Search results for: generative model

Authors: Heeba A. Gurku

Abstract:

Introduction: Cone Beam CT(CBCT) images play an integral part in proper patient positioning in cancer patients undergoing radiation therapy treatment. But these images are low in quality. The purpose of this study is to generate high-quality synthetic CT images from CBCT using generative models. Material and Methods: This study utilized two datasets from The Cancer Imaging Archive (TCIA) 1) Lung cancer dataset of 20 patients (with full view CBCT images) and 2) Pancreatic cancer dataset of 40 patients (only 27 patients having limited view images were included in the study). Cycle Generative Adversarial Networks (GAN) and its variant Attention Guided Generative Adversarial Networks (AGGAN) models were used to generate the synthetic CTs. Models were evaluated by visual evaluation and on four metrics, Structural Similarity Index Measure (SSIM), Peak Signal Noise Ratio (PSNR) Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), to compare the synthetic CT and original CT images. Results: For pancreatic dataset with limited view CBCT images, our study showed that in Cycle GAN model, MAE, RMSE, PSNR improved from 12.57to 8.49, 20.94 to 15.29 and 21.85 to 24.63, respectively but structural similarity only marginally increased from 0.78 to 0.79. Similar, results were achieved with AGGAN with no improvement over Cycle GAN. However, for lung dataset with full view CBCT images Cycle GAN was able to reduce MAE significantly from 89.44 to 15.11 and AGGAN was able to reduce it to 19.77. Similarly, RMSE was also decreased from 92.68 to 23.50 in Cycle GAN and to 29.02 in AGGAN. SSIM and PSNR also improved significantly from 0.17 to 0.59 and from 8.81 to 21.06 in Cycle GAN respectively while in AGGAN SSIM increased to 0.52 and PSNR increased to 19.31. In both datasets, GAN models were able to reduce artifacts, reduce noise, have better resolution, and better contrast enhancement. Conclusion and Recommendation: Both Cycle GAN and AGGAN were significantly able to reduce MAE, RMSE and PSNR in both datasets. However, full view lung dataset showed more improvement in SSIM and image quality than limited view pancreatic dataset.

Keywords: CT images, CBCT images, cycle GAN, AGGAN

Procedia PDF Downloads 55

Authors: Rao Fu, Jie Chen, Shutian Zeng, Yiping Zhuang, Agus Sudjianto

Abstract:

Generative Adversarial Net (GAN) has proved to be a powerful machine learning tool in image data analysis and generation. In this paper, we propose to use Conditional Generative Adversarial Net (CGAN) to learn and simulate time series data. The conditions include both categorical and continuous variables with different auxiliary information. Our simulation studies show that CGAN has the capability to learn different types of normal and heavy-tailed distributions, as well as dependent structures of different time series. It also has the capability to generate conditional predictive distributions consistent with training data distributions. We also provide an in-depth discussion on the rationale behind GAN and the neural networks as hierarchical splines to establish a clear connection with existing statistical methods of distribution generation. In practice, CGAN has a wide range of applications in market risk and counterparty risk analysis: it can be applied to learn historical data and generate scenarios for the calculation of Value-at-Risk (VaR) and Expected Shortfall (ES), and it can also predict the movement of the market risk factors. We present a real data analysis including a backtesting to demonstrate that CGAN can outperform Historical Simulation (HS), a popular method in market risk analysis to calculate VaR. CGAN can also be applied in economic time series modeling and forecasting. In this regard, we have included an example of hypothetical shock analysis for economic models and the generation of potential CCAR scenarios by CGAN at the end of the paper.

Keywords: conditional generative adversarial net, market and credit risk management, neural network, time series

Procedia PDF Downloads 110

Authors: Touila Ahmed, Elie Louis, Hamza Gharbi

Abstract:

State of the art question answering models have recently shown impressive performance especially in a zero-shot setting. This approach is particularly useful when confronted with a highly diverse domain such as the legal field, in which it is increasingly difficult to have a dataset covering every notion and concept. In this work, we propose a flexible generative question answering approach to contract analysis as well as a weakly supervised procedure to leverage unannotated data and boost our models’ performance in general, and their zero-shot performance in particular.

Keywords: question answering, contract analysis, zero-shot, natural language processing, generative models, self-supervision

Procedia PDF Downloads 144

Authors: Luz Estrada

Abstract:

The currents of evolutionary materialistic thought have argued that knowledge about an object is not obtained through the abstractive method. That is, the object cannot come to be understood if founded upon itself, nor does it take place by the encounter between form and matter. According to this affirmation, the research presented here identified as a problematic situation the absence of comprehension of the formal creation as a generative operation. This has been referred to as a recurrent lack in the production of objects and corresponds to the need to conceive the configurative process from the reality of its genesis. In this case, it is of interest to explore ways of creation that consider the object as if it were a living organism, as well as responding to the object’s experience as embodied in the designer since it unfolds its genesis simultaneously to the ways of existence of those who are involved in the generative experience.

