OpenGL Interoperability for immediate rendering. The goal of our experiment is to benchmark this implementation compared to the existing ones in terms of speed, memory usage, configurability and scalability. In our comparison we focus on the most important Lenia implementations, selected for their prominence, accessibility and widespread use in the scientific community. The implementations include MATLAB, JavaScript, ShaderToy GLSL, Jupyter, Rust and R. The list is not exhaustive but provides a broad view of the principal current approaches and their respective strengths and weaknesses. Our comparison primarily considers computational performance and memory efficiency, as these factors are critical for large-scale simulations, but we also investigate the ease of use and configurability. The experimental runs conducted so far demonstrate that the CUDA C++ implementation outperforms the other implementations by one order of magnitude or more. The benefits of using the GPU become apparent especially with larger grids and convolution kernels. However, our research is still ongoing. We are currently exploring the impact of several software design choices and optimization techniques, such as convolution with Fast Fourier Transforms (FFT), various GPU memory management scenarios, and the trade-off between speed and accuracy using single versus double precision floating point arithmetic. The results will give valuable insights into the practice of parallel programming of the Lenia algorithm, and all conclusions will be thoroughly presented in the conference paper. The final version of our CUDA C++ implementation will be published on github and made freely accessible to the Alife community for further development.

Keywords: artificial life, cellular automaton, GPU optimization, Lenia, comparative analysis.

Procedia PDF Downloads 41

Authors: Shaukat Ali Shahee, Usha Ananthakumar

Abstract:

A data set exhibits class imbalance problem when one class has very few examples compared to the other class, and this is also referred to as between class imbalance. The traditional classifiers fail to classify the minority class examples correctly due to its bias towards the majority class. Apart from between-class imbalance, imbalance within classes where classes are composed of a different number of sub-clusters with these sub-clusters containing different number of examples also deteriorates the performance of the classifier. Previously, many methods have been proposed for handling imbalanced dataset problem. These methods can be classified into four categories: data preprocessing, algorithmic based, cost-based methods and ensemble of classifier. Data preprocessing techniques have shown great potential as they attempt to improve data distribution rather than the classifier. Data preprocessing technique handles class imbalance either by increasing the minority class examples or by decreasing the majority class examples. Decreasing the majority class examples lead to loss of information and also when minority class has an absolute rarity, removing the majority class examples is generally not recommended. Existing methods available for handling class imbalance do not address both between-class imbalance and within-class imbalance simultaneously. In this paper, we propose a method that handles between class imbalance and within class imbalance simultaneously for binary classification problem. Removing between class imbalance and within class imbalance simultaneously eliminates the biases of the classifier towards bigger sub-clusters by minimizing the error domination of bigger sub-clusters in total error. The proposed method uses model-based clustering to find the presence of sub-clusters or sub-concepts in the dataset. The number of examples oversampled among the sub-clusters is determined based on the complexity of sub-clusters. The method also takes into consideration the scatter of the data in the feature space and also adaptively copes up with unseen test data using Lowner-John ellipsoid for increasing the accuracy of the classifier. In this study, neural network is being used as this is one such classifier where the total error is minimized and removing the between-class imbalance and within class imbalance simultaneously help the classifier in giving equal weight to all the sub-clusters irrespective of the classes. The proposed method is validated on 9 publicly available data sets and compared with three existing oversampling techniques that rely on the spatial location of minority class examples in the euclidean feature space. The experimental results show the proposed method to be statistically significantly superior to other methods in terms of various accuracy measures. Thus the proposed method can serve as a good alternative to handle various problem domains like credit scoring, customer churn prediction, financial distress, etc., that typically involve imbalanced data sets.

Keywords: classification, imbalanced dataset, Lowner-John ellipsoid, model based clustering, oversampling

Procedia PDF Downloads 418

Authors: Martin Pellon Consunji

Abstract:

Due to the rise in popularity of Bitcoin and other crypto assets as a store of wealth and speculative investment, there is an ever-growing demand for automated trading tools, such as bots, in order to gain an advantage over the market. Traditionally, trading in the stock market was done by professionals with years of training who understood patterns and exploited market opportunities in order to gain a profit. However, nowadays a larger portion of market participants are at minimum aided by market-data processing bots, which can generally generate more stable signals than the average human trader. The rise in trading bot usage can be accredited to the inherent advantages that bots have over humans in terms of processing large amounts of data, lack of emotions of fear or greed, and predicting market prices using past data and artificial intelligence, hence a growing number of approaches have been brought forward to tackle this task. However, the general limitation of these approaches can still be broken down to the fact that limited historical data doesn’t always determine the future, and that a lot of market participants are still human emotion-driven traders. Moreover, developing markets such as those of the cryptocurrency space have even less historical data to interpret than most other well-established markets. Due to this, some human traders have gone back to the tried-and-tested traditional technical analysis tools for exploiting market patterns and simplifying the broader spectrum of data that is involved in making market predictions. This paper proposes a method which uses neuro evolution techniques on both sentimental data and, the more traditionally human-consumed, technical analysis data in order to gain a more accurate forecast of future market behavior and account for the way both automated bots and human traders affect the market prices of Bitcoin and other cryptocurrencies. This study’s approach uses evolutionary algorithms to automatically develop increasingly improved populations of bots which, by using the latest inflows of market analysis and sentimental data, evolve to efficiently predict future market price movements. The effectiveness of the approach is validated by testing the system in a simulated historical trading scenario, a real Bitcoin market live trading scenario, and testing its robustness in other cryptocurrency and stock market scenarios. Experimental results during a 30-day period show that this method outperformed the buy and hold strategy by over 260% in terms of net profits, even when taking into consideration standard trading fees.

