Search results for: generative AI

Authors: Shubham Gogri, Lucia Florescu

Abstract:

Medical diagnosis and radiotherapy treatment planning using Computed Tomography (CT) rely on the quantitative accuracy and quality of the CT images. At the same time, requirements for CT imaging include reducing the radiation dose exposure to patients and minimizing scanning time. A solution to this is the sparse-view CT technique, based on a reduced number of projection views. This, however, introduces a new problem— the incomplete projection data results in lower quality of the reconstructed images. To tackle this issue, deep learning methods have been applied to enhance the quality of the sparse-view CT images. A first approach involved employing Mir-Net, a dedicated deep neural network designed for image enhancement. This showed promise, utilizing an intricate architecture comprising encoder and decoder networks, along with the incorporation of the Charbonnier Loss. However, this approach was computationally demanding. Subsequently, a specialized Generative Adversarial Network (GAN) architecture, rooted in the Pix2Pix framework, was implemented. This GAN framework involves a U-Net-based Generator and a Discriminator based on Convolutional Neural Networks. To bolster the GAN's performance, both Charbonnier and Wasserstein loss functions were introduced, collectively focusing on capturing minute details while ensuring training stability. The integration of the perceptual loss, calculated based on feature vectors extracted from the VGG16 network pretrained on the ImageNet dataset, further enhanced the network's ability to synthesize relevant images. A series of comprehensive experiments with clinical CT data were conducted, exploring various GAN loss functions, including Wasserstein, Charbonnier, and perceptual loss. The outcomes demonstrated significant image quality improvements, confirmed through pertinent metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) between the corrected images and the ground truth. Furthermore, learning curves and qualitative comparisons added evidence of the enhanced image quality and the network's increased stability, while preserving pixel value intensity. The experiments underscored the potential of deep learning frameworks in enhancing the visual interpretation of CT scans, achieving outcomes with SSIM values close to one and PSNR values reaching up to 76.

Keywords: generative adversarial networks, sparse view computed tomography, CT image correction, Mir-Net

Procedia PDF Downloads 159

Authors: Yifan Fan, Xudong Luo, Pingping Lin

Abstract:

Nowadays, dialogue systems increasingly become the way for humans to access many computer systems. So, humans can interact with computers in natural language. A dialogue system consists of three parts: understanding what humans say in natural language, managing dialogue, and generating responses in natural language. In this paper, we survey deep learning based methods for dialogue management, response generation and dialogue evaluation. Specifically, these methods are based on neural network, long short-term memory network, deep reinforcement learning, pre-training and generative adversarial network. We compare these methods and point out the further research directions.

Keywords: dialogue management, response generation, deep learning, evaluation

Procedia PDF Downloads 166

Authors: Xiaochen Mu

Abstract:

Whether in common law countries that adhere to the "commercial copyright theory" or in civil law countries that center around "author's rights," the standards for judging originality have undergone continuous adjustments in response to the development of information technology. The adherence to originality standards does not arbitrarily dictate that all types of works be judged according to a single standard of originality, nor does it rigidly ignore the changes in creative methods and dissemination models brought about by technology. Adjustments and interpretations should be allowed based on the different forms of expression of works. Appropriate adjustments and interpretations are our response to technological advancements. However, what should be upheld are the principles and bottom lines of these adjustments and interpretations, namely the legislative intent and purpose of copyright law, which are to encourage the creation and dissemination of outstanding cultural works and to promote the flourishing of culture.

Keywords: generative artificial intelligence, originality, works, copyright

Procedia PDF Downloads 41

Authors: Akash Amalan, Rui Wang, Yanqi Qiao, Emmanouil Panaousis, Kaitai Liang

Abstract:

Federated learning is an emerging concept in the domain of distributed machine learning. This concept has enabled General Adversarial Networks (GANs) to benefit from the rich distributed training data while preserving privacy. However, in a non-IID setting, current federated GAN architectures are unstable, struggling to learn the distinct features, and vulnerable to mode collapse. In this paper, we propose an architecture MULTI-FLGAN to solve the problem of low-quality images, mode collapse, and instability for non-IID datasets. Our results show that MULTI-FLGAN is four times as stable and performant (i.e., high inception score) on average over 20 clients compared to baseline FLGAN.

Keywords: federated learning, generative adversarial network, inference attack, non-IID data distribution

Procedia PDF Downloads 158

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 196

Authors: Alba Francesca Canta

Abstract:

