Search results for: KaraAgroAI cocoa dataset

Authors: Hannes Koester

Abstract:

Driven by difficult market conditions and increasing regulations, many banks are making the strategic decision to restructure their portfolio by divesting several business segments. Using a unique dataset of 727 portfolio restructuring announcements by 161 international listed banks over the period 1999 to 2015, we investigate the impact of restructuring measurements on the stock performance as well as on the banks’ profitability and risk. Employing the event study methodology, we detect positive stock market reactions on the announcement of restructuring measurements. These positive stock market reactions indicate that shareholders reward banks’ specialization activities. However, the results of the system GMM regressions show a negative relation between restructuring measurements and banks’ return on assets and a positive relation towards the individual and systemic risk of banks. These empirical results indicate that there is no guarantee that portfolio restructurings will result in a more profitable and less risky institution.

Keywords: bank performance, bank risk, divestiture, restructuring, systemic risk

Procedia PDF Downloads 317

Authors: Obe Olumide Olayinka, Victor Balanica, Eugen Neagoe

Abstract:

The effects of hypertension are often lethal thus its early detection and prevention is very important for everybody. In this paper, a neural network (NN) model was developed and trained based on a dataset of hypertension causative parameters in order to forecast the likelihood of occurrence of hypertension in patients. Our research goal was to analyze the potential of the presented NN to predict, for a period of time, the risk of hypertension or the risk of developing this disease for patients that are or not currently hypertensive. The results of the analysis for a given patient can support doctors in taking pro-active measures for averting the occurrence of hypertension such as recommendations regarding the patient behavior in order to lower his hypertension risk. Moreover, the paper envisages a set of three example scenarios in order to determine the age when the patient becomes hypertensive, i.e. determine the threshold for hypertensive age, to analyze what happens if the threshold hypertensive age is set to a certain age and the weight of the patient if being varied, and, to set the ideal weight for the patient and analyze what happens with the threshold of hypertensive age.

Keywords: neural network, hypertension, data set, training set, supervised learning

Procedia PDF Downloads 392

Authors: Amir Taghizadeh Vahed, Adithya Thaduri

Abstract:

Shovels are the main components in the mining transportation system. The productivity of the mines depends on the availability of shovels due to its high capital and operating costs. The unplanned failure/shutdowns of a shovel results in higher repair costs, increase in downtime, as well as increasing indirect cost (i.e. loss of production and company’s reputation). In order to mitigate these failures, predictive maintenance can be useful approach using failure prediction. The modern mining machinery or shovels collect huge datasets automatically; it consists of reliability and maintenance data. However, the gathered datasets are useless until the information and knowledge of data are extracted. Machine learning as well as data mining, which has a major role in recent studies, has been used for the knowledge discovery process. In this study, data mining and machine learning approaches are implemented to detect not only anomalies but also patterns from a dataset and further detection of failures.

Keywords: maintenance, machine learning, shovel, conditional based monitoring

Procedia PDF Downloads 220

Authors: Adrià Arbués-Sangüesa, Coloma Ballester, Gloria Haro

Abstract:

Tracking sports players is a widely challenging scenario, specially in single-feed videos recorded in tight courts, where cluttering and occlusions cannot be avoided. This paper presents an analysis of several geometric and semantic visual features to detect and track basketball players. An ablation study is carried out and then used to remark that a robust tracker can be built with Deep Learning features, without the need of extracting contextual ones, such as proximity or color similarity, nor applying camera stabilization techniques. The presented tracker consists of: (1) a detection step, which uses a pretrained deep learning model to estimate the players pose, followed by (2) a tracking step, which leverages pose and semantic information from the output of a convolutional layer in a VGG network. Its performance is analyzed in terms of MOTA over a basketball dataset with more than 10k instances.

Keywords: basketball, deep learning, feature extraction, single-camera, tracking

Procedia PDF Downloads 138

Authors: Aymen Sekhri, Emmanuel Kwabena Frimpong, Bolaji Mubarak Ayeyemi, Aleksi Hirvonen, Matias Hirvonen, Tedros Tesfay Andemichael

Abstract:

Fixation identification, which involves isolating and identifying fixations and saccades in eye-tracking protocols, is an important aspect of eye-movement data processing that can have a big impact on higher-level analyses. However, fixation identification techniques are frequently discussed informally and rarely compared in any meaningful way. With two state-of-the-art algorithms, we will implement fixation detection and analysis in this work. The velocity threshold fixation algorithm is the first algorithm, and it identifies fixation based on a threshold value. For eye movement detection, the second approach is U'n' Eye, a deep neural network algorithm. The goal of this project is to analyze and visualize eye-tracking data from an eye gaze dataset that has been provided. The data was collected in a scenario in which individuals were shown photos and asked whether or not they recognized them. The results of the two-fixation detection approach are contrasted and visualized in this paper.

Keywords: human-computer interaction, eye-tracking, CNN, fixations, saccades

Procedia PDF Downloads 135

Authors: Pei-Gi Shu, Yin-Hua Yeh, Chao-Ting Chen

Abstract:

Using a dataset consisting of 7,120 firm-year observations from the Taiwan stock market over the 2007-2011 sample period, we find a significantly negative relationship between board external connectedness and firm diversification. We propose a learningeffect hypothesis indicating that an externally connected board member’s experiences in other companies directly affect his recommendations regarding the underlying firm’s diversification. The partial correlation between diversification and the performance of firms with externally connected board members is used as a proxy for the learning effect. The empirical results show that the learning effect is asymmetrically embedded in firm diversification, with negative experiences having a greater effect on firm diversification than positive experiences. Externally connected board members are associated with reduced diversification in one firm after they learn that diversification is detrimental to value in other companies. Moreover, the diversification of a firm due to board external connectedness is moderated by the controlling owner’s interest alignment and entrenchment.

Keywords: board, external, connectedness, diversification

Procedia PDF Downloads 462

Authors: Mohammad Mohammadi, Mojtaba Askari, Mohammad Farouq Hossaini

Abstract:

An intrinsic property of rock mass known as rock bolt supporting factor (RSF) or rock bolting capability of rock mass was developed and used for explanation of the mechanism of rock bolting practice. Based on the theory of RSF, numeral values can be assigned to each given rock mass to show the capability of that rock mass to be reinforced by rock bolting. For determination of shotcrete thickness, both safety and cost must be taken into account. The present paper introduces a scientific approach for determination of the necessary shotcrete thickness in underground structures for support purposes using the concept of rock bolt supporting factor (RSF). The proposed approach makes the outcome of shotcrete design one step more accurate than before. The actual dataset of 500 meters of Alborz Tunnel length is used as an example of the application of the approach.

Keywords: rock bolt supporting factor (RSF), shotcrete design, underground excavation, Alborz Tunnel

Procedia PDF Downloads 320

Authors: Mouhammd Alkasassbeh, Alaa Lasasmeh

Abstract:

We live in an era of rapid development as a result of significant scientific growth. Like other technologies, wireless sensor networks (WSNs) are playing one of the main roles. Based on WSNs, ZigBee adds many features to devices, such as minimum cost and power consumption, and increasing the range and connect ability of sensor nodes. ZigBee technology has come to be used in various fields, including science, engineering, and networks, and even in medicinal aspects of intelligence building. In this work, we generated two main datasets, the first being based on tree topology and the second on star topology. The datasets were evaluated by three machine learning (ML) algorithms: J48, meta.j48 and multilayer perceptron (MLP). Each topology was classified into normal and abnormal (attack) network traffic. The dataset used in our work contained simulated data from network simulation 2 (NS2). In each database, the Bayesian network meta.j48 classifier achieved the highest accuracy level among other classifiers, of 99.7% and 99.2% respectively.

Keywords: IDS, Machine learning, WSN, ZigBee technology

Procedia PDF Downloads 543

Authors: Bhaveek Maini, Sanjay Dhanka, Surita Maini

Abstract:

Electroencephalography (EEG) is used to classify several epileptic seizures worldwide. It is a very crucial task for the neurologist to identify the epileptic seizure with manual EEG analysis, as it takes lots of effort and time. Human error is always at high risk in EEG, as acquiring signals needs manual intervention. Disease diagnosis using machine learning (ML) has continuously been explored since its inception. Moreover, where a large number of datasets have to be analyzed, ML is acting as a boon for doctors. In this research paper, authors proposed two different ML models, i.e., logistic regression (LR) and Naïve Bayes (NB), to predict epileptic seizures based on general parameters. These two techniques are applied to the epileptic seizures recognition dataset, available on the UCI ML repository. The algorithms are implemented on an 80:20 train test ratio (80% for training and 20% for testing), and the performance of the model was validated by 10-fold cross-validation. The proposed study has claimed accuracy of 81.87% and 95.49% for LR and NB, respectively.

Keywords: epileptic seizure recognition, logistic regression, Naïve Bayes, machine learning

Procedia PDF Downloads 61

Authors: Tanu Aneja, Harsha Malaviya

Abstract:

Capturing pose information in cricket shots poses several challenges, such as low-resolution videos, noisy data, and joint occlusions caused by the nature of the shots. In response to these challenges, we propose a CondDGConv-based framework specifically for cricket shot prediction. By analyzing the spatial-temporal relationships in batsman shot sequences from an annotated 2D cricket dataset, our model achieves a 97% accuracy in predicting shot types. This performance is made possible by conditioning the graph network on batsman 2D poses, allowing for precise prediction of shot outcomes based on pose dynamics. Our approach highlights the potential for enhancing shot prediction in cricket analytics, offering a robust solution for overcoming pose-related challenges in sports analysis.

Keywords: action recognition, cricket. sports video analytics, computer vision, graph convolutional networks

Procedia PDF Downloads 18

Authors: Devendra Kr Tayal, Vrinda Gupta, Aastha Bansal, Khushi Singh, Sristi Sharma, Hunny Gaur

Abstract:

In today's world, suicide is a serious problem. In order to save lives, early suicide attempt detection and prevention should be addressed. A good number of at-risk people utilize social media platforms to talk about their issues or find knowledge on related chores. Twitter and Reddit are two of the most common platforms that are used for expressing oneself. Extensive research has already been done in this field. Through supervised classification techniques like Nave Bayes, Bernoulli Nave Bayes, and Multiple Layer Perceptron on a Reddit dataset, we demonstrate the early recognition of suicidal ideation. We also performed comparative analysis on these approaches and used accuracy, recall score, F1 score, and precision score for analysis.

Keywords: machine learning, suicide ideation detection, supervised classification, natural language processing

Procedia PDF Downloads 90

Authors: Sam Khozama, Ali M. Mayya

Abstract:

Breast cancer is one of the most common types in women. Early prediction of breast cancer helps physicians detect cancer in its early stages. Big cancer data needs a very powerful tool to analyze and extract predictions. Machine learning and deep learning are two of the most efficient tools for predicting cancer based on textual data. In this study, we developed a fusion model of two machine learning and deep learning models. To obtain the final prediction, Long-Short Term Memory (LSTM) and ensemble learning with hyper parameters optimization are used, and score-level fusion is used. Experiments are done on the Breast Cancer Surveillance Consortium (BCSC) dataset after balancing and grouping the class categories. Five different training scenarios are used, and the tests show that the designed fusion model improved the performance by 3.3% compared to the individual models.

Keywords: machine learning, deep learning, cancer prediction, breast cancer, LSTM, fusion

Procedia PDF Downloads 163

Authors: Asyraf Abdul Halim, Mohd Edil Abd Sukor, Obiyathulla Ismath Bacha

Abstract:

This paper attempts to investigate the capital structure behavior of Shariah compliant firms of various levels as well those firms who are consistently Shariah non-compliant in Malaysia. The paper utilizes a unique dataset of firms of the heterogeneous level of Shariah-compliancy status over a 20 year period from the year 1997 to 2016. The paper focuses on the effects of dynamic forces behind capital structure variation such as the optimal capital structure behavior based on the trade-off, pecking order, market timing and firmly fixed effect models of capital structure. This study documents significant evidence in support of the trade-off theory with a high speed of adjustment (SOA) as well as for the time-invariant firm fixed effects across all Shariah compliance group.

Keywords: capital structure, market timing, trade-off theory, equity risk premium, Shariah-compliant firms

Procedia PDF Downloads 313

Authors: Michael Vallely

Abstract:

This paper uses data from waves 1 to 10 (2009-2019) of the UK Household Longitudinal Study to examine the social origin pay gap in the UK labour market. We find that regardless of how we proxy social origin, whether it be using the dominance approach, total parental occupation, parental education, total parental education, or the higher parental occupation and higher parental education, the results have one thing in common; in all cases, we observe a significant social origin pay gap for those from the lower social origins with the largest pay gap observed for those from the ‘lowest’ social origin. The results may indicate that when we consider the occupational status and education of both parents, previous estimates of social origin pay gaps and the number of individuals affected may have been underestimated. We also observe social origin pay gaps within educational attainment groups, such as degree holders, and within professional and managerial occupations. Therefore, this paper makes a valuable contribution to the social origin pay gap literature as it provides empirical evidence of a social origin pay gap using a large-scale UK dataset and challenges the argument that education is the great ‘social leveller’.

Keywords: social class, social origin, pay gaps, wage inequality

Procedia PDF Downloads 144

Authors: Emad A. Mohammed

Abstract:

Utilizing the capabilities of Data Mining and Artificial Intelligence in the prediction of the minimum miscibility pressure (MMP) required for multi-contact miscible (MCM) displacement of reservoir petroleum by hydrocarbon gas flooding using Fuzzy Logic models and Artificial Neural Network models will help a lot in giving accurate results. The factors affecting the (MMP) as it is proved from the literature and from the dataset are as follows: XC2-6: Intermediate composition in the oil-containing C2-6, CO2 and H2S, in mole %, XC1: Amount of methane in the oil (%),T: Temperature (°C), MwC7+: Molecular weight of C7+ (g/mol), YC2+: Mole percent of C2+ composition in injected gas (%), MwC2+: Molecular weight of C2+ in injected gas. Fuzzy Logic and Neural Networks have been used widely in prediction and classification, with relatively high accuracy, in different fields of study. It is well known that the Fuzzy Inference system can handle uncertainty within the inputs such as in our case. The results of this work showed that our proposed models perform better with higher performance indices than other emprical correlations.

Keywords: MMP, gas flooding, artificial intelligence, correlation

Procedia PDF Downloads 144

Authors: Kianoosh Ahadi, Mustafa Ergil

Abstract:

In this study, the event of vibrations caused by vortexes and the distribution of induced hydrodynamic forces due to vortexes on the sliding gate valves has been investigated. For this reason, a sliding valve with the help of computational fluid dynamics (CFD) software was simulated in two-dimensional )2D(, where the flow and turbulence equations were solved for three different valve openings (full, half, and 16.7 %) models. The variety of vortexes formed within the vicinity of the valve structure was investigated based on time where the trend of fluctuations and their occurrence regions have been detected. From the gathered solution dataset of the numerical simulations, the pressure coefficient (CP), the lift force coefficient (CL), the drag force coefficient (CD), and the momentum coefficient due to hydrodynamic forces (CM) were examined, and relevant figures were generated were from these results, the vortex-induced vibrations were analyzed.

Keywords: induced vibrations, computational fluid dynamics, sliding gate valves, vortexes

Procedia PDF Downloads 120

Authors: Haojie Wang, Xun Li, Rui Yin

Abstract:

In order to solve the memorization overfitting in the meta-learning MAML algorithm, a method of generating mutually exclusive tasks based on data augmentation is proposed. This method generates a mutex task by corresponding one feature of the data to multiple labels, so that the generated mutex task is inconsistent with the data distribution in the initial dataset. Because generating mutex tasks for all data will produce a large number of invalid data and, in the worst case, lead to exponential growth of computation, this paper also proposes a key data extraction method, that only extracts part of the data to generate the mutex task. The experiments show that the method of generating mutually exclusive tasks can effectively solve the memorization overfitting in the meta-learning MAML algorithm.

Keywords: data augmentation, mutex task generation, meta-learning, text classification.

Procedia PDF Downloads 94

Authors: Khalil Bachiri, Ali Yahyaouy, Nicoleta Rogovschi

Abstract:

Improved user experience is a goal of contemporary recommender systems. Recommender systems are starting to incorporate reinforcement learning since it easily satisfies this goal of increasing a user’s reward every session. In this paper, we examine the most effective Reinforcement Learning agent tactics on the Movielens (1M) dataset, balancing precision and a variety of recommendations. The absence of variability in final predictions makes simplistic techniques, although able to optimize ranking quality criteria, worthless for consumers of the recommendation system. Utilizing the stochasticity of Leonard-Ornstein processes, our suggested strategy encourages the agent to investigate its surroundings. Research demonstrates that raising the NDCG (Discounted Cumulative Gain) and HR (HitRate) criterion without lowering the Ornstein-Uhlenbeck process drift coefficient enhances the diversity of suggestions.

Keywords: recommender systems, reinforcement learning, deep learning, DDPG, Leonard-Ornstein process

Procedia PDF Downloads 143

Authors: Dalwinder Singh, Amandeep Verma, Manpreet Kaur, Birmohan Singh

Abstract:

The detection of pre-cancerous changes using a Pap smear test of cervical cell is the important step for the early diagnosis of cervical cancer. The Pap smear test consists of a sample of human cells taken from the cervix which are analysed to detect cancerous and pre-cancerous stage of the given subject. The manual analysis of these cells is labor intensive and time consuming process which relies on expert cytotechnologist. In this paper, a computer assisted system for the automated analysis of the cervical cells has been proposed. We propose a morphology based approach to the nucleus detection and segmentation of the cytoplasmic region of the given single or multiple overlapped cell. Further, various texture and region based features are calculated from these cells to classify these into normal and abnormal cell. Experimental results on public available dataset show that our system has achieved satisfactory success rate.

Keywords: cervical cancer, cervical tissue, mathematical morphology, texture features

Procedia PDF Downloads 527

Authors: Yufen Qin

Abstract:

Topic-to-Essay generation is a challenging task in Natural language processing, which aims to generate novel, diverse, and topic-related text based on user input. Previous research has overlooked the generation of articles under the constraints of event elements, resulting in issues such as incomplete event elements and logical inconsistencies in the generated results. To fill this gap, this paper proposes an event-constrained approach for a topic-to-essay generation that enforces the completeness of event elements during the generation process. Additionally, a language model is employed to verify the logical consistency of the generated results. Experimental results demonstrate that the proposed model achieves a better BLEU-2 score and performs better than the baseline in terms of subjective evaluation on a real dataset, indicating its capability to generate higher-quality topic-related text.

Keywords: event element, language model, natural language processing, topic-to-essay generation.

Procedia PDF Downloads 236

Authors: Guang-Ho Cha

Abstract:

This paper presents a context-sensitive media similarity search algorithm. One of the central problems regarding media search is the semantic gap between the low-level features computed automatically from media data and the human interpretation of them. This is because the notion of similarity is usually based on high-level abstraction but the low-level features do not sometimes reflect the human perception. Many media search algorithms have used the Minkowski metric to measure similarity between image pairs. However those functions cannot adequately capture the aspects of the characteristics of the human visual system as well as the nonlinear relationships in contextual information given by images in a collection. Our search algorithm tackles this problem by employing a similarity measure and a ranking strategy that reflect the nonlinearity of human perception and contextual information in a dataset. Similarity search in an image database based on this contextual information shows encouraging experimental results.

Keywords: context-sensitive search, image search, similarity ranking, similarity search

Procedia PDF Downloads 365

Authors: O. J. Akinsola, Yinan Zheng, Rose Anorlu, F. T. Ogunsola, Lifang Hou, Robert Leo-Murphy

Abstract:

Cervical Cancer (CC) is the 2nd most common cancer among women living in low and middle-income countries, with no associated symptoms during formative periods. With the advancement and innovative medical research, there are numerous preventive measures being utilized, but the incidence of cervical cancer cannot be truncated with the application of only screening tests. The mortality associated with this invasive cervical cancer can be nipped in the bud through the important role of early-stage detection. This study research selected an array of different top features selection techniques which was aimed at developing a model that could validly diagnose the risk factors of cervical cancer. A retrospective clinic-based cohort study was conducted on 178 HIV-associated cervical cancer patients in Lagos University teaching Hospital, Nigeria (U54 data repository) in April 2022. The outcome measure was the automated prediction of the HIV-associated cervical cancer cases, while the predictor variables include: demographic information, reproductive history, birth control, sexual history, cervical cancer screening history for invasive cervical cancer. The proposed technique was assessed with R and Python programming software to produce the model by utilizing the classification algorithms for the detection and diagnosis of cervical cancer disease. Four machine learning classification algorithms used are: the machine learning model was split into training and testing dataset into ratio 80:20. The numerical features were also standardized while hyperparameter tuning was carried out on the machine learning to train and test the data. Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), and K-Nearest Neighbor (KNN). Some fitting features were selected for the detection and diagnosis of cervical cancer diseases from selected characteristics in the dataset using the contribution of various selection methods for the classification cervical cancer into healthy or diseased status. The mean age of patients was 49.7±12.1 years, mean age at pregnancy was 23.3±5.5 years, mean age at first sexual experience was 19.4±3.2 years, while the mean BMI was 27.1±5.6 kg/m2. A larger percentage of the patients are Married (62.9%), while most of them have at least two sexual partners (72.5%). Age of patients (OR=1.065, p<0.001**), marital status (OR=0.375, p=0.011**), number of pregnancy live-births (OR=1.317, p=0.007**), and use of birth control pills (OR=0.291, p=0.015**) were found to be significantly associated with HIV-associated cervical cancer. On top ten 10 features (variables) considered in the analysis, RF claims the overall model performance, which include: accuracy of (72.0%), the precision of (84.6%), a recall of (84.6%) and F1-score of (74.0%) while LR has: an accuracy of (74.0%), precision of (70.0%), recall of (70.0%) and F1-score of (70.0%). The RF model identified 10 features predictive of developing cervical cancer. The age of patients was considered as the most important risk factor, followed by the number of pregnancy livebirths, marital status, and use of birth control pills, The study shows that data mining techniques could be used to identify women living with HIV at high risk of developing cervical cancer in Nigeria and other sub-Saharan African countries.

Keywords: associated cervical cancer, data mining, random forest, logistic regression

Procedia PDF Downloads 84

Authors: Shuqi Zhang

Abstract:

Federated learning allows users to participate in collaborative model training without sending data to third-party servers, reducing the risk of user data privacy leakage, and is widely used in smart finance and smart healthcare. However, the distributed architecture design of federation learning itself and the existence of secure aggregation protocols make it inherently vulnerable to backdoor attacks. To solve this problem, the federated learning backdoor defense framework FLEX based on group aggregation, cluster analysis, and neuron pruning is proposed, and inter-compatibility with secure aggregation protocols is achieved. The good performance of FLEX is verified by building a horizontal federated learning framework on the CIFAR-10 dataset for experiments, which achieves 98% success rate of backdoor detection and reduces the success rate of backdoor tasks to 0% ~ 10%.

Keywords: federated learning, secure aggregation, backdoor attack, cluster analysis, neuron pruning

Procedia PDF Downloads 96

Authors: Shao-Di Feng, Von-Wun Soo

Abstract:

To generate symbolic music with specific emotions is a challenging task due to symbolic music datasets that have emotion labels are scarce and incomplete. This research aims to generate more refined emotions based on the training datasets that are only labeled with four quadrants in Russel’s 2D emotion model. We focus on the theory of Music Fadernet and map arousal and valence to the low-level attributes, and build a symbolic music generation model by combining transformer and GM-VAE. We adopt an in-attention mechanism for the model and improve it by allowing modulation by conditional information. And we show the music generation model could control the generation of music according to the emotions specified by users in terms of high-level linguistic expression and by manipulating their corresponding low-level musical attributes. Finally, we evaluate the model performance using a pre-trained emotion classifier against a pop piano midi dataset called EMOPIA, and by subjective listening evaluation, we demonstrate that the model could generate music with more refined emotions correctly.

Keywords: music generation, music emotion controlling, deep learning, semi-supervised learning

Procedia PDF Downloads 90

Authors: Evy Rombaut, Marie-Anne Guerry

Abstract:

Predicting voluntary turnover of employees is an important topic of study, both in academia and industry. Researchers try to uncover determinants for a broader understanding and possible prevention of turnover. In the current study, we use a data set based approach to reveal determinants for turnover, differing for blue and white collar workers. Our data set based approach made it possible to study actual turnover for more than 500000 employees in 15692 Belgian corporations. We use logistic regression to calculate individual turnover probabilities and test the goodness of our model with the AUC (area under the ROC-curve) method. The results of the study confirm the relationship of known determinants to employee turnover such as age, seniority, pay and work distance. In addition, the study unravels unknown and verifies known differences between blue and white collar workers. It shows opposite relationships to turnover for gender, marital status, the number of children, nationality, and pay.

Keywords: employee turnover, blue collar, white collar, dataset analysis

Procedia PDF Downloads 291

Authors: Haojie Wang, Xun Li, Rui Yin

Abstract:

In order to solve the memorization overfitting in the model-agnostic meta-learning MAML algorithm, a method of generating mutually exclusive tasks based on data augmentation is proposed. This method generates a mutex task by corresponding one feature of the data to multiple labels so that the generated mutex task is inconsistent with the data distribution in the initial dataset. Because generating mutex tasks for all data will produce a large number of invalid data and, in the worst case, lead to an exponential growth of computation, this paper also proposes a key data extraction method that only extract part of the data to generate the mutex task. The experiments show that the method of generating mutually exclusive tasks can effectively solve the memorization overfitting in the meta-learning MAML algorithm.

Keywords: mutex task generation, data augmentation, meta-learning, text classification.

Procedia PDF Downloads 143

Authors: Ousmane Youme, Jean Marie Dembele, Eugene Ezin, Christophe Cambier

Abstract:

In recent years, the convolutional neural networks (CNN) architectures designed by evolution algorithms have proven to be competitive with handcrafted architectures designed by experts. However, these algorithms need a lot of computational power, which is beyond the capabilities of most researchers and engineers. To overcome this problem, we propose an evolution architecture under length constraints. It consists of two algorithms: a search length strategy to find an optimal space and a search architecture strategy based on a genetic algorithm to find the best individual in the optimal space. Our algorithms drastically reduce resource costs and also keep good performance. On the Cifar-10 dataset, our framework presents outstanding performance with an error rate of 5.12% and only 4.6 GPU a day to converge to the optimal individual -22 GPU a day less than the lowest cost automatic evolutionary algorithm in the peer competition.

Keywords: CNN architecture, genetic algorithm, evolution algorithm, length constraints

Procedia PDF Downloads 128

Authors: Phawin Sangsuvan, Chutimet Srinilta

Abstract:

This Paper presents the applications of Particle Swarm Optimization (PSO) Method for Thai optical character recognition (OCR). OCR consists of the pre-processing, character recognition and post-processing. Before enter into recognition process. The Character must be “Prepped” by pre-processing process. The PSO is an optimization method that belongs to the swarm intelligence family based on the imitation of social behavior patterns of animals. Route of each particle is determined by an individual data among neighborhood particles. The interaction of the particles with neighbors is the advantage of Particle Swarm to determine the best solution. So PSO is interested by a lot of researchers in many difficult problems including character recognition. As the previous this research used a Projection Histogram to extract printed digits features and defined the simple Fitness Function for PSO. The results reveal that PSO gives 67.73% for testing dataset. So in the future there can be explored enhancement the better performance of PSO with improve the Fitness Function.

Keywords: character recognition, histogram projection, particle swarm optimization, pattern recognition techniques

Procedia PDF Downloads 477

Authors: Rema Daher, Theodor Chakhachiro, Daniel Asmar

Abstract:

In the era of autonomous robot mapping, assessing the goodness of the generated maps is important, and is usually performed by aligning them to ground truth. Map alignment is difficult for two reasons: first, the query maps can be significantly distorted from ground truth, and second, establishing what constitutes ground truth for different settings is challenging. Most map alignment techniques to this date have addressed the first problem, while paying too little importance to the second. In this paper, we propose a benchmark dataset, which consists of synthetically transformed maps with their corresponding displacement fields. Furthermore, we propose a new system for comparison, where the displacement field of any map alignment technique can be computed and compared to the ground truth using statistical measures. The local information in displacement fields renders the evaluation system applicable to any alignment technique, whether it is linear or not. In our experiments, the proposed method was applied to different alignment methods from the literature, allowing for a comparative assessment between them all.

Keywords: assessment methods, benchmark, image deformation, map alignment, robot mapping, robot motion

Procedia PDF Downloads 120

Authors: Sheldon Liu, Gordon Wang, Lei Liu, Xuefeng Liu

Abstract:

Anomaly detection, also referred to as one-class classification, plays a crucial role in identifying product images that deviate from the expected distribution. This study introduces Data-centric Anomaly Detection with Diffusion Models (DCADDM), presenting a systematic strategy for data collection and further diversifying the data with image generation via diffusion models. The algorithm addresses data collection challenges in real-world scenarios and points toward data augmentation with the integration of generative AI capabilities. The paper explores the generation of normal images using diffusion models. The experiments demonstrate that with 30% of the original normal image size, modeling in an unsupervised setting with state-of-the-art approaches can achieve equivalent performances. With the addition of generated images via diffusion models (10% equivalence of the original dataset size), the proposed algorithm achieves better or equivalent anomaly localization performance.

Keywords: diffusion models, anomaly detection, data-centric, generative AI

Procedia PDF Downloads 83