检索结果-内蒙古大学图书馆

ChatGPT and Generative AI Guidelines for Addressing Academic Integrity and Augmenting Pre-Existing Chatbots

学校读者我要写书评

暂无评论

ChatGPT and Generative AI Guidelines for Addressing Academic...

IEEE International Conference on Industrial Technology (ICIT)

作者： Daswin De Silva Nishan Mills Mona El-Ayoubi Milos Manic Damminda Alahakoon Centre for Data Analytics and Cognition La Trobe University Melbourne Australia Education Services La Trobe University Melbourne Australia Department of Computer Science Virginia Commonwealth University Richmond USA

Chat Generative Pretrained Transformer (Chat-GPT) and related Generative AI models are leading a paradigm shift in the acceptance and application of Artificial Intelligence (AI) across all disciplines and industry sectors. Despite the criticisms of an ‘intelligence without knowledge or reasoning or the notions of truth’, ChatGPT is highly effective at human-like conversation with seemingly sophisticated and useful responses to questions, summarization, classification, extraction and generation tasks. Unlike similar large AI models in the modalities of image, audio and video, text-based conversation is straightforward and familiar to a large audience of regular users of the Internet and smartphone applications. This is further accentuated by the large-scale adoption of ‘standard’ chatbot technologies for trivial conversations in task-specific automation, across every industry sector. This rare combination of highly effective human-like conversation, familiarity of foundational technology and versatility of intelligent application, has led to several challenges and opportunities in leveraging generative AI. A primary challenge is its impact on the academic integrity of scholarly work, where AI-generated content can be useful and detrimental in both teaching and research. On the other hand, ChatGPT presents a unique opportunity in augmenting preexisting (‘standard’) chatbots with human-like conversation for advanced intelligent automation, across all application domains. Although diametrically opposed, the challenge of addressing academic integrity and the opportunity of augmenting pre-existing chatbots are grounded in the conversational AI capabilities of ChatGPT and similar generative AI models. In this paper, we investigate these formative capabilities and present guidelines for leveraging ChatGPT and similar generative AI models.

关键词：

Suspension Analysis and Selective Continuation-Passing Style for Universal Probabilistic Programming Languages

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Lundén, Daniel Hummelgren, Lars Kudlicka, Jan Eriksson, Oscar Broman, David Oracle Stockholm Sweden EECS and Digital Futures KTH Royal Institute of Technology Stockholm Sweden Department of Data Science and Analytics BI Norwegian Business School Oslo Norway Computer Science Department Stanford University CA United States

Universal probabilistic programming languages (PPLs) make it relatively easy to encode and automatically solve statistical inference problems. To solve inference problems, PPL implementations often apply Monte