The present is one of those times when what normally seems to constitute a reason for living vanishes, particularly in times of crisis, during which certainties of all times crumble, and critical issues emerge, especially in already problematic areas such as the role of women and children. This paper aims to explore the issue of gender and highlight the importance of education for people’s development and well-being. The study is part of the broader framework of the capability approach, a multidimensional approach based on the need to consider a person’s wealth by virtue of their opportunity and freedom to live a ‘life of worth. The results of empirical research conducted in 2020 will be presented, the main objective of which was to measure, through qualitative (project techniques, focus groups, interviews with key informants) and quantitative (questionnaire) methods, the level of empowerment of women in two Italian territories and the consequent well-being of their children. By means of the relationship study, the present research results show that a higher level of women’s empowerment corresponds to a higher level of children’s well-being in a positive virtuous process. The opportunity structure and education are the main driving guide both to women’s empowerment and children’s well-being, emphasizing the importance of education to gender culture as a key factor for the development of the whole society. Among all the traumatic events that broke the harmony of the world and caused an abrupt turn in all areas of society, the crisis of democracy and education are some of the harshest. Nevertheless, education continues to be a fundamental pillar of Global Development Agendas, and above all, democratic education is the main factor in the development of a generative society, capable of forming people who know how to live in society. In this context, recovering democratic and inclusive education can be the key to a breakthrough. In the capability approach Sen, and other Scholars, point out education from two different perspectives: a. education as a fundamental right capable of influencing other real fields of people’s life (i.e., being educated to prevent illness, to vote, etc.) and b. spread communitarian education, tolerance, inclusive, democratic, and respectful, capable of forming human beings. This kind of educational system can directly lead to a general process of gender education that presupposes respect for essential principles: equality, uniqueness, and the participation of all in the processes of defining a democratic society. Many practices of women and children’s exclusions essentially derive from social factors (norms, values, quality of institutions, relations of power, educational and cultural practices) that can build strong barriers. Respect for these principles and education for gender culture could foster the renewal of society and the acquisition of fundamental skills for a generative and inclusive society, such as critical skills, cosmopolitan skills, and narrative imagination.

Keywords: capability approach, children’s well-being, education, women’s empowerment

Procedia PDF Downloads 66

Authors: S. Jagadeesh Kumar

Abstract:

This expose presents a luculent of recent machine learning practices as employed in the modern and as pertinent to prospective automatic speech recognition schemes. The aspiration is to promote additional traverse ablution among the machine learning and automatic speech recognition factions that have transpired in the precedent. The manuscript is structured according to the chief machine learning archetypes that are furthermore trendy by now or have latency for building momentous hand-outs to automatic speech recognition expertise. The standards offered and convoluted in this article embraces adaptive and multi-task learning, active learning, Bayesian learning, discriminative learning, generative learning, supervised and unsupervised learning. These learning archetypes are aggravated and conferred in the perspective of automatic speech recognition tools and functions. This manuscript bequeaths and surveys topical advances of deep learning and learning with sparse depictions; further limelight is on their incessant significance in the evolution of automatic speech recognition.

Keywords: automatic speech recognition, deep learning methods, machine learning archetypes, Bayesian learning, supervised and unsupervised learning

Procedia PDF Downloads 446

Authors: St. Subaedah, A. Takdir, Netty, D. Hidrawati

Abstract:

Maize development at the rainfed lowland after rice is often confronted with the occurrence of drought stress at the time of entering the generative phase, which will cause be hampered crop production. Consequently, in the utilization of the rainfed lowland areas optimally, an effort that can be done using the varieties of early maturity to minimize crop failures due to its short rainy season. The aim of this research was evaluating the potential yield of genotypes of candidates of maize early maturity in the rainfed lowland areas. The study was conducted during May to August 2016 at South Sulawesi, Indonesia. The study used randomized block design to compare 12 treatments and consists of 8 genotypes namely CH1, CH2, CH3, CH4, CH5, CH6, CH7, CH8 and the use of four varieties, namely Bima 3, Bima 7, Lamuru and Gumarang. The results showed that genotype of CH2, CH3, CH5, CH 6, CH7 and CH8 harvesting has less than 90 days. There are two genotypes namely genotypes of CH7 and CH8 that have a fairly high production respectively of 7.16 tons / ha and 8.11 tons/ ha and significantly not different from the superior varieties Bima3.

Keywords: evaluation, early maturity, maize, yield potential

Procedia PDF Downloads 192

Authors: Rui Yin, Haojie Wang, Xun Li

Abstract:

Recently, mainstream dense retrieval methods have obtained state-of-the-art results on some datasets and tasks. However, they require large amounts of training data, which is not available in most domains. The severe performance degradation of dense retrievers on new data domains has limited the use of dense retrieval methods to only a few domains with large training datasets. In this paper, we propose an unsupervised domain-adaptive approach based on query generation. First, a generative model is used to generate relevant queries for each passage in the target corpus, and then the generated queries are used for mining negative passages. Finally, the query-passage pairs are labeled with a cross-encoder and used to train a domain-adapted dense retriever. Experiments show that our approach is more robust than previous methods in target domains that require less unlabeled data.

Keywords: dense retrieval, query generation, unsupervised training, text retrieval

Procedia PDF Downloads 71

Authors: R. R. Govind

Abstract:

In the digital age, artificial intelligence (AI) is having a significant impact on the rapid changes that cities are experiencing. This study explores the profound impact of AI on urban morphology, especially with regard to promoting friendly design choices. It addresses a significant research gap by examining the real-world effects of integrating AI into urban design and management. The main objective is to outline a framework for integrating AI to transform urban settings. The study employs an urban design framework to effectively navigate complicated urban environments, emphasize the need for urban management, and provide efficient planning and design strategies. Taking Gangtok's informal settlements as a focal point, the study employs AI methodologies such as machine learning, predictive analytics, and generative AI to tackle issues of 'urban informality'. The insights garnered not only offer valuable perspectives but also unveil AI's transformative potential in addressing contemporary urban challenges.