Keywords: architecture, theoretical graphics, evolutionary genetics, formal perception

Procedia PDF Downloads 86

Authors: Minh-Thuan Nguyen, Phuong-Thai Nguyen, Van-Vinh Nguyen, Quang-Minh Nguyen

Abstract:

Research on automatically generating descriptions for e-commerce products is gaining increasing attention in recent years. However, the generated descriptions of their systems are often less informative and attractive because of lacking training datasets or the limitation of these approaches, which often use templates or statistical methods. In this paper, we explore a method to generate production descriptions by using the GPT-2 model. In addition, we apply text paraphrasing and task-adaptive pretraining techniques to improve the qualify of descriptions generated from the GPT-2 model. Experiment results show that our models outperform the baseline model through automatic evaluation and human evaluation. Especially, our methods achieve a promising result not only on the seen test set but also in the unseen test set.

Keywords: GPT-2, product description, transformer, task-adaptive, language model, pretraining

Procedia PDF Downloads 169

Authors: Rita Chattopadhyay, Vivek Anand Thoutam

Abstract:

Speech is the most acceptable means of communication where we can quickly exchange our feelings and thoughts. Quite often, people can communicate orally but cannot interact or work with computers or devices. It’s easy and quick to give speech commands than typing commands to computers. In the same way, it’s easy listening to audio played from a device than extract output from computers or devices. Especially with Robotics being an emerging market with applications in warehouses, the hospitality industry, consumer electronics, assistive technology, etc., speech-based human-machine interaction is emerging as a lucrative feature for robot manufacturers. Considering this factor, the objective of this paper is to design the “Audio-Visual Co-Data Processing Pipeline.” This pipeline is an integrated version of Automatic speech recognition, a Natural language model for text understanding, object detection, and text-to-speech modules. There are many Deep Learning models for each type of the modules mentioned above, but OpenVINO Model Zoo models are used because the OpenVINO toolkit covers both computer vision and non-computer vision workloads across Intel hardware and maximizes performance, and accelerates application development. A speech command is given as input that has information about target objects to be detected and start and end times to extract the required interval from the video. Speech is converted to text using the Automatic speech recognition QuartzNet model. The summary is extracted from text using a natural language model Generative Pre-Trained Transformer-3 (GPT-3). Based on the summary, essential frames from the video are extracted, and the You Only Look Once (YOLO) object detection model detects You Only Look Once (YOLO) objects on these extracted frames. Frame numbers that have target objects (specified objects in the speech command) are saved as text. Finally, this text (frame numbers) is converted to speech using text to speech model and will be played from the device. This project is developed for 80 You Only Look Once (YOLO) labels, and the user can extract frames based on only one or two target labels. This pipeline can be extended for more than two target labels easily by making appropriate changes in the object detection module. This project is developed for four different speech command formats by including sample examples in the prompt used by Generative Pre-Trained Transformer-3 (GPT-3) model. Based on user preference, one can come up with a new speech command format by including some examples of the respective format in the prompt used by the Generative Pre-Trained Transformer-3 (GPT-3) model. This pipeline can be used in many projects like human-machine interface, human-robot interaction, and surveillance through speech commands. All object detection projects can be upgraded using this pipeline so that one can give speech commands and output is played from the device.

Keywords: OpenVINO, automatic speech recognition, natural language processing, object detection, text to speech

Procedia PDF Downloads 49

Authors: Zhen Cheng, Xinyu Dai, Shujian Huang, Jiajun Chen

Abstract:

Recently, explanatory natural language inference has attracted much attention for the interpretability of logic relationship prediction, which is also known as explanation generation for Natural Language Inference (NLI). Existing explanation generators based on discriminative Encoder-Decoder architecture have achieved noticeable results. However, we find that these discriminative generators usually generate explanations with correct evidence but incorrect logic semantic. It is due to that logic information is implicitly encoded in the premise-hypothesis pairs and difficult to model. Actually, logic information identically exists between premise-hypothesis pair and explanation. And it is easy to extract logic information that is explicitly contained in the target explanation. Hence we assume that there exists a latent space of logic information while generating explanations. Specifically, we propose a generative model called Variational Explanation Generator (VariationalEG) with a latent variable to model this space. Training with the guide of explicit logic information in target explanations, latent variable in VariationalEG could capture the implicit logic information in premise-hypothesis pairs effectively. Additionally, to tackle the problem of posterior collapse while training VariaztionalEG, we propose a simple yet effective approach called Logic Supervision on the latent variable to force it to encode logic information. Experiments on explanation generation benchmark—explanation-Stanford Natural Language Inference (e-SNLI) demonstrate that the proposed VariationalEG achieves significant improvement compared to previous studies and yields a state-of-the-art result. Furthermore, we perform the analysis of generated explanations to demonstrate the effect of the latent variable.