Keywords: neuro-evolution, Bitcoin, trading bots, artificial neural networks, technical analysis, evolutionary algorithms

Procedia PDF Downloads 123

Authors: Rafael do Amaral Teodoro, Leandro Augusto da Silva

Abstract:

Originated from graphite, graphene is a two-dimensional (2D) material that promises to revolutionize technology in many different areas, such as energy, telecommunications, civil construction, aviation, textile, and medicine. This is possible because its structure, formed by carbon bonds, provides desirable optical, thermal, and mechanical characteristics that are interesting to multiple areas of the market. Thus, several research and development centers are studying different manufacturing methods and material applications of graphene, which are often compromised by the scarcity of more agile and accurate methodologies to characterize the material – that is to determine its composition, shape, size, and the number of layers and crystals. To engage in this search, this study proposes a computational methodology that applies deep learning to identify graphene oxide crystals in order to characterize samples by crystal sizes. To achieve this, a fully convolutional neural network called U-net has been trained to segment SEM graphene oxide images. The segmentation generated by the U-net is fine-tuned with a standard deviation technique by classes, which allows crystals to be distinguished with different labels through an object delimitation algorithm. As a next step, the characteristics of the position, area, perimeter, and lateral measures of each detected crystal are extracted from the images. This information generates a database with the dimensions of the crystals that compose the samples. Finally, graphs are automatically created showing the frequency distributions by area size and perimeter of the crystals. This methodological process resulted in a high capacity of segmentation of graphene oxide crystals, presenting accuracy and F-score equal to 95% and 94%, respectively, over the test set. Such performance demonstrates a high generalization capacity of the method in crystal segmentation, since its performance considers significant changes in image extraction quality. The measurement of non-overlapping crystals presented an average error of 6% for the different measurement metrics, thus suggesting that the model provides a high-performance measurement for non-overlapping segmentations. For overlapping crystals, however, a limitation of the model was identified. To overcome this limitation, it is important to ensure that the samples to be analyzed are properly prepared. This will minimize crystal overlap in the SEM image acquisition and guarantee a lower error in the measurements without greater efforts for data handling. All in all, the method developed is a time optimizer with a high measurement value, considering that it is capable of measuring hundreds of graphene oxide crystals in seconds, saving weeks of manual work.

Keywords: characterization, graphene oxide, nanomaterials, U-net, deep learning

Procedia PDF Downloads 160

Authors: A. Degale Desta, Tamirat Kebamo

Abstract:

Background: Recent advances in computer vision and pattern recognition have significantly improved activity recognition through video analysis, particularly with the application of Deep Convolutional Neural Networks (CNNs). One-stage detectors now enable efficient video-based recognition by simultaneously predicting object categories and locations. Such advancements are highly relevant in educational settings where CCTV surveillance could automatically monitor academic activities, enhancing security and classroom management. However, current datasets and recognition systems lack the specific focus on campus environments necessary for practical application in these settings.Objective: This study aims to address this gap by developing a dataset and testing an automated activity recognition system specifically tailored for educational campuses. The EthioCAD dataset was created to capture various classroom activities and teacher-student interactions, facilitating reliable recognition of academic activities using deep learning models. Method: EthioCAD, a novel video-based dataset, was created with a design science research approach to encompass teacher-student interactions across three domains and 18 distinct classroom activities. Using the Roboflow AI framework, the data was processed, with 4.224 KB of frames and 33.485 MB of images managed for frame extraction, labeling, and organization. The Ultralytics YOLOv8 model was then implemented within Google Colab to evaluate the dataset’s effectiveness, achieving high mean Average Precision (mAP) scores. Results: The YOLOv8 model demonstrated robust activity recognition within campus-like settings, achieving an mAP50 of 90.2% and an mAP50-95 of 78.6%. These results highlight the potential of EthioCAD, combined with YOLOv8, to provide reliable detection and classification of classroom activities, supporting automated surveillance needs on educational campuses. Discussion: The high performance of YOLOv8 on the EthioCAD dataset suggests that automated activity recognition for surveillance is feasible within educational environments. This system addresses current limitations in campus-specific data and tools, offering a tailored solution for academic monitoring that could enhance the effectiveness of CCTV systems in these settings. Conclusion: The EthioCAD dataset, alongside the YOLOv8 model, provides a promising framework for automated campus activity recognition. This approach lays the groundwork for future advancements in CCTV-based educational surveillance systems, enabling more refined and reliable monitoring of classroom activities.

Keywords: deep CNN, EthioCAD, deep learning, YOLOv8, activity recognition

Procedia PDF Downloads 10

Authors: Sevde Altuntas, Fatih Buyukserin

Abstract:

Alzheimer’s disease is responsible for irreversible neural damage of brain parts. One of the disease markers is Amyloid-β 1-42 protein that accumulates in the brain in the form plaques. The basic problem for detection of the protein is the low amount of protein that cannot be detected properly in body liquids such as blood, saliva or urine. To solve this problem, tests like ELISA or PCR are proposed which are expensive, require specialized personnel and can contain complex protocols. Therefore, Surface-enhanced Raman Spectroscopy (SERS) a good candidate for detection of Amyloid-β 1-42 protein. Because the spectroscopic technique can potentially allow even single molecule detection from liquid and solid surfaces. Besides SERS signal can be improved by using nanopattern surface and also is specific to molecules. In this context, our study proposes to fabricate diagnostic test models that utilize Au-coated nanopatterned polycarbonate (PC) surfaces modified with Thioflavin - T to detect low concentrations of Amyloid-β 1-42 protein in water and artificial saliva medium by the enhancement of protein SERS signal. The nanopatterned PC surface that was used to enhance SERS signal was fabricated by using Anodic Alumina Membranes (AAM) as a template. It is possible to produce AAMs with different column structures and varying thicknesses depending on voltage and anodization time. After fabrication process, the pore diameter of AAMs can be arranged with dilute acid solution treatment. In this study, two different columns structures were prepared. After a surface modification to decrease their surface energy, AAMs were treated with PC solution. Following the solvent evaporation, nanopatterned PC films with tunable pillared structures were peeled off from the membrane surface. The PC film was then modified with Au and Thioflavin-T for the detection of Amyloid-β 1-42 protein. The protein detection studies were conducted first in water via this biosensor platform. Same measurements were conducted in artificial saliva to detect the presence of Amyloid Amyloid-β 1-42 protein. SEM, SERS and contact angle measurements were carried out for the characterization of different surfaces and further demonstration of the protein attachment. SERS enhancement factor calculations were also completed via experimental results. As a result, our research group fabricated diagnostic test models that utilize Au-coated nanopatterned polycarbonate (PC) surfaces modified with Thioflavin-T to detect low concentrations of Alzheimer’s Amiloid – β protein in water and artificial saliva medium. This work was supported by The Scientific and Technological Research Council of Turkey (TUBITAK) Grant No: 214Z167.