Keywords: urban design, artificial intelligence, urban challenges, machine learning, urban informality

Procedia PDF Downloads 60

Authors: Ukpaka Paschal

Abstract:

Margret Boden argues that a creative product is one that is new, surprising, and valuable as a result of the combination, exploration, or transformation involved in producing it. Boden uses examples of artificial intelligence systems that fit all of these criteria and argues that real creativity involves autonomy, intentionality, valuation, emotion, and consciousness. This paper provides an analysis of all these elements in order to try to understand whether they are sufficient to account for creativity, especially human creativity. This paper focuses on Generative Adversarial Networks (GANs), which is a class of artificial intelligence algorithms that are said to have disproved the common perception that creativity is something that only humans possess. This paper will then argue that Boden’s listed properties of creativity, which capture the creativity exhibited by GANs, are not sufficient to account for human creativity, and this paper will further identify “first-person phenomenological experience” as an essential property of human creativity. The rationale behind the proposed essential property is that if creativity involves comprehending our experience of the world around us into a form of self-expression, then our experience of the world really matters with regard to creativity.

Keywords: artificial intelligence, creativity, GANs, first-person experience

Procedia PDF Downloads 134

Authors: Debarag Banerjee, Pooja Singh, Arjun Avadhanam, Saksham Srivastava

Abstract:

Autonomous conversational agents, i.e., chatbots, are becoming an increasingly common mechanism for enterprises to provide support to customers and partners. In order to rate chatbots, especially ones powered by Generative AI tools like Large Language Models (LLMs), we need to be able to accurately assess their performance. This is where chatbot benchmarking becomes important. In this paper, authors propose the use of a benchmark that they call the E2E (End to End) benchmark and show how the E2E benchmark can be used to evaluate the accuracy and usefulness of the answers provided by chatbots, especially ones powered by LLMs. The authors evaluate an example chatbot at different levels of sophistication based on both our E2E benchmark as well as other available metrics commonly used in the state of the art and observe that the proposed benchmark shows better results compared to others. In addition, while some metrics proved to be unpredictable, the metric associated with the E2E benchmark, which uses cosine similarity, performed well in evaluating chatbots. The performance of our best models shows that there are several benefits of using the cosine similarity score as a metric in the E2E benchmark.

Keywords: chatbot benchmarking, end-to-end (E2E) benchmarking, large language model, user centric evaluation.

Procedia PDF Downloads 65

Authors: Amarendar Reddy Addula

Abstract:

Definition of Artificial Intelligence: Artificial intelligence is the simulation of mortal intelligence processes by machines, especially computer systems. Explicit operations of AI comprise expert systems, natural language processing, and speech recognition, and machine vision. Artificial Intelligence (AI) is an original medium for digital business, according to a new report by Gartner. The last 10 times represent an advance period in AI’s development, prodded by the confluence of factors, including the rise of big data, advancements in cipher structure, new machine literacy ways, the materialization of pall computing, and the vibrant open- source ecosystem. Influence of AI to a broader set of use cases and druggies and its gaining fashionability because it improves AI’s versatility, effectiveness, and rigidity. Edge AI will enable digital moments by employing AI for real- time analytics closer to data sources. Gartner predicts that by 2025, further than 50 of all data analysis by deep neural networks will do at the edge, over from lower than 10 in 2021. Responsible AI is a marquee term for making suitable business and ethical choices when espousing AI. It requires considering business and societal value, threat, trust, translucency, fairness, bias mitigation, explainability, responsibility, safety, sequestration, and nonsupervisory compliance. Responsible AI is ever more significant amidst growing nonsupervisory oversight, consumer prospects, and rising sustainability pretensions. Generative AI is the use of AI to induce new vestiges and produce innovative products. To date, generative AI sweats have concentrated on creating media content similar as photorealistic images of people and effects, but it can also be used for law generation, creating synthetic irregular data, and designing medicinals and accoutrements with specific parcels. AI is the subject of a wide- ranging debate in which there's a growing concern about its ethical and legal aspects. Constantly, the two are varied and nonplussed despite being different issues and areas of knowledge. The ethical debate raises two main problems the first, abstract, relates to the idea and content of ethics; the alternate, functional, and concerns its relationship with the law. Both set up models of social geste, but they're different in compass and nature. The juridical analysis is grounded on anon-formalistic scientific methodology. This means that it's essential to consider the nature and characteristics of the AI as a primary step to the description of its legal paradigm. In this regard, there are two main issues the relationship between artificial and mortal intelligence and the question of the unitary or different nature of the AI. From that theoretical and practical base, the study of the legal system is carried out by examining its foundations, the governance model, and the nonsupervisory bases. According to this analysis, throughout the work and in the conclusions, International Law is linked as the top legal frame for the regulation of AI.