Keywords: natural language inference, explanation generation, variational auto-encoder, generative model

Procedia PDF Downloads 122

Authors: Malik Muhammad Arslan, Muneeb Ullah, Dai Shihan, Daniyal Haider, Xiaodong Yang

Abstract:

Chest X-rays are instrumental in the detection and monitoring of a wide array of diseases, including viral infections such as COVID-19, tuberculosis, pneumonia, lung cancer, and various cardiac and pulmonary conditions. To enhance the accuracy of diagnosis, artificial intelligence (AI) algorithms, particularly deep learning models like Convolutional Neural Networks (CNNs), are employed. However, these deep learning models demand a substantial and varied dataset to attain optimal precision. Generative Adversarial Networks (GANs) can be employed to create new data, thereby supplementing the existing dataset and enhancing the accuracy of deep learning models. Nevertheless, GANs have their limitations, such as issues related to stability, convergence, and the ability to distinguish between authentic and fabricated data. In order to overcome these challenges and advance the detection and classification of CXR normal and abnormal images, this study introduces a distinctive technique known as DCWGAN (Diverse Conditional Wasserstein GAN) for generating synthetic chest X-ray (CXR) images. The study evaluates the effectiveness of this Idiosyncratic DCWGAN technique using the ResNet50 model and compares its results with those obtained using the traditional GAN approach. The findings reveal that the ResNet50 model trained on the DCWGAN-generated dataset outperformed the model trained on the classic GAN-generated dataset. Specifically, the ResNet50 model utilizing DCWGAN synthetic images achieved impressive performance metrics with an accuracy of 0.961, precision of 0.955, recall of 0.970, and F1-Measure of 0.963. These results indicate the promising potential for the early detection of diseases in CXR images using this Inimitable approach.

Keywords: CNN, classification, deep learning, GAN, Resnet50

Procedia PDF Downloads 47

Authors: P. Karthick, K. Mahesh

Abstract:

Video has become an increasingly significant component of our digital everyday contact. With the advancement of greater contents and shows of the resolution, its significant volume poses serious obstacles to the objective of receiving, distributing, compressing, and revealing video content of high quality. In this paper, we propose the primary beginning to complete a deep video compression model that jointly upgrades all video compression components. The video compression method involves splitting the video into frames, comparing the images using convolutional neural networks (CNN) to remove duplicates, repeating the single image instead of the duplicate images by recognizing and detecting minute changes using generative adversarial network (GAN) and recorded with long short-term memory (LSTM). Instead of the complete image, the small changes generated using GAN are substituted, which helps in frame level compression. Pixel wise comparison is performed using K-nearest neighbours (KNN) over the frame, clustered with K-means, and singular value decomposition (SVD) is applied for each and every frame in the video for all three color channels [Red, Green, Blue] to decrease the dimension of the utility matrix [R, G, B] by extracting its latent factors. Video frames are packed with parameters with the aid of a codec and converted to video format, and the results are compared with the original video. Repeated experiments on several videos with different sizes, duration, frames per second (FPS), and quality results demonstrate a significant resampling rate. On average, the result produced had approximately a 10% deviation in quality and more than 50% in size when compared with the original video.

Keywords: video compression, K-means clustering, convolutional neural network, generative adversarial network, singular value decomposition, pixel visualization, stochastic gradient descent, frame per second extraction, RGB channel extraction, self-detection and deciding system

Procedia PDF Downloads 161

Authors: Andrew Fox, Tao Zhang, Yuanhong Zhao, Qingping Yang

Abstract:

The furniture manufacturing industry has been slow in general to adopt the latest manufacturing technologies, historically relying heavily upon specialised conventional machinery. This approach not only requires high levels of specialist process knowledge, training, and capital investment but also suffers from significant subtractive manufacturing waste and high logistics costs due to the requirement for centralised manufacturing, with high levels of furniture product not re-cycled or re-used. This paper aims to address the problems by introducing suitable digital manufacturing technologies to create step changes in furniture manufacturing design, as the traditional design practices have been reported as building in 80% of environmental impact. In this paper, a 3D printing robot for furniture manufacturing is reported. The 3D printing robot mainly comprises a KUKA industrial robot, an Arduino microprocessor, and a self-assembled screw fed extruder. Compared to traditional 3D printer, the 3D printing robot has larger motion range and can be easily upgraded to enlarge the maximum size of the printed object. Generative design is also investigated in this paper, aiming to establish a combined design methodology that allows assessment of goals, constraints, materials, and manufacturing processes simultaneously. ‘Matrixing’ for part amalgamation and product performance optimisation is enabled. The generative design goals of integrated waste reduction increased manufacturing efficiency, optimised product performance, and reduced environmental impact institute a truly lean and innovative future design methodology. In addition, there is massive future potential to leverage Single Minute Exchange of Die (SMED) theory through generative design post-processing of geometry for robot manufacture, resulting in ‘mass customised’ furniture with virtually no setup requirements. These generatively designed products can be manufactured using the robot based additive manufacturing. Essentially, the 3D printing robot is already functional; some initial goals have been achieved and are also presented in this paper.

Keywords: additive manufacturing, generative design, robot, sustainability

Procedia PDF Downloads 94

Authors: Rony Reátegui, Cesar Chácara, Benjamin Castañeda, Rafael Aguilar

Abstract:

Nowadays, heritage building information modeling (HBIM) is considered an efficient tool to represent and manage information of cultural heritage (CH). The basis of this tool relies on a 3D model generally obtained from a cloud-to-BIM procedure. There are different methods to create an HBIM model that goes from manual modeling based on the point cloud to the automatic detection of shapes and the creation of objects. The selection of these methods depends on the desired level of development (LOD), level of information (LOI), grade of generation (GOG), as well as on the availability of commercial software. This paper presents the 3D modeling of a stone masonry chapel using Recap Pro, Revit, and Dynamo interface following a three-step methodology. The first step consists of the manual modeling of simple structural (e.g., regular walls, columns, floors, wall openings, etc.) and architectural (e.g., cornices, moldings, and other minor details) elements using the point cloud as reference. Then, Dynamo is used for generative modeling of complex structural elements such as vaults, infills, and domes. Finally, semantic information (e.g., materials, typology, state of conservation, etc.) and pathologies are added within the HBIM model as text parameters and generic models families, respectively. The application of this methodology allows the documentation of CH following a relatively simple to apply process that ensures adequate LOD, LOI, and GOG levels. In addition, the easy implementation of the method as well as the fact of using only one BIM software with its respective plugin for the scan-to-BIM modeling process means that this methodology can be adopted by a larger number of users with intermediate knowledge and limited resources since the BIM software used has a free student license.