Keywords: alzheimer, anodic aluminum oxide, nanotopography, surface enhanced Raman spectroscopy

Procedia PDF Downloads 291

Authors: Shayla He

Abstract:

Background and Purpose: According to Chamie (2017), it’s estimated that no less than 150 million people, or about 2 percent of the world’s population, are homeless. The homeless population in the United States has grown rapidly in the past four decades. In New York City, the sheltered homeless population has increased from 12,830 in 1983 to 62,679 in 2020. Knowing the trend on the homeless population is crucial at helping the states and the cities make affordable housing plans, and other community service plans ahead of time to better prepare for the situation. This study utilized the data from New York City, examined the key factors associated with the homelessness, and developed systematic modeling to predict homeless populations of the future. Using the best model developed, named HP-RNN, an analysis on the homeless population change during the months of 2020 and 2021, which were impacted by the COVID-19 pandemic, was conducted. Moreover, HP-RNN was tested on the data from Seattle. Methods: The methodology involves four phases in developing robust prediction methods. Phase 1 gathered and analyzed raw data of homeless population and demographic conditions from five urban centers. Phase 2 identified the key factors that contribute to the rate of homelessness. In Phase 3, three models were built using Linear Regression, Random Forest, and Recurrent Neural Network (RNN), respectively, to predict the future trend of society's homeless population. Each model was trained and tuned based on the dataset from New York City for its accuracy measured by Mean Squared Error (MSE). In Phase 4, the final phase, the best model from Phase 3 was evaluated using the data from Seattle that was not part of the model training and tuning process in Phase 3. Results: Compared to the Linear Regression based model used by HUD et al (2019), HP-RNN significantly improved the prediction metrics of Coefficient of Determination (R2) from -11.73 to 0.88 and MSE by 99%. HP-RNN was then validated on the data from Seattle, WA, which showed a peak %error of 14.5% between the actual and the predicted count. Finally, the modeling results were collected to predict the trend during the COVID-19 pandemic. It shows a good correlation between the actual and the predicted homeless population, with the peak %error less than 8.6%. Conclusions and Implications: This work is the first work to apply RNN to model the time series of the homeless related data. The Model shows a close correlation between the actual and the predicted homeless population. There are two major implications of this result. First, the model can be used to predict the homeless population for the next several years, and the prediction can help the states and the cities plan ahead on affordable housing allocation and other community service to better prepare for the future. Moreover, this prediction can serve as a reference to policy makers and legislators as they seek to make changes that may impact the factors closely associated with the future homeless population trend.

Keywords: homeless, prediction, model, RNN

Procedia PDF Downloads 120

Authors: Mohammad H. Taheri, David J. Brown, Nasser Sherkat

Abstract:

Engagement is one of the most important factors in determining successful outcomes and deep learning in students. Existing approaches to detect student engagement involve periodic human observations that are subject to inter-rater reliability. Our solution uses real-time multimodal multisensor data labeled by objective performance outcomes to infer the engagement of students. The study involves four students with a combined diagnosis of cerebral palsy and a learning disability who took part in a 3-month trial over 59 sessions. Multimodal multisensor data were collected while they participated in a continuous performance test. Eye gaze, electroencephalogram, body pose, and interaction data were used to create a model of student engagement through objective labeling from the continuous performance test outcomes. In order to achieve this, a type of continuous performance test is introduced, the Seek-X type. Nine features were extracted including high-level handpicked compound features. Using leave-one-out cross-validation, a series of different machine learning approaches were evaluated. Overall, the random forest classification approach achieved the best classification results. Using random forest, 93.3% classification for engagement and 42.9% accuracy for disengagement were achieved. We compared these results to outcomes from different models: AdaBoost, decision tree, k-Nearest Neighbor, naïve Bayes, neural network, and support vector machine. We showed that using a multisensor approach achieved higher accuracy than using features from any reduced set of sensors. We found that using high-level handpicked features can improve the classification accuracy in every sensor mode. Our approach is robust to both sensor fallout and occlusions. The single most important sensor feature to the classification of engagement and distraction was shown to be eye gaze. It has been shown that we can accurately predict the level of engagement of students with learning disabilities in a real-time approach that is not subject to inter-rater reliability, human observation or reliant on a single mode of sensor input. This will help teachers design interventions for a heterogeneous group of students, where teachers cannot possibly attend to each of their individual needs. Our approach can be used to identify those with the greatest learning challenges so that all students are supported to reach their full potential.

Keywords: affective computing in education, affect detection, continuous performance test, engagement, flow, HCI, interaction, learning disabilities, machine learning, multimodal, multisensor, physiological sensors, student engagement

Procedia PDF Downloads 94

Authors: Piotr K. Szewczyk, Arkadiusz Gradys, Sungkyun Kim, Luana Persano, Mateusz M. Marzec, Oleksander Kryshtal, Andrzej Bernasik, Sohini Kar-Narayan, Pawel Sajkiewicz, Urszula Stachewicz

Abstract:

Piezoelectric polymers have received great attention in smart textiles, wearables, and flexible electronics. Their potential applications range from devices that could operate without traditional power sources, through self-powering sensors, up to implantable biosensors. Semi-crystalline PVDF is often proposed as the main candidate for industrial-scale applications as it exhibits exceptional energy harvesting efficiency compared to other polymers combined with high mechanical strength and thermal stability. Plenty of approaches have been proposed for obtaining PVDF rich in the desired β-phase with electric polling, thermal annealing, and mechanical stretching being the most prevalent. Electrospinning is a highly tunable technique that provides a one-step process of obtaining highly piezoelectric PVDF fibers without the need for post-treatment. In this study, voltage polarity and relative humidity influence on electrospun PVDF, fibers were investigated with the main focus on piezoelectric β-phase contents and piezoelectric performance. Morphology and internal structure of fibers were investigated using scanning (SEM) and transmission electron microscopy techniques (TEM). Fourier Transform Infrared Spectroscopy (FITR), wide-angle X-ray scattering (WAXS) and differential scanning calorimetry (DSC) were used to characterize the phase composition of electrospun PVDF. Additionally, surface chemistry was verified with X-ray photoelectron spectroscopy (XPS). Piezoelectric performance of individual electrospun PVDF fibers was measured using piezoresponse force microscopy (PFM), and the power output from meshes was analyzed via custom-built equipment. To prepare the solution for electrospinning, PVDF pellets were dissolved in dimethylacetamide and acetone solution in a 1:1 ratio to achieve a 24% solution. Fibers were electrospun with a constant voltage of +/-15kV applied to the stainless steel nozzle with the inner diameter of 0.8mm. The flow rate was kept constant at 6mlh⁻¹. The electrospinning of PVDF was performed at T = 25°C and relative humidity of 30 and 60% for PVDF30+/- and PVDF60+/- samples respectively in the environmental chamber. The SEM and TEM analysis of fibers produced at a lower relative humidity of 30% (PVDF30+/-) showed a smooth surface in opposition to fibers obtained at 60% relative humidity (PVDF60+/-), which had wrinkled surface and additionally internal voids. XPS results confirmed lower fluorine content at the surface of PVDF- fibers obtained by electrospinning with negative voltage polarity comparing to the PVDF+ obtained with positive voltage polarity. Changes in surface composition measured with XPS were found to influence the piezoelectric performance of obtained fibers what was further confirmed by PFM as well as by custom-built fiber-based piezoelectric generator. For PVDF60+/- samples humidity led to an increase of β-phase contents in PVDF fibers as confirmed by FTIR, WAXS, and DSC measurements, which showed almost two times higher concentrations of β-phase. A combination of negative voltage polarity with high relative humidity led to fibers with the highest β-phase contents and the best piezoelectric performance of all investigated samples. This study outlines the possibility to produce electrospun PVDF fibers with tunable piezoelectric performance in a one-step electrospinning process by controlling relative humidity and voltage polarity conditions. Acknowledgment: This research was conducted within the funding from m the Sonata Bis 5 project granted by National Science Centre, No 2015/18/E/ST5/00230, and supported by the infrastructure at International Centre of Electron Microscopy for Materials Science (IC-EM) at AGH University of Science and Technology. The PFM measurements were supported by an STSM Grant from COST Action CA17107.

Keywords: crystallinity, electrospinning, PVDF, voltage polarity

Procedia PDF Downloads 134

Authors: Marcos Silva Tavares, Jamile Raquel Regazzo, Edson José de Souza Sardinha, Murilo Mesquita Baesso

Abstract:

Common beans are a widely cultivated and consumed legume globally, serving as a staple food for humans, especially in developing countries, due to their nutritional characteristics. Nitrogen (N) is the most limiting nutrient for productivity, and foliar analysis is crucial to ensure balanced nitrogen fertilization. Excessive N applications can cause, either isolated or cumulatively, soil and water contamination, plant toxicity, and increase their susceptibility to diseases and pests. However, the quantification of N using conventional methods is time-consuming and costly, demanding new technologies to optimize the adequate supply of N to plants. Thus, it becomes necessary to establish constant monitoring of the foliar content of this macronutrient in plants, mainly at the V4 stage, aiming at precision management of nitrogen fertilization. In this work, the objective was to evaluate the performance of a deep learning model, Resnet-50, in the classification of foliar nitrogen in common beans using RGB images. The BRS Estilo cultivar was sown in a greenhouse in a completely randomized design with four nitrogen doses (T1 = 0 kg N ha-1, T2 = 25 kg N ha-1, T3 = 75 kg N ha-1, and T4 = 100 kg N ha-1) and 12 replications. Pots with 5L capacity were used with a substrate composed of 43% soil (Neossolo Quartzarênico), 28.5% crushed sugarcane bagasse, and 28.5% cured bovine manure. The water supply of the plants was done with 5mm of water per day. The application of urea (45% N) and the acquisition of images occurred 14 and 32 days after sowing, respectively. A code developed in Matlab© R2022b was used to cut the original images into smaller blocks, originating an image bank composed of 4 folders representing the four classes and labeled as T1, T2, T3, and T4, each containing 500 images of 224x224 pixels obtained from plants cultivated under different N doses. The Matlab© R2022b software was used for the implementation and performance analysis of the model. The evaluation of the efficiency was done by a set of metrics, including accuracy (AC), F1-score (F1), specificity (SP), area under the curve (AUC), and precision (P). The ResNet-50 showed high performance in the classification of foliar N levels in common beans, with AC values of 85.6%. The F1 for classes T1, T2, T3, and T4 was 76, 72, 74, and 77%, respectively. This study revealed that the use of RGB images combined with deep learning can be a promising alternative to slow laboratory analyses, capable of optimizing the estimation of foliar N. This can allow rapid intervention by the producer to achieve higher productivity and less fertilizer waste. Future approaches are encouraged to develop mobile devices capable of handling images using deep learning for the classification of the nutritional status of plants in situ.