Keywords: artificial intelligence, ethics & human rights issues, laws, international laws

Procedia PDF Downloads 93

Authors: Gangmin Li, Fan Yang

Abstract:

Personalized recommendation is crucial for any recommendation system. One of the techniques for personalized recommendation is to identify the intention. Traditional user intention identification uses the user’s selection when facing multiple items. This modeling relies primarily on historical behaviour data resulting in challenges such as the cold start, unintended choice, and failure to capture intention when items are new. Motivated by recent advancements in Large Language Models (LLMs) like ChatGPT, we present an approach for user intention identification by embracing LLMs with Chain-of-Thought (CoT) prompting. We use the initial user profile as input to LLMs and design a collection of prompts to align the LLM's response through various recommendation tasks encompassing rating prediction, search and browse history, user clarification, etc. Our tests on real-world datasets demonstrate the improvements in recommendation by explicit user intention identification and, with that intention, merged into a user model.

Keywords: personalized recommendation, generative user modelling, user intention identification, large language models, chain-of-thought prompting

Procedia PDF Downloads 52

Authors: Claudio Burguez, María Castelló, Mikaela Pisani, Marcos Umpiérrez

Abstract:

BIOFEEDBACK exhibition offers a unique experience for each visitor, combining art, neuroscience, and technology in an interactive way. Using a headband that captures the bioelectric activity of the brain, the visitors are able to generate sound and images in a sequence loop, making them an integral part of the artwork. Through this interactive exhibit, visitors gain a deeper appreciation of the beauty and complexity of the brain. As a special takeaway, visitors will receive an NFT as a present, allowing them to continue their engagement with the exhibition beyond the physical space. We used the EEG Biofeedback technique following a closed-loop neuroscience approach, transforming EEG data captured by a Muse S headband in real-time into audiovisual stimulation. PureData is used for sound generation and Generative Adversarial Networks (GANs) for image generation. Thirty participants have experienced the exhibition. For some individuals, it was easier to focus than others. Participants who said they could focus during the exhibit stated that at one point, they felt that they could control the sound, while images were more abstract, and they did not feel that they were able to control them.

Keywords: art, audiovisual, biofeedback, EEG, NFT, neuroscience, technology

Procedia PDF Downloads 71

Authors: Olufemi Olupe

Abstract:

Within linguistics, various approaches have been adopted to the study of language. One of such approaches is the syntax. The syntax is an aspect of the grammar of the language which deals with how words are put together to form phrases and sentences and how such structures are interpreted in language. Ambiguity, which is also germane in this discourse is about the uncertainty of meaning as a result of the possibility of a phrase or sentence being understood and interpreted in more than one way. In the light of the above, this paper attempts a syntactic study of syntactic ambiguities in The English Language, using the Transformational Generative Grammar (TGG) Approach. In doing this, phrases and sentences were raised with each description followed by relevant analysis. Finding in the work reveals that ambiguity cannot always be disambiguated by the means of syntactic analysis alone without recourse to semantic interpretation. The further finding shows that some syntactical ambiguities structures cannot be analysed on two surface structures in spite of the fact that there are more than one deep structures. The paper concludes that in as much as ambiguity remains in language; it will continue to pose a problem of understanding to a second language learner. Users of English as a second language, must, however, make a conscious effort to avoid its usage to achieve effective communication.

Keywords: language, syntax, semantics, morphology, ambiguity

Procedia PDF Downloads 392

Authors: Weisi Guo

Abstract:

It is well established that there may be running correlations in sequential exam marks due to students sitting in the order of course registration patterns. As such, a random and non-sequential sampling of exam marks is a standard recommended practice. Here, the paper examines a large number of exam data stretching several years across different modules to see the degree to which it is true. Using the real mark distribution as a generative process, it was found that random simulated data had no more sequential randomness than the real data. That is to say, the running correlations that one often observes are statistically identical to chance. Digging deeper, it was found that some high running correlations have students that indeed share a common course history and make similar mistakes. However, at the statistical scale of a module question, the combined effect is statistically similar to the random shuffling of papers. As such, there may not be the need to take random samples for marks, but it still remains good practice to mark papers in a random sequence to reduce the repetitive marking bias and errors.

Keywords: data analysis, empirical study, exams, marking

Procedia PDF Downloads 180

Authors: Hossam Abdelrahman, Sama Rostom, Reem Yassein, Yara Mohamed, Salma Salah, Nour Awny

Abstract:

In today’s environment, people are becoming increasingly interested in their appearance. However, they are afraid of their unknown appearance after a plastic surgery or treatment. Accidents, burns and genetic problems such as bowing of body parts of people have a negative impact on their mental health with their appearance and this makes them feel uncomfortable and underestimated. The approach presents a revolutionary deep learning-based image inpainting method that analyses the various picture structures and corrects damaged images. In this study, A model is proposed based on the in-painting of medical images with Stable Diffusion Inpainting method. Reconstructing missing and damaged sections of an image is known as image inpainting is a key progress facilitated by deep neural networks. The system uses the input of the user of an image to indicate a problem, the system will then modify the image and output the fixed image, facilitating for the patient to see the final result.