Keywords: cloud-to-BIM, cultural heritage, generative modeling, HBIM, parametric modeling, Revit

Procedia PDF Downloads 118

Authors: John M. Lyne, K. Andrea Scott

Abstract:

In fluid dynamics, direct numerical simulations (DNS) of turbulent flows require large amounts of nodes to appropriately resolve all scales of energy transfer. Due to the size of these databases, sharing these datasets amongst the academic community is a challenge. Recent work has been done to investigate the use of super-resolution to enable database sharing, where a low-resolution flow field is super-resolved to high resolutions using a neural network. Recently, Generative Adversarial Networks (GAN) have grown in popularity with impressive results in the generation of faces, landscapes, and more. This work investigates the generation of unique high-resolution channel flow velocity fields from a low-dimensional latent space using a GAN. The training objective of the GAN is to generate samples in which the distribution of the generated samplesis ideally indistinguishable from the distribution of the training data. In this study, the network is trained using samples drawn from a statistically stationary channel flow at a Reynolds number of 560. Results show that the turbulent statistics and energy spectra of the generated flow fields are within reasonable agreement with those of the DNS data, demonstrating that GANscan produce the intricate multi-scale phenomena of turbulence.

Keywords: computational fluid dynamics, channel flow, turbulence, generative adversarial network

Procedia PDF Downloads 171

Authors: Getaneh Berie Tarekegn

Abstract:

A precise localization system is crucial for many artificial intelligence Internet of Things (AI-IoT) applications in the era of smart cities. Their applications include traffic monitoring, emergency alarming, environmental monitoring, location-based advertising, intelligent transportation, and smart health care. The most common method for providing continuous positioning services in outdoor environments is by using a global navigation satellite system (GNSS). Due to nonline-of-sight, multipath, and weather conditions, GNSS systems do not perform well in dense urban, urban, and suburban areas.This paper proposes a generative AI-based positioning scheme for large-scale wireless settings using fingerprinting techniques. In this article, we presented a semi-supervised deep convolutional generative adversarial network (S-DCGAN)-based radio map construction method for real-time device localization. It also employed a reliable signal fingerprint feature extraction method with t-distributed stochastic neighbor embedding (t-SNE), which extracts dominant features while eliminating noise from hybrid WLAN and long-term evolution (LTE) fingerprints. The proposed scheme reduced the workload of site surveying required to build the fingerprint database by up to 78.5% and significantly improved positioning accuracy. The results show that the average positioning error of GAILoc is less than 0.39 m, and more than 90% of the errors are less than 0.82 m. According to numerical results, SRCLoc improves positioning performance and reduces radio map construction costs significantly compared to traditional methods.

Keywords: location-aware services, feature extraction technique, generative adversarial network, long short-term memory, support vector machine

Procedia PDF Downloads 18

Authors: Guanghua Zhang, Fubao Wang, Weijun Duan

Abstract:

Convolution neural network (CNN) has attracted more and more attention on recent years. Especially in the field of computer vision and image classification. However, unsupervised learning with CNN has received less attention than supervised learning. In this work, we use a new powerful tool which is deep convolutional generative adversarial networks (DCGANs) to generate images from Sloan Digital Sky Survey. Training by various star and galaxy images, it shows that both the generator and the discriminator are good for unsupervised learning. In this paper, we also took several experiments to choose the best value for hyper-parameters and which could help to stabilize the training process and promise a good quality of the output.

Keywords: convolution neural network, discriminator, generator, unsupervised learning

Procedia PDF Downloads 236

Authors: Samer Muthana Sarsam, Abdul Samad Shibghatullah, Chit Su Mon, Abd Aziz Alias, Hosam Al-Samarraie

Abstract:

Generative Artificial Intelligence chatbots (GAI chatbots) have emerged as promising tools in various domains, including higher education. However, their specific role within the educational context and the level of legal support for their implementation remain unclear. Therefore, this study aims to investigate the role of Bard, a newly developed GAI chatbot, in higher education. To achieve this objective, English tweets were collected from Twitter's free streaming Application Programming Interface (API). The Latent Dirichlet Allocation (LDA) algorithm was applied to extract latent topics from the collected tweets. User sentiments, including disgust, surprise, sadness, anger, fear, joy, anticipation, and trust, as well as positive and negative sentiments, were extracted using the NRC Affect Intensity Lexicon and SentiStrength tools. This study explored the benefits, challenges, and future implications of integrating GAI chatbots in higher education. The findings shed light on the potential power of such tools, exemplified by Bard, in enhancing the learning process and providing support to students throughout their educational journey.