Keywords: convolutional neural network, residual network 50, nutritional status, artificial intelligence

Procedia PDF Downloads 19

Authors: Fabiana Franceschi, Martha Cobo, Manuel Figueredo

Abstract:

Bogotá, the capital of Colombia, is its largest city and one of the most polluted in Latin America due to the fast economic growth over the last ten years. Bogotá has been affected by high pollution events which led to the high concentration of PM10 and NO2, exceeding the local 24-hour legal limits (100 and 150 g/m3 each). The most important pollutants in the city are PM10 and PM2.5 (which are associated with respiratory and cardiovascular problems) and it is known that their concentrations in the atmosphere depend on the local meteorological factors. Therefore, it is necessary to establish a relationship between the meteorological variables and the concentrations of the atmospheric pollutants such as PM10, PM2.5, CO, SO2, NO2 and O3. This study aims to determine the interrelations between meteorological variables and air pollutants in Bogotá, using data mining techniques. Data from 13 monitoring stations were collected from the Bogotá Air Quality Monitoring Network within the period 2010-2015. The Principal Component Analysis (PCA) algorithm was applied to obtain primary relations between all the parameters, and afterwards, the K-means clustering technique was implemented to corroborate those relations found previously and to find patterns in the data. PCA was also used on a per shift basis (morning, afternoon, night and early morning) to validate possible variation of the previous trends and a per year basis to verify that the identified trends have remained throughout the study time. Results demonstrated that wind speed, wind direction, temperature, and NO2 are the most influencing factors on PM10 concentrations. Furthermore, it was confirmed that high humidity episodes increased PM2,5 levels. It was also found that there are direct proportional relationships between O3 levels and wind speed and radiation, while there is an inverse relationship between O3 levels and humidity. Concentrations of SO2 increases with the presence of PM10 and decreases with the wind speed and wind direction. They proved as well that there is a decreasing trend of pollutant concentrations over the last five years. Also, in rainy periods (March-June and September-December) some trends regarding precipitations were stronger. Results obtained with K-means demonstrated that it was possible to find patterns on the data, and they also showed similar conditions and data distribution among Carvajal, Tunal and Puente Aranda stations, and also between Parque Simon Bolivar and las Ferias. It was verified that the aforementioned trends prevailed during the study period by applying the same technique per year. It was concluded that PCA algorithm is useful to establish preliminary relationships among variables, and K-means clustering to find patterns in the data and understanding its distribution. The discovery of patterns in the data allows using these clusters as an input to an Artificial Neural Network prediction model.

Keywords: air pollution, air quality modelling, data mining, particulate matter

Procedia PDF Downloads 258

Authors: Anirudh Lahiri

Abstract:

Neuromorphic computing, inspired by the intricate operations of biological neural networks, offers a revolutionary approach to overcoming the limitations of traditional computing architectures. This research proposes the integration of spintronics with neuromorphic systems, aiming to enhance computational performance, scalability, and energy efficiency. Traditional computing systems, based on the Von Neumann architecture, struggle with scalability and efficiency due to the segregation of memory and processing functions. In contrast, the human brain exemplifies high efficiency and adaptability, processing vast amounts of information with minimal energy consumption. This project explores the use of spintronics, which utilizes the electron's spin rather than its charge, to create more energy-efficient computing systems. Spintronic devices, such as magnetic tunnel junctions (MTJs) manipulated through spin-transfer torque (STT) and spin-orbit torque (SOT), offer a promising pathway to reducing power consumption and enhancing the speed of data processing. The integration of these devices within a neuromorphic framework aims to replicate the efficiency and adaptability of biological systems. The research is structured into three phases: an exhaustive literature review to build a theoretical foundation, laboratory experiments to test and optimize the theoretical models, and iterative refinements based on experimental results to finalize the system. The initial phase focuses on understanding the current state of neuromorphic and spintronic technologies. The second phase involves practical experimentation with spintronic devices and the development of neuromorphic systems that mimic synaptic plasticity and other biological processes. The final phase focuses on refining the systems based on feedback from the testing phase and preparing the findings for publication. The expected contributions of this research are twofold. Firstly, it aims to significantly reduce the energy consumption of computational systems while maintaining or increasing processing speed, addressing a critical need in the field of computing. Secondly, it seeks to enhance the learning capabilities of neuromorphic systems, allowing them to adapt more dynamically to changing environmental inputs, thus better mimicking the human brain's functionality. The integration of spintronics with neuromorphic computing could revolutionize how computational systems are designed, making them more efficient, faster, and more adaptable. This research aligns with the ongoing pursuit of energy-efficient and scalable computing solutions, marking a significant step forward in the field of computational technology.

Keywords: material science, biological engineering, mechanical engineering, neuromorphic computing, spintronics, energy efficiency, computational scalability, synaptic plasticity.

Procedia PDF Downloads 43

Authors: Ahmad S. Darwish, Tarek Lemaoui, Hanifa Taher, Inas M. AlNashef, Fawzi Banat

Abstract:

This research reports a systematic top-down approach for designing neoteric hydrophobic solvents –particularly, deep eutectic solvents (DES) and ionic liquids (IL)– as furfural extractants from aqueous media for the application of sustainable biomass conversion. The first stage of the framework entailed screening 32 neoteric solvents to determine their efficacy against toluene as the application’s conventional benchmark for comparison. The selection criteria for the best solvents encompassed not only their efficiency in extracting furfural but also low viscosity and minimal toxicity levels. Additionally, for the DESs, their natural origins, availability, and biodegradability were also taken into account. From the screening pool, two neoteric solvents were selected: thymol:decanoic acid 1:1 (Thy:DecA) and trihexyltetradecyl phosphonium bis(trifluoromethylsulfonyl) imide [P₁₄,₆,₆,₆][NTf₂]. These solvents outperformed the toluene benchmark, achieving efficiencies of 94.1% and 97.1% respectively, compared to toluene’s 81.2%, while also possessing the desired properties. These solvents were then characterized thoroughly in terms of their physical properties, thermal properties, critical properties, and cross-contamination solubilities. The selected neoteric solvents were then extensively tested under various operating conditions, and an exceptional stable performance was exhibited, maintaining high efficiency across a broad range of temperatures (15–100 °C), pH levels (1–13), and furfural concentrations (0.1–2.0 wt%) with a remarkable equilibrium time of only 2 minutes, and most notably, demonstrated high efficiencies even at low solvent-to-feed ratios. The durability of the neoteric solvents was also validated to be stable over multiple extraction-regeneration cycles, with limited leachability to the aqueous phase (≈0.1%). Moreover, the extraction performance of the solvents was then modeled through machine learning, specifically multiple non-linear regression (MNLR) and artificial neural networks (ANN). The models demonstrated high accuracy, indicated by their low absolute average relative deviations with values of 2.74% and 2.28% for Thy:DecA and [P₁₄,₆,₆,₆][NTf₂], respectively, using MNLR, and 0.10% for Thy:DecA and 0.41% for [P₁₄,₆,₆,₆][NTf₂] using ANN, highlighting the significantly enhanced predictive accuracy of the ANN. The neoteric solvents presented herein offer noteworthy advantages over traditional organic solvents, including their high efficiency in both extraction and regeneration processes, their stability and minimal leachability, making them particularly suitable for applications involving aqueous media. Moreover, these solvents are more environmentally friendly, incorporating renewable and sustainable components like thymol and decanoic acid. This exceptional efficacy of the newly developed neoteric solvents signifies a significant advancement, providing a green and sustainable alternative for furfural production from biowaste.