Keywords: generative adversarial network, large mask inpainting, stable diffusion inpainting, plastic surgery

Procedia PDF Downloads 73

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 352

Authors: Luminiţa Duţică, Gheorghe Duţică

Abstract:

One of the most powerful language innovation in the twentieth century music was the heterophony–hypostasis of the vertical syntax entered into the sphere of interest of many composers, such as George Enescu, Pierre Boulez, Mauricio Kagel, György Ligeti and others. The heterophonic syntax has a history of its growth, which means a succession of different concepts and writing techniques. The trajectory of settling this phenomenon does not necessarily take into account the chronology: there are highly complex primary stages and advanced stages of returning to the simple forms of writing. In folklore, the plurimelodic simultaneities are free or random and originate from the (unintentional) differences/‘deviations’ from the state of unison, through a variety of ornaments, melismas, imitations, elongations and abbreviations, all in a flexible rhythmic and non-periodic/immeasurable framework, proper to the parlando-rubato rhythmics. Within the general framework of the multivocal organization, the heterophonic syntax in elaborate (academic) version has imposed itself relatively late compared with polyphony and homophony. Of course, the explanation is simple, if we consider the causal relationship between the sound vocabulary elements – in this case, the modalism – and the typologies of vertical organization appropriate for it. Therefore, adding up the ‘classic’ pathway of the writing typologies (monody – polyphony – homophony), heterophony - applied equally to the structures of modal, serial or synthesis vocabulary – reclaims necessarily an own macrotemporal form, in the sense of the analogies enshrined by the evolution of the musical styles and languages: polyphony→fugue, homophony→sonata. Concerned about the prospect of edifying a new musical ontology, the composer Ştefan Niculescu experienced – along with the mathematical organization of heterophony according to his own original methods – the possibility of extrapolation of this phenomenon in macrostructural plan, reaching this way to the unique form of ‘synchrony’. Founded on coincidentia oppositorum principle (involving the ‘one-multiple’ binom), the sound architecture imagined by Ştefan Niculescu consists in one (temporal) model / algorithm of articulation of two sound states: 1. monovocality state (principle of identity) and 2. multivocality state (principle of difference). In this context, the heterophony becomes an (auto)generative mechanism, with macrotemporal amplitude, strategy that will be grown by the composer, practically throughout his creation (see the works: Ison I, Ison II, Unisonos I, Unisonos II, Duplum, Triplum, Psalmus, Héterophonies pour Montreux (Homages to Enescu and Bartók etc.). For the present demonstration, we selected one of the most edifying works of Ştefan Niculescu – Simphony II, Opus dacicum – where the form of (heterophony-)synchrony acquires monumental-symphonic features, representing an emblematic case for the complexity level achieved by this type of vertical syntax in the twentieth century music.

Keywords: heterophony, modalism, serialism, synchrony, syntax

Procedia PDF Downloads 341

Authors: Rui Yin, Haojie Wang, Xun Li

Abstract:

Recently, mainstream dense retrieval methods have obtained state-of-the-art results on some datasets and tasks. However, they require large amounts of training data, which is not available in most domains. The severe performance degradation of dense retrievers on new data domains has limited the use of dense retrieval methods to only a few domains with large training datasets. In this paper, we propose an unsupervised domain-adaptive approach based on query generation. First, a generative model is used to generate relevant queries for each passage in the target corpus, and then, the generated queries are used for mining negative passages. Finally, the query-passage pairs are labeled with a cross-encoder and used to train a domain-adapted dense retriever. We also explore contrastive learning as a method for training domain-adapted dense retrievers and show that it leads to strong performance in various retrieval settings. Experiments show that our approach is more robust than previous methods in target domains that require less unlabeled data.

Keywords: dense retrieval, query generation, contrastive learning, unsupervised training

Procedia PDF Downloads 101

Authors: Yakubu Bala Mohammed, Nadire Chavus, Mohammed Bulama

Abstract:

Generative Pre-trained Transformer (ChatGPT) represents an artificial intelligence (AI) tool capable of swiftly generating comprehensive responses to prompts and follow-up inquiries. This emerging AI tool was introduced in November 2022 by OpenAI firm, an American AI research laboratory, utilizing substantial language models. This present study aims to delve into the potential future consequences of ChatGPT usage in education using AI-based algorithms. The paper will bring forth the likely potential risks of ChatGBT utilization, such as academic integrity concerns, unfair learning assessments, excessive reliance on AI, and dissemination of inaccurate information using four machine learning algorithms: eXtreme-Gradient Boosting (XGBoost), Support vector machine (SVM), Emotional artificial neural network (EANN), and Random forest (RF) would be used to analyze the study collected data due to their robustness. Finally, the findings of the study will assist education stakeholders in understanding the future implications of ChatGPT usage in education and propose solutions and directions for upcoming studies.