Keywords: generative artificial intelligence chatbots, bard, higher education, topic modelling, sentiment analysis

Procedia PDF Downloads 48

Authors: Getaneh Berie Tarekegn

Abstract:

A precise localization system is crucial for many artificial intelligence Internet of Things (AI-IoT) applications in the era of smart cities. Their applications include traffic monitoring, emergency alarming, environmental monitoring, location-based advertising, intelligent transportation, and smart health care. The most common method for providing continuous positioning services in outdoor environments is by using a global navigation satellite system (GNSS). Due to nonline-of-sight, multipath, and weather conditions, GNSS systems do not perform well in dense urban, urban, and suburban areas.This paper proposes a generative AI-based positioning scheme for large-scale wireless settings using fingerprinting techniques. In this article, we presented a novel semi-supervised deep convolutional generative adversarial network (S-DCGAN)-based radio map construction method for real-time device localization. We also employed a reliable signal fingerprint feature extraction method with t-distributed stochastic neighbor embedding (t-SNE), which extracts dominant features while eliminating noise from hybrid WLAN and long-term evolution (LTE) fingerprints. The proposed scheme reduced the workload of site surveying required to build the fingerprint database by up to 78.5% and significantly improved positioning accuracy. The results show that the average positioning error of GAILoc is less than 39 cm, and more than 90% of the errors are less than 82 cm. That is, numerical results proved that, in comparison to traditional methods, the proposed SRCLoc method can significantly improve positioning performance and reduce radio map construction costs.

Keywords: location-aware services, feature extraction technique, generative adversarial network, long short-term memory, support vector machine

Procedia PDF Downloads 41

Authors: Rongbo Shen, Jianhua Yao, Kezhou Yan, Kuan Tian, Cheng Jiang, Ke Zhou

Abstract:

This study proposes a generative adversarial networks (GAN) framework to perform synthetic sampling in feature space, i.e., feature augmentation, to address the class imbalance problem in medical image analysis. A feature extraction network is first trained to convert images into feature space. Then the GAN framework incorporates adversarial learning to train a feature generator for the minority class through playing a minimax game with a discriminator. The feature generator then generates features for minority class from arbitrary latent distributions to balance the data between the majority class and the minority class. Additionally, a data cleaning technique, i.e., Tomek link, is employed to clean up undesirable conflicting features introduced from the feature augmentation and thus establish well-defined class clusters for the training. The experiment section evaluates the proposed method on two medical image analysis tasks, i.e., mass classification on mammogram and cancer metastasis classification on histopathological images. Experimental results suggest that the proposed method obtains superior or comparable performance over the state-of-the-art counterparts. Compared to all counterparts, our proposed method improves more than 1.5 percentage of accuracy.

Keywords: class imbalance, synthetic sampling, feature augmentation, generative adversarial networks, data cleaning

Procedia PDF Downloads 103

Authors: Getaneh Berie Tarekegn

Abstract:

A precise localization system is crucial for many artificial intelligence Internet of Things (AI-IoT) applications in the era of smart cities. Their applications include traffic monitoring, emergency alarming, environmental monitoring, location-based advertising, intelligent transportation, and smart health care. The most common method for providing continuous positioning services in outdoor environments is by using a global navigation satellite system (GNSS). Due to nonline-of-sight, multipath, and weather conditions, GNSS systems do not perform well in dense urban, urban, and suburban areas.This paper proposes a generative AI-based positioning scheme for large-scale wireless settings using fingerprinting techniques. In this article, we presented a novel semi-supervised deep convolutional generative adversarial network (S-DCGAN)-based radio map construction method for real-time device localization. We also employed a reliable signal fingerprint feature extraction method with t-distributed stochastic neighbor embedding (t-SNE), which extracts dominant features while eliminating noise from hybrid WLAN and long-term evolution (LTE) fingerprints. The proposed scheme reduced the workload of site surveying required to build the fingerprint database by up to 78.5% and significantly improved positioning accuracy. The results show that the average positioning error of GAILoc is less than 39 cm, and more than 90% of the errors are less than 82 cm. That is, numerical results proved that, in comparison to traditional methods, the proposed SRCLoc method can significantly improve positioning performance and reduce radio map construction costs.

Keywords: location-aware services, feature extraction technique, generative adversarial network, long short-term memory, support vector machine

Procedia PDF Downloads 38

Authors: Hirofumi Miyajima, Kazuya Kishida, Noritaka Shigei, Hiromi Miyajima

Abstract:

Most of self-tuning fuzzy systems, which are automatically constructed from learning data, are based on the steepest descent method (SDM). However, this approach often requires a large convergence time and gets stuck into a shallow local minimum. One of its solutions is to use fuzzy rule modules with a small number of inputs such as DIRMs (Double-Input Rule Modules) and SIRMs (Single-Input Rule Modules). In this paper, we consider a (generalized) DIRMs model composed of double and single-input rule modules. Further, in order to reduce the redundant modules for the (generalized) DIRMs model, pruning and generative learning algorithms for the model are suggested. In order to show the effectiveness of them, numerical simulations for function approximation, Box-Jenkins and obstacle avoidance problems are performed.