Keywords: sustainable biomass conversion, furfural extraction, ionic liquids, deep eutectic solvents

Procedia PDF Downloads 69

Authors: Skyler Kim

Abstract:

An early diagnosis of leukemia has always been a challenge to doctors and hematologists. On a worldwide basis, it was reported that there were approximately 350,000 new cases in 2012, and diagnosing leukemia was time-consuming and inefficient because of an endemic shortage of flow cytometry equipment in current clinical practice. As the number of medical diagnosis tools increased and a large volume of high-quality data was produced, there was an urgent need for more advanced data analysis methods. One of these methods was the AI approach. This approach has become a major trend in recent years, and several research groups have been working on developing these diagnostic models. However, designing and implementing a leukemia diagnostic system in real clinical environments based on a deep learning approach with larger sets remains complex. Leukemia is a major hematological malignancy that results in mortality and morbidity throughout different ages. We decided to select acute lymphocytic leukemia to develop our diagnostic system since acute lymphocytic leukemia is the most common type of leukemia, accounting for 74% of all children diagnosed with leukemia. The results from this development work can be applied to all other types of leukemia. To develop our model, the Kaggle dataset was used, which consists of 15135 total images, 8491 of these are images of abnormal cells, and 5398 images are normal. In this paper, we design and implement a leukemia diagnostic system in a real clinical environment based on deep learning approaches with larger sets. The proposed diagnostic system has the function of detecting and classifying leukemia. Different from other AI approaches, we explore hybrid architectures to improve the current performance. First, we developed two independent convolutional neural network models: VGG19 and ResNet50. Then, using both VGG19 and ResNet50, we developed a hybrid deep learning architecture employing transfer learning techniques to extract features from each input image. In our approach, fusing the features from specific abstraction layers can be deemed as auxiliary features and lead to further improvement of the classification accuracy. In this approach, features extracted from the lower levels are combined into higher dimension feature maps to help improve the discriminative capability of intermediate features and also overcome the problem of network gradient vanishing or exploding. By comparing VGG19 and ResNet50 and the proposed hybrid model, we concluded that the hybrid model had a significant advantage in accuracy. The detailed results of each model’s performance and their pros and cons will be presented in the conference.

Keywords: acute lymphoblastic leukemia, hybrid model, leukemia diagnostic system, machine learning

Procedia PDF Downloads 187

Authors: Atbin Mahabbati, Jason Beringer, Matthias Leopold

Abstract:

To address the global challenges of climate and environmental changes, there is a need for quantifying and reducing uncertainties in environmental data, including observations of carbon, water, and energy. Global eddy covariance flux tower networks (FLUXNET), and their regional counterparts (i.e., OzFlux, AmeriFlux, China Flux, etc.) were established in the late 1990s and early 2000s to address the demand. Despite the capability of eddy covariance in validating process modelling analyses, field surveys and remote sensing assessments, there are some serious concerns regarding the challenges associated with the technique, e.g. data gaps and uncertainties. To address these concerns, this research has developed an ensemble model to fill the data gaps of CO₂ flux to avoid the limitations of using a single algorithm, and therefore, provide less error and decline the uncertainties associated with the gap-filling process. In this study, the data of five towers in the OzFlux Network (Alice Springs Mulga, Calperum, Gingin, Howard Springs and Tumbarumba) during 2013 were used to develop an ensemble machine learning model, using five feedforward neural networks (FFNN) with different structures combined with an eXtreme Gradient Boosting (XGB) algorithm. The former methods, FFNN, provided the primary estimations in the first layer, while the later, XGB, used the outputs of the first layer as its input to provide the final estimations of CO₂ flux. The introduced model showed slight superiority over each single FFNN and the XGB, while each of these two methods was used individually, overall RMSE: 2.64, 2.91, and 3.54 g C m⁻² yr⁻¹ respectively (3.54 provided by the best FFNN). The most significant improvement happened to the estimation of the extreme diurnal values (during midday and sunrise), as well as nocturnal estimations, which is generally considered as one of the most challenging parts of CO₂ flux gap-filling. The towers, as well as seasonality, showed different levels of sensitivity to improvements provided by the ensemble model. For instance, Tumbarumba showed more sensitivity compared to Calperum, where the differences between the Ensemble model on the one hand and the FFNNs and XGB, on the other hand, were the least of all 5 sites. Besides, the performance difference between the ensemble model and its components individually were more significant during the warm season (Jan, Feb, Mar, Oct, Nov, and Dec) compared to the cold season (Apr, May, Jun, Jul, Aug, and Sep) due to the higher amount of photosynthesis of plants, which led to a larger range of CO₂ exchange. In conclusion, the introduced ensemble model slightly improved the accuracy of CO₂ flux gap-filling and robustness of the model. Therefore, using ensemble machine learning models is potentially capable of improving data estimation and regression outcome when it seems to be no more room for improvement while using a single algorithm.

Keywords: carbon flux, Eddy covariance, extreme gradient boosting, gap-filling comparison, hybrid model, OzFlux network

Procedia PDF Downloads 139

Authors: Mia Bahar, Özlem Bozkurt

Abstract:

Obsessive-compulsive disorder (OCD) is a typical, persistent, and long-lasting mental health condition in which a person experiences uncontrollable, recurrent thoughts (or "obsessions") and/or activities (or "compulsions") that they feel compelled to engage in repeatedly. Obsessive-compulsive disorder is both underdiagnosed and undertreated. It frequently manifests in a variety of medical settings and is persistent, expensive, and burdensome. Obsessive-compulsive neurosis was long believed to be a condition that offered valuable insight into the inner workings of the unconscious mind. Obsessive-compulsive disorder is now recognized as a prime example of a neuropsychiatric condition susceptible to particular pharmacotherapeutic and psychotherapy therapies and mediated by pathology in particular neural circuits. An obsessive-compulsive disorder which is called OCD, usually has two components, one cognitive and the other behavioral, although either can occur alone. Obsessions are often repetitive and intrusive thoughts that invade consciousness. These obsessions are incredibly hard to control or dismiss. People who have OCD often engage in rituals to reduce anxiety associated with intrusive thoughts. Once the ritual is formed, the person may feel extreme relief and be free from anxiety until the thoughts of contamination intrude once again. These thoughts are strengthened through a manifestation of negative reinforcement because they allow the person to avoid anxiety and obscurity. These thoughts are described as autogenous, meaning they most likely come from nowhere. These unwelcome thoughts are related to actions which we can describe as Thought Action Fusion. The thought becomes equated with an action, such as if they refuse to perform the ritual, something bad might happen, and so people perform the ritual to escape the intrusive thought. In almost all cases of OCD, the person's life gets extremely disturbed by compulsions and obsessions. Studies show OCD is an estimated 1.1% prevalence, making it a challenging issue with high co-morbidities with other issues like depressive episodes, panic disorders, and specific phobias. The first to reveal brain anomalies in OCD were numerous CT investigations, although the results were inconsistent. A few studies have focused on the orbitofrontal cortex (OFC), anterior cingulate gyrus (AC), and thalamus, structures also implicated in the pathophysiology of OCD by functional neuroimaging studies, but few have found consistent results. However, some studies have found abnormalities in the basal ganglion. There have also been some discussions that OCD might be genetic. OCD has been linked to families in studies of family aggregation, and findings from twin studies show that this relationship is somewhat influenced by genetic variables. Some Research has shown that OCD is a heritable, polygenic condition that can result from de novo harmful mutations as well as common and unusual variants. Numerous studies have also presented solid evidence in favor of a significant additive genetic component to OCD risk, with distinct OCD symptom dimensions showing both common and individual genetic risks.

Keywords: compulsions, obsessions, neuropsychiatric, genetic

Procedia PDF Downloads 64

Authors: Supria Sarkar, Vasit Sagan, Sourav Bhadra, Meghnath Pokharel, Felix B.Fritschi

Abstract:

Soybean and their derivatives are very important agricultural commodities around the world because of their wide applicability in human food, animal feed, biofuel, and industries. However, the significance of soybean production depends on the quality of the soybean seeds rather than the yield alone. Seed composition is widely dependent on plant physiological properties, aerobic and anaerobic environmental conditions, nutrient content, and plant phenological characteristics, which can be captured by high temporal resolution remote sensing datasets. Planet scope (PS) satellite images have high potential in sequential information of crop growth due to their frequent revisit throughout the world. In this study, we estimate soybean seed composition while the plants are in the field by utilizing PlanetScope (PS) satellite images and different machine learning algorithms. Several experimental fields were established with varying genotypes and different seed compositions were measured from the samples as ground truth data. The PS images were processed to extract 462 hand-crafted vegetative and textural features. Four machine learning algorithms, i.e., partial least squares (PLSR), random forest (RFR), gradient boosting machine (GBM), support vector machine (SVM), and two recurrent neural network architectures, i.e., long short-term memory (LSTM) and gated recurrent unit (GRU) were used in this study to predict oil, protein, sucrose, ash, starch, and fiber of soybean seed samples. The GRU and LSTM architectures had two separate branches, one for vegetative features and the other for textures features, which were later concatenated together to predict seed composition. The results show that sucrose, ash, protein, and oil yielded comparable prediction results. Machine learning algorithms that best predicted the six seed composition traits differed. GRU worked well for oil (R-Squared: of 0.53) and protein (R-Squared: 0.36), whereas SVR and PLSR showed the best result for sucrose (R-Squared: 0.74) and ash (R-Squared: 0.60), respectively. Although, the RFR and GBM provided comparable performance, the models tended to extremely overfit. Among the features, vegetative features were found as the most important variables compared to texture features. It is suggested to utilize many vegetation indices for machine learning training and select the best ones by using feature selection methods. Overall, the study reveals the feasibility and efficiency of PS images and machine learning for plot-level seed composition estimation. However, special care should be given while designing the plot size in the experiments to avoid mixed pixel issues.

Keywords: agriculture, computer vision, data science, geospatial technology

Procedia PDF Downloads 137

Authors: Rasha Waheeb, Bjorn Andersen, Rafa Shakir

Abstract:

Purpose The purpose of this study is to avoid delays that occur in emergency reconstruction projects especially in post disaster circumstances whether if they were natural or manmade due to their particular national and humanitarian importance. We presented a theoretical and practical concepts for projects management in the field of construction industry that deal with a range of global and local trails. This study aimed to identify the factors of effective delay in construction projects in Iraq that affect the time and the specific quality cost, and find the best solutions to address delays and solve the problem by setting parameters to restore balance in this study. 30 projects were selected in different areas of construction were selected as a sample for this study. Design/methodology/approach This study discusses the reconstruction strategies and delay in time and cost caused by different delay factors in some selected projects in Iraq (Baghdad as a case study).A case study approach was adopted, with thirty construction projects selected from the Baghdad region, of different types and sizes. Project participants from the case projects provided data about the projects through a data collection instrument distributed through a survey. Mixed approach and methods were applied in this study. Mathematical data analysis was used to construct models to predict delay in time and cost of projects before they started. The artificial neural networks analysis was selected as a mathematical approach. These models were mainly to help decision makers in construction project to find solutions to these delays before they cause any inefficiency in the project being implemented and to strike the obstacles thoroughly to develop this industry in Iraq. This approach was practiced using the data collected through survey and questionnaire data collection as information form. Findings The most important delay factors identified leading to schedule overruns were contractor failure, redesigning of designs/plans and change orders, security issues, selection of low-price bids, weather factors, and owner failures. Some of these are quite in line with findings from similar studies in other countries/regions, but some are unique to the Iraqi project sample, such as security issues and low-price bid selection. Originality/value we selected ANN’s analysis first because ANN’s was rarely used in project management , and never been used in Iraq to finding solutions for problems in construction industry. Also, this methodology can be used in complicated problems when there is no interpretation or solution for a problem. In some cases statistical analysis was conducted and in some cases the problem is not following a linear equation or there was a weak correlation, thus we suggested using the ANN’s because it is used for nonlinear problems to find the relationship between input and output data and that was really supportive.

Keywords: construction projects, delay factors, emergency reconstruction, innovation ANN, post disasters, project management

Procedia PDF Downloads 165

Authors: André Augusto Ceballos Melo

Abstract:

Design to facilitate the recognition of congruent prosthetic movements, context-to-motion translations guided by image, verbal prompt, users nonverbal communication such as facial expressions, gestures, paralinguistics, scene context, and object recognition contributes to this process though it can also be applied to other tasks, such as walking, Prosthetic limbs as assistive technology through gestures, sound codes, signs, facial, body expressions, and scene context The context-to-motion model is a machine learning approach that is designed to improve the control and dexterity of prosthetic limbs. It works by using sensory input from the prosthetic limb to learn about the dynamics of the environment and then using this information to generate smooth, stable movements. This can help to improve the performance of the prosthetic limb and make it easier for the user to perform a wide range of tasks. There are several key benefits to using the context-to-motion model for prosthetic limb control. First, it can help to improve the naturalness and smoothness of prosthetic limb movements, which can make them more comfortable and easier to use for the user. Second, it can help to improve the accuracy and precision of prosthetic limb movements, which can be particularly useful for tasks that require fine motor control. Finally, the context-to-motion model can be trained using a variety of different sensory inputs, which makes it adaptable to a wide range of prosthetic limb designs and environments. Stable diffusion is a machine learning method that can be used to improve the control and stability of movements in robotic and prosthetic systems. It works by using sensory feedback to learn about the dynamics of the environment and then using this information to generate smooth, stable movements. One key aspect of stable diffusion is that it is designed to be robust to noise and uncertainty in the sensory feedback. This means that it can continue to produce stable, smooth movements even when the sensory data is noisy or unreliable. To implement stable diffusion in a robotic or prosthetic system, it is typically necessary to first collect a dataset of examples of the desired movements. This dataset can then be used to train a machine learning model to predict the appropriate control inputs for a given set of sensory observations. Once the model has been trained, it can be used to control the robotic or prosthetic system in real-time. The model receives sensory input from the system and uses it to generate control signals that drive the motors or actuators responsible for moving the system. Overall, the use of the context-to-motion model has the potential to significantly improve the dexterity and performance of prosthetic limbs, making them more useful and effective for a wide range of users Hand Gesture Body Language Influence Communication to social interaction, offering a possibility for users to maximize their quality of life, social interaction, and gesture communication.

Keywords: stable diffusion, neural interface, smart prosthetic, augmenting

Procedia PDF Downloads 101

Authors: Sijie Chen, Xiaofang Chen, Juan Wang, Yingying Zhang, Yu Hong, Wanyu Zhuang, Xinxin Huang, Ping Ou, Longsheng Huang

Abstract:

Purpose: Autistic symptom improvement can be observed in children treated with acupuncture, but the mechanism is still being explored. In the present study, we used scalp acupuncture to treat autism rat model, and then their improvement in the abnormal behaviors and specific mechanisms behind were revealed by detecting animal behaviors, analyzing the RNA sequencing of the prefrontal cortex(PFC), and observing the ultrastructure of PFC neurons under the transmission electron microscope. Methods: On gestational day 12.5, Wistar rats were given valproic acid (VPA) by intraperitoneal injection, and their offspring were considered to be reliable rat models of autism. They were randomized to VPA or VPA-acupuncture group (n=8). Offspring of Wistar pregnant rats that were simultaneously injected with saline were randomly selected as the wild-type group (WT). VPA_acupuncture group rats received acupuncture intervention at 23 days of age for 4 weeks, and the other two groups followed without intervention. After the intervention, all experimental rats underwent behavioral tests. Immediately afterward, they were euthanized by cervical dislocation, and their prefrontal cortex was isolated for RNA sequencing and transmission electron microscopy. Results: The main results are as follows: 1. Animal behavioural tests: VPA group rats showed more anxiety-like behaviour and repetitive, stereotyped behaviour than WT group rats. While VPA group rats showed less spatial exploration ability, activity level, social interaction, and social novelty preference than WT group rats. It was gratifying to observe that acupuncture indeed improved these abnormal behaviors of autism rat model. 2. RNA-sequencing: The three groups of rats differed in the expression and enrichment pathways of multiple genes related to synaptic function, neural signal transduction, and circadian rhythm regulation. Our experiments indicated that acupuncture can alleviate the major symptoms of ASD by improving these neurological abnormalities. 3. Under the transmission electron microscopy, several lysosomes and mitochondrial structural abnormalities were observed in the prefrontal neurons of VPA group rats, which were manifested as atrophy of the mitochondrial membran, blurring or disappearance of the mitochondrial cristae, and even vacuolization. Moreover, the number of synapses and synaptic vesicles was relatively small. Conversely, the mitochondrial structure of rats in the WT group and VPA_acupuncture was normal, and the number of synapses and synaptic vesicles was relatively large. Conclusion: Acupuncture effectively improved the abnormal behaviors of autism rat model and the ultrastructure of the PFC neurons, which might worked by improving their abnormal synaptic function, synaptic plasticity and promoting neuronal signal transduction.

Keywords: autism spectrum disorder, acupuncture, animal behavior, RNA sequencing, transmission electron microscope

Procedia PDF Downloads 45