Keywords: machine learning, ChatGPT, education, learning, implications

Procedia PDF Downloads 230

Authors: K. O'Malley

Abstract:

Background: The emergence and rapid evolution of large language models (LLMs) (i.e., models of generative artificial intelligence, or AI) has been unprecedented. ChatGPT is one of the most widely used LLM platforms. Using natural language processing technology, it generates customized responses to user prompts, enabling it to mimic human conversation. Responses are generated using predictive modeling of vast internet text and data swathes and are further refined and reinforced through user feedback. The popularity of LLMs is increasing, with a growing number of students utilizing these platforms for study and revision purposes. Notwithstanding its many novel applications, LLM technology is inherently susceptible to bias and error. This poses a significant challenge in the educational setting, where academic integrity may be undermined. This study aims to evaluate the performance of the latest iteration of ChatGPT (ChatGPT4.0) in standardized state medical licensing examinations. Methods: A considered search strategy was used to interrogate the PubMed electronic database. The keywords ‘ChatGPT’ AND ‘medical education’ OR ‘medical school’ OR ‘medical licensing exam’ were used to identify relevant literature. The search included all peer-reviewed literature published in the past five years. The search was limited to publications in the English language only. Eligibility was ascertained based on the study title and abstract and confirmed by consulting the full-text document. Data was extracted into a Microsoft Excel document for analysis. Results: The search yielded 345 publications that were screened. 225 original articles were identified, of which 11 met the pre-determined criteria for inclusion in a narrative synthesis. These studies included performance assessments in national medical licensing examinations from the United States, United Kingdom, Saudi Arabia, Poland, Taiwan, Japan and Germany. ChatGPT 4.0 achieved scores ranging from 67.1 to 88.6 percent. The mean score across all studies was 82.49 percent (SD= 5.95). In all studies, ChatGPT exceeded the threshold for a passing grade in the corresponding exam. Conclusion: The capabilities of ChatGPT in standardized academic assessment in medicine are robust. While this technology can potentially revolutionize higher education, it also presents several challenges with which educators have not had to contend before. The overall strong performance of ChatGPT, as outlined above, may lend itself to unfair use (such as the plagiarism of deliverable coursework) and pose unforeseen ethical challenges (arising from algorithmic bias). Conversely, it highlights potential pitfalls if users assume LLM-generated content to be entirely accurate. In the aforementioned studies, ChatGPT exhibits a margin of error between 11.4 and 32.9 percent, which resonates strongly with concerns regarding the quality and veracity of LLM-generated content. It is imperative to highlight these limitations, particularly to students in the early stages of their education who are less likely to possess the requisite insight or knowledge to recognize errors, inaccuracies or false information. Educators must inform themselves of these emerging challenges to effectively address them and mitigate potential disruption in academic fora.

Keywords: artificial intelligence, ChatGPT, generative ai, large language models, licensing exam, medical education, medicine, university

Procedia PDF Downloads 30

Authors: Xulu Yao, Moi Hoon Yap, Yanlong Zhang

Abstract:

As a branch of artificial neural network, deep learning is widely used in the field of image recognition, but the lack of its dataset leads to imperfect model learning. By analysing the data scale requirements of deep learning and aiming at the application in GUI generation, it is found that the collection of GUI dataset is a time-consuming and labor-consuming project, which is difficult to meet the needs of current deep learning network. To solve this problem, this paper proposes a semi-supervised deep learning model that relies on the original small-scale datasets to produce a large number of reliable data sets. By combining the cyclic neural network with the generated countermeasure network, the cyclic neural network can learn the sequence relationship and characteristics of data, make the generated countermeasure network generate reasonable data, and then expand the Rico dataset. Relying on the network structure, the characteristics of collected data can be well analysed, and a large number of reasonable data can be generated according to these characteristics. After data processing, a reliable dataset for model training can be formed, which alleviates the problem of dataset shortage in deep learning.

Keywords: GUI, deep learning, GAN, data augmentation

Procedia PDF Downloads 183

Authors: Joseph M. Rich, Vinay A. Duddalwar, Assad A. Oberai

Abstract:

Prostate adenocarcinoma is the most common cancer in males, with osseous metastases as the commonest site of metastatic prostate carcinoma (mPC). Treatment monitoring is based on the evaluation and characterization of lesions on multiple imaging studies, including Computed Tomography (CT). Monitoring of the osseous disease burden, including follow-up of lesions and identification and characterization of new lesions, is a laborious task for radiologists. Deep learning algorithms are increasingly used to perform tasks such as identification and segmentation for osseous metastatic disease and provide accurate information regarding metastatic burden. Here, nnUNet was used to produce a model which can segment CT scan images of prostate adenocarcinoma vertebral bone metastatic lesions. nnUNet is an open-source Python package that adds optimizations to deep learning-based UNet architecture but has not been extensively combined with transfer learning techniques due to the absence of a readily available functionality of this method. The IRB-approved study data set includes imaging studies from patients with mPC who were enrolled in clinical trials at the University of Southern California (USC) Health Science Campus and Los Angeles County (LAC)/USC medical center. Manual segmentation of metastatic lesions was completed by an expert radiologist Dr. Vinay Duddalwar (20+ years in radiology and oncologic imaging), to serve as ground truths for the automated segmentation. Despite nnUNet’s success on some medical segmentation tasks, it only produced an average Dice Similarity Coefficient (DSC) of 0.31 on the USC dataset. DSC results fell in a bimodal distribution, with most scores falling either over 0.66 (reasonably accurate) or at 0 (no lesion detected). Applying more aggressive data augmentation techniques dropped the DSC to 0.15, and reducing the number of epochs reduced the DSC to below 0.1. Datasets have been identified for transfer learning, which involve balancing between size and similarity of the dataset. Identified datasets include the Pancreas data from the Medical Segmentation Decathlon, Pelvic Reference Data, and CT volumes with multiple organ segmentations (CT-ORG). Some of the challenges of producing an accurate model from the USC dataset include small dataset size (115 images), 2D data (as nnUNet generally performs better on 3D data), and the limited amount of public data capturing annotated CT images of bone lesions. Optimizations and improvements will be made by applying transfer learning and generative methods, including incorporating generative adversarial networks and diffusion models in order to augment the dataset. Performance with different libraries, including MONAI and custom architectures with Pytorch, will be compared. In the future, molecular correlations will be tracked with radiologic features for the purpose of multimodal composite biomarker identification. Once validated, these models will be incorporated into evaluation workflows to optimize radiologist evaluation. Our work demonstrates the challenges of applying automated image segmentation to small medical datasets and lays a foundation for techniques to improve performance. As machine learning models become increasingly incorporated into the workflow of radiologists, these findings will help improve the speed and accuracy of vertebral metastatic lesions detection.