Keywords: Box-Jenkins's problem, double-input rule module, fuzzy inference model, obstacle avoidance, single-input rule module

Procedia PDF Downloads 327

Authors: Yifan Fan, Xudong Luo, Pingping Lin

Abstract:

An essential task in the field of artificial intelligence is to allow computers to interact with people through natural language. Therefore, researches such as virtual assistants and dialogue systems have received widespread attention from industry and academia. The response generation plays a crucial role in dialogue systems, so to push forward the research on this topic, this paper surveys various methods for response generation. We sort out these methods into three categories. First one includes finite state machine methods, framework methods, and instance methods. The second contains full-text indexing methods, ontology methods, vast knowledge base method, and some other methods. The third covers retrieval methods and generative methods. We also discuss some hybrid methods based knowledge and deep learning. We compare their disadvantages and advantages and point out in which ways these studies can be improved further. Our discussion covers some studies published in leading conferences such as IJCAI and AAAI in recent years.

Keywords: deep learning, generative, knowledge, response generation, retrieval

Procedia PDF Downloads 104

Authors: B. Mombayeva

Abstract:

In the deserts area of south east of Kazakhstan, 16 species of Coleoptera from 6 families and 12 genus of insects damaging Saxaul have been revealed. The vast number of species belong to the Cerambycidae familyCapricorn Beetle (4 species) and Hemlock Borer of Melanophila genus and 3 species of weevils and flea-beetles, and 1 species of coctsinelids and carrion beetle. Some of them cause appreciable harm, and sometimes very heavy damageto saxaul. According to food specialization they are divided into polyphages and - oligophages. According to the confinement to saxaul parts, registered beetles insects mainly feed on generative parts (11 species) and leaves (5 species). 9 species from them feed on roots, leaves and generative organs. They are scarablike beetle’s larvae (Apatophysismongolica Semenov., Tursmenigenavarentzovi Melg., Phytoecia (Opsilla) coerulescens Scopoli., Apatophysismongolica Semenov.), Jewel beetles (Julodis (s. Str.) Variolaris (Pallas), Sphenoptera (s. Str.) cuprina Motschulsky, S. (s. str.) exarata (Fischer), SphenopterapotaniniJak.) and some weevil (Barisartemisiae Hbst.). The larvae eat the roots and the imago - generative organs. Their feeding noticeably has its effect on the condition of saxaul. Beetles also slightlygnaw vegetative organs of plants. Among the harmful species the desert Capricorn Beetle Julodisvariolaris (Pallas) deserved attention. Its larvae live in the soil and cause harm to the roots of Saxaul and other pasture plants. In addition, the larvae of Sphenopterapotanini, S.punctatissima colonize the roots, trunk and branches of Haloxylon. In the spring Saxaul flowers are much damaged by Ladybeetle Bulaealichatchovi.

Keywords: saxaul, coleoptera, insecta, haloxylon

Procedia PDF Downloads 233

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 39

Authors: Vlad Andrei Raducanu, Mariana Lucia Angelescu, Ion Cinca, Vasile Danut Cojocaru, Doina Raducanu

Abstract:

By using smart digital tools, such as generative design (GD) and digital fabrication (DF), problems of high actuality concerning resources optimization (materials, energy, time) can be solved and applications or products of free-form type can be created. In the new digital technology materials are active, designed in response to a set of performance requirements, which impose a total rethinking of old material practices. The article presents the design procedure key steps of a free-form architectural object - a column type one with connections to get an adaptive 3D surface, by using the parametric design methodology and by exploiting the properties of conventional metallic materials. In parametric design the form of the created object or space is shaped by varying the parameters values and relationships between the forms are described by mathematical equations. Digital parametric design is based on specific procedures, as shape grammars, Lindenmayer - systems, cellular automata, genetic algorithms or swarm intelligence, each of these procedures having limitations which make them applicable only in certain cases. In the paper the design process stages and the shape grammar type algorithm are presented. The generative design process relies on two basic principles: the modeling principle and the generative principle. The generative method is based on a form finding process, by creating many 3D spatial forms, using an algorithm conceived in order to apply its generating logic onto different input geometry. Once the algorithm is realized, it can be applied repeatedly to generate the geometry for a number of different input surfaces. The generated configurations are then analyzed through a technical or aesthetic selection criterion and finally the optimal solution is selected. Endless range of generative capacity of codes and algorithms used in digital design offers various conceptual possibilities and optimal solutions for both technical and environmental increasing demands of building industry and architecture. Constructions or spaces generated by parametric design can be specifically tuned, in order to meet certain technical or aesthetical requirements. The proposed approach has direct applicability in sustainable architecture, offering important potential economic advantages, a flexible design (which can be changed until the end of the design process) and unique geometric models of high performance.

Keywords: parametric design, algorithmic procedures, free-form architectural object, sustainable architecture

Procedia PDF Downloads 344

Authors: Wesley Teskey, Vedran Glavas, Julian Wegener

Abstract:

Generating battery cathode microstructures is an important area of research, given the proliferation of the use of automotive batteries. Currently, finite element analysis (FEA) is often used for simulations of battery cathode microstructures before physical batteries can be manufactured and tested to verify the simulation results. Unfortunately, a key drawback of using FEA is that this method of simulation is very slow in terms of computational runtime. Generative AI offers the key advantage of speed when compared to FEA, and because of this, generative AI is capable of evaluating very large numbers of candidate microstructures. Given AI generated candidate microstructures, a subset of the promising microstructures can be selected for further validation using FEA. Leveraging the speed advantage of AI allows for a better final microstructural selection because high speed allows for the evaluation of many more candidate microstructures. For the approach presented, battery cathode 3D candidate microstructures are generated using Gaussian Mixture Models (GMMs) and pix2pix. This approach first uses GMMs to generate a population of spheres (representing the “active material” of the cathode). Once spheres have been sampled from the GMM, they are placed within a microstructure. Subsequently, the pix2pix sweeps over the 3D microstructure (iteratively) slice by slice and adds details to the microstructure to determine what portions of the microstructure will become electrolyte and what part of the microstructure will become binder. In this manner, each subsequent slice of the microstructure is evaluated using pix2pix, where the inputs into pix2pix are the previously processed layers of the microstructure. By feeding into pix2pix previously fully processed layers of the microstructure, pix2pix can be used to ensure candidate microstructures represent a realistic physical reality. More specifically, in order for the microstructure to represent a realistic physical reality, the locations of electrolyte and binder in each layer of the microstructure must reasonably match the locations of electrolyte and binder in previous layers to ensure geometric continuity. Using the above outlined approach, a 10x to 100x speed increase was possible when generating candidate microstructures using AI when compared to using a FEA only approach for this task. A key metric for evaluating microstructures was the battery specific power value that the microstructures would be able to produce. The best generative AI result obtained was a 12% increase in specific power for a candidate microstructure when compared to what a FEA only approach was capable of producing. This 12% increase in specific power was verified by FEA simulation.

Keywords: finite element analysis, gaussian mixture models, generative design, Pix2Pix, structural design

Procedia PDF Downloads 79

Authors: Po-Chun Wang, Yan-Ru Lai, Sophia I. C. Lin, Alvin W. Y. Su

Abstract:

The Generative Theory of Tonal Music (GTTM) provides systematic approaches to recognizing local boundaries of music. The rules have been implemented in some automated melody segmentation algorithms. Besides, there are also deep learning methods with GTTM features applied to boundary detection tasks. However, these studies might face constraints such as a lack of or inconsistent label data. The GTTM database is currently the most widely used GTTM database, which includes manually labeled GTTM rules and local boundaries. Even so, we found some problems with these labels. They are sometimes discrepancies with GTTM rules. In addition, since it is labeled at different times by multiple musicians, they are not within the same scope in some cases. Therefore, in this paper, we examine this database with musicians from the aspect of classical music and relabel the scores. The relabeled database - GTTM Database v2.0 - will be released for academic research usage. Despite the experimental and statistical results showing that the relabeled database is more consistent, the improvement in boundary detection is not substantial. It seems that we need more clues than GTTM rules for boundary detection in the future.

Keywords: dataset, GTTM, local boundary, neural network

Procedia PDF Downloads 104

Authors: Saqib Aziz, Brad Alexander, Christoph Gengnagel, Stefan Weinzierl

Abstract:

This paper explores a generative design, simulation, and optimization workflow for the integration of acoustical diffuser and/or absorber geometry with embedded coupled Helmholtz-resonators for full-scale 3D printed building components. Large-scale additive manufacturing in conjunction with algorithmic CAD design tools enables a vast amount of control when creating geometry. This is advantageous regarding the increasing demands of comfort standards for indoor spaces and the use of more resourceful and sustainable construction methods and materials. The presented methodology highlights these new technological advancements and offers a multimodal and integrative design solution with the potential for an immediate application in the AEC-Industry. In principle, the methodology can be applied to a wide range of structural elements that can be manufactured by additive manufacturing processes. The current paper focuses on a case study of an application for a biaxial load-bearing beam grillage made of reinforced concrete, which allows for a variety of applications through the combination of additive prefabricated semi-finished parts and in-situ concrete supplementation. The semi-prefabricated parts or formwork bodies form the basic framework of the supporting structure and at the same time have acoustic absorption and diffusion properties that are precisely acoustically programmed for the space underneath the structure. To this end, a hybrid validation strategy is being explored using a digital and cross-platform simulation environment, verified with physical prototyping. The iterative workflow starts with the generation of a parametric design model for the acoustical geometry using the algorithmic visual scripting editor Grasshopper3D inside the building information modeling (BIM) software Revit. Various geometric attributes (i.e., bottleneck and cavity dimensions) of the resonator are parameterized and fed to a numerical optimization algorithm which can modify the geometry with the goal of increasing absorption at resonance and increasing the bandwidth of the effective absorption range. Using Rhino.Inside and LiveLink for Revit, the generative model was imported directly into the Multiphysics simulation environment COMSOL. The geometry was further modified and prepared for simulation in a semi-automated process. The incident and scattered pressure fields were simulated from which the surface normal absorption coefficients were calculated. This reciprocal process was repeated to further optimize the geometric parameters. Subsequently the numerical models were compared to a set of 3D concrete printed physical twin models, which were tested in a .25 m x .25 m impedance tube. The empirical results served to improve the starting parameter settings of the initial numerical model. The geometry resulting from the numerical optimization was finally returned to grasshopper for further implementation in an interdisciplinary study.

Keywords: acoustical design, additive manufacturing, computational design, multimodal optimization

Procedia PDF Downloads 134

Authors: Roberto Cabezas H

Abstract:

The current work shows the methodology developed to create narrative lighting spaces for the multimedia performance piece 'cluster: the vanished paradise.' This empirical research is focused on exploring unconventional roles for machines in subjective creative processes, by delving into the semantics of data and machine intelligence algorithms in hybrid technological, creative contexts to expand epistemic domains trough human-machine cooperation. The creative process in scenic and performing arts is guided mostly by intuition; from that idea, we developed an approach to embed collective intuition in computational creative systems, by joining the properties of Generative Adversarial Networks (GAN’s) and Fuzzy Clustering based on a semi-supervised data creation and analysis pipeline. The model makes use of GAN’s to learn from phenomenological data (data generated from experience with lighting scenography) and algorithmic design data (augmented data by procedural design methods), fuzzy logic clustering is then applied to artificially created data from GAN’s to define narrative transitions built on membership index; this process allowed for the creation of simple and complex spaces with expressive capabilities based on position and light intensity as the parameters to guide the narrative. Hybridization comes not only from the human-machine symbiosis but also on the integration of different techniques for the implementation of the aided design system. Machine intelligence tools as proposed in this work are well suited to redefine collaborative creation by learning to express and expand a conglomerate of ideas and a wide range of opinions for the creation of sensory experiences. We found in GAN’s and Fuzzy Logic an ideal tool to develop new computational models based on interaction, learning, emotion and imagination to expand the traditional algorithmic model of computation.

Keywords: fuzzy clustering, generative adversarial networks, human-machine cooperation, hybrid collective data, multimedia performance

Procedia PDF Downloads 115

Authors: Arya Prakash Padhi, Anupam Chakrabarti, Rajib Chowdhury

Abstract:

Topology or layout optimization often gives better performing economic structures and is very helpful in the conceptual design phase. But traditionally it is being done in finite element-based optimization schemes which, although gives a good result, is very time-consuming especially in 3D structures. Among other alternatives machine learning, especially deep learning-based methods, have a very good potential in resolving this computational issue. Here convolutional neural network (3D-CNN) based variational auto encoder (VAE) is trained using a dataset generated from commercially available topology optimization code ABAQUS Tosca using solid isotropic material with penalization (SIMP) method for compliance minimization. The encoded data in latent space is then fed to a 3D generative adversarial network (3D-GAN) to generate the outcome in 64x64x64 size. Here the network consists of 3D volumetric CNN with rectified linear unit (ReLU) activation in between and sigmoid activation in the end. The proposed network is seen to provide almost optimal results with significantly reduced computational time, as there is no iteration involved.

Keywords: 3D generative adversarial network, deep learning, structural topology optimization, variational auto encoder

Procedia PDF Downloads 137

Authors: Hamed Alqahtani, Manolya Kavakli-Thorne

Abstract:

The large pose discrepancy is one of the critical challenges in face recognition during video surveillance. Due to the entanglement of pose attributes with identity information, the conventional approaches for pose-independent representation lack in providing quality results in recognizing largely posed faces. In this paper, we propose a practical approach to disentangle the pose attribute from the identity information followed by synthesis of a face using a classifier network in latent space. The proposed approach employs a modified generative adversarial network framework consisting of an encoder-decoder structure embedded with a classifier in manifold space for carrying out factorization on the latent encoding. It can be further generalized to other face and non-face attributes for real-life video frames containing faces with significant attribute variations. Experimental results and comparison with state of the art in the field prove that the learned representation of the proposed approach synthesizes more compelling perceptual images through a combination of adversarial and classification losses.

Keywords: disentanglement, face detection, generative adversarial networks, video surveillance

Procedia PDF Downloads 92

Authors: Yoganathan Karthik, Guhanathan Poravi

Abstract:

Colorizing near-infrared (NIR) images poses unique challenges due to the absence of color information and the nuances in light absorption. In this paper, we present an approach to NIR image colorization utilizing a synthetic dataset generated from visible light images. Our method addresses two major challenges encountered in NIR image colorization: accurately colorizing objects with color variations and avoiding over/under saturation in dimly lit scenes. To tackle these challenges, we propose a Generative Adversarial Network (GAN)-based framework that learns to map NIR images to their corresponding colorized versions. The synthetic dataset ensures diverse color representations, enabling the model to effectively handle objects with varying hues and shades. Furthermore, the GAN architecture facilitates the generation of realistic colorizations while preserving the integrity of dimly lit scenes, thus mitigating issues related to over/under saturation. Experimental results on benchmark NIR image datasets demonstrate the efficacy of our approach in producing high-quality colorizations with improved color accuracy and naturalness. Quantitative evaluations and comparative studies validate the superiority of our method over existing techniques, showcasing its robustness and generalization capability across diverse NIR image scenarios. Our research not only contributes to advancing NIR image colorization but also underscores the importance of synthetic datasets and GANs in addressing domain-specific challenges in image processing tasks. The proposed framework holds promise for various applications in remote sensing, medical imaging, and surveillance where accurate color representation of NIR imagery is crucial for analysis and interpretation.

Keywords: computer vision, near-infrared images, automatic image colorization, generative adversarial networks, synthetic data

Procedia PDF Downloads 4