Keywords: deep learning, image segmentation, medicine, nnUNet, prostate carcinoma, radiomics

Procedia PDF Downloads 95

Authors: Al Karim Datoo

Abstract:

Abstract— The presentation argues for an urgent case to center and integrate critical sociology in enriching potency of educational thought and practice to counteract inequalities and social injustices. COVID phenomenon has starkly exposed burgeoning of social-economic inequalities and widening marginalities which have been historically and politically constructed through deep-seated social and power imbalances and injustices in the world. What potent role could education possibly play to combat these issues? A point of departure for this paper highlights increasing reductionist and exclusionary ‘mind-set’ of education that has been developed through trends in education such as: the commodification of knowledge, standardisation, homogenization, and reification which are products of the positivist ideology of knowledge coopted to serve capitalist interests. To redress these issues of de-contextualization and de-humanization of education, it is emphasized that there is an urgent need to center the role of interpretive and critical epistemologies and pedagogies of social sciences. In this regard, notions of problem-posing versus problem-solving, generative themes, instrumental versus emancipatory reasoning will be discussed. The presentation will conclude by illustrating the pedagogic utility of these critically oriented notions to counteract the social reproduction of exclusionary and inequality in and through education.

Keywords: Critical pedagogy, social justice, inclusion , education

Procedia PDF Downloads 112

Authors: Milind Patil

Abstract:

Present paper is a discussion on Hindi-Marathi Morphological Analysis and generating rules for Machine Translation on the basis of Machine Readable Dictionary (MRD). This used Transformative Generative Grammar (TGG) rules to design the MRD. As per TGG rules, the suffix of a particular root word is based on its Tense, Aspect, Modality and Voice. That's why the suffix is very important for the word meanings (or root meanings). The Hindi and Marathi Language both have relation with Indo-Aryan language family. Both have been derived from Sanskrit language and their script is 'Devnagari'. But there are lots of differences in terms of semantics and grammatical level too. In Marathi, there are three genders, but in Hindi only two (Masculine and Feminine), the Natural gender is absent in Hindi. Likewise other grammatical categories also differ in their level of use. For MRD the suffixes (or Morpheme) are of particular root word for GNP (Gender, Number and Person) are based on its natural phenomena. A particular Suffix and Morphine change as per the need of person, number and gender. The design of MRD also based on this format. In first, Person, Number, Gender and Tense are key points than root words and suffix of particular Person, Number Gender (PNG). After that the inferences are drawn on the basis of rules that is (V.stem) (Pre.T/Past.T) (x) + (Aux-Pre.T) (x) → (V.Stem.) + (SP.TM) (X).

Keywords: MRD, TGG, stem, morph, morpheme, suffix, PNG, TAM&V, root

Procedia PDF Downloads 324

Authors: Shriya Shukla, Lachin Fernando

Abstract:

Pneumonia, a leading cause of mortality in young children globally, presents significant diagnostic challenges, particularly in resource-limited settings. This study presents an approach to diagnosing pediatric pneumonia using Chest X-Ray (CXR) images, employing a lightweight MobileNetV2 model enhanced with synthetic data augmentation. Addressing the challenge of dataset scarcity and imbalance, the study used a Variational Autoencoder-Generative Adversarial Network (VAE-GAN) to generate synthetic CXR images, improving the representation of normal cases in the pediatric dataset. This approach not only addresses the issues of data imbalance and scarcity prevalent in medical imaging but also provides a more accessible and reliable diagnostic tool for early pneumonia detection. The augmented data improved the model’s accuracy and generalization, achieving an overall accuracy of 95% in pneumonia detection. These findings highlight the efficacy of the MobileNetV2 model, offering a computationally efficient yet robust solution well-suited for resource-constrained environments such as mobile health applications. This study demonstrates the potential of synthetic data augmentation in enhancing medical image analysis for critical conditions like pediatric pneumonia.

Keywords: pneumonia, MobileNetV2, image classification, GAN, VAE, deep learning

Procedia PDF Downloads 123

Authors: M. Ben Jrad, A. Belhaj Mohamed, S. Riahi, I. Bouazizi, M. Saidi, M. Soussi

Abstract:

The generative potential, depositional environment, thermal maturity and oil seeps of the organic-rich Bou Dabbous Formation (Ypresian) from the thrust belt northwestern Tunisia, were determined by Rock Eval and molecular analyses. The paleo-tectonic units in the area show some similarities with equivalent facies in Mediterranean Sea and Sicilian. The Bou Dabbous Formation displays variable source rock characteristics through the various units Tellian and Numidian nappes Units. Organic matter contents and petroleum potentials are fair to high (reaching 1.95% and 6 kg of HC/t of rock respectively) marine type II kerogen. An increasing SE-NW maturity gradient is well documented in the study area. The Bou Dabbous organic-rich facies are marginally mature stage in the Tellian Unit (Kasseb domain), whilst they are mature-late mature stage within Nefza-Ain Allega tectonic windows. A long and north of Cap Serrat-Ghardimaou Master Fault these facies are overmature. Oil/Oil and Oil/source rock correlation, based on biomarker and carbon isotopic composition, shows a positive genetic correlation between the oil seeps and Bou Dabbous source rock.

Keywords: biomarkers, Bou Dabbous Formation, Northern Tunisia, source rock

Procedia PDF Downloads 485

Authors: Mian Huang, Chi Ma, Junyu Lin, William Lu

Abstract:

Introduction: Bone health attracts more attention recently and an intelligent question and answer (QA) chatbot for osteoporosis is helpful for science popularization. With Generative Pretrained Transformer (GPT) technology developing, we build an osteoporosis corpus dataset and then fine-tune LLaMA, a famous open-source GPT foundation large language model(LLM), on our self-constructed osteoporosis corpus. Evaluated by clinical orthopedic experts, our fine-tuned model outperforms vanilla LLaMA on osteoporosis QA task in Chinese. Three-dimensional quantitative computed tomography (QCT) measured bone mineral density (BMD) is considered as more accurate than DXA for BMD measurement in recent years. We develop an automatic Phantom-less QCT(PL-QCT) that is more efficient for BMD measurement since no need of an external phantom for calibration. Combined with LLM on osteoporosis, our PL-QCT provides efficient and accurate BMD measurement for our chatbot users. Material and Methods: We build an osteoporosis corpus containing about 30,000 Chinese literatures whose titles are related to osteoporosis. The whole process is done automatically, including crawling literatures in .pdf format, localizing text/figure/table region by layout segmentation algorithm and recognizing text by OCR algorithm. We train our model by continuous pre-training with Low-rank Adaptation (LoRA, rank=10) technology to adapt LLaMA-7B model to osteoporosis domain, whose basic principle is to mask the next word in the text and make the model predict that word. The loss function is defined as cross-entropy between the predicted and ground-truth word. Experiment is implemented on single NVIDIA A800 GPU for 15 days. Our automatic PL-QCT BMD measurement adopt AI-associated region-of-interest (ROI) generation algorithm for localizing vertebrae-parallel cylinder in cancellous bone. Due to no phantom for BMD calibration, we calculate ROI BMD by CT-BMD of personal muscle and fat. Results & Discussion: Clinical orthopaedic experts are invited to design 5 osteoporosis questions in Chinese, evaluating performance of vanilla LLaMA and our fine-tuned model. Our model outperforms LLaMA on over 80% of these questions, understanding ‘Expert Consensus on Osteoporosis’, ‘QCT for osteoporosis diagnosis’ and ‘Effect of age on osteoporosis’. Detailed results are shown in appendix. Future work may be done by training a larger LLM on the whole orthopaedics with more high-quality domain data, or a multi-modal GPT combining and understanding X-ray and medical text for orthopaedic computer-aided-diagnosis. However, GPT model gives unexpected outputs sometimes, such as repetitive text or seemingly normal but wrong answer (called ‘hallucination’). Even though GPT give correct answers, it cannot be considered as valid clinical diagnoses instead of clinical doctors. The PL-QCT BMD system provided by Bone’s QCT(Bone’s Technology(Shenzhen) Limited) achieves 0.1448mg/cm2(spine) and 0.0002 mg/cm2(hip) mean absolute error(MAE) and linear correlation coefficient R2=0.9970(spine) and R2=0.9991(hip)(compared to QCT-Pro(Mindways)) on 155 patients in three-center clinical trial in Guangzhou, China. Conclusion: This study builds a Chinese osteoporosis corpus and develops a fine-tuned and domain-adapted LLM as well as a PL-QCT BMD measurement system. Our fine-tuned GPT model shows better capability than LLaMA model on most testing questions on osteoporosis. Combined with our PL-QCT BMD system, we are looking forward to providing science popularization and early morning screening for potential osteoporotic patients.

Keywords: GPT, phantom-less QCT, large language model, osteoporosis

Procedia PDF Downloads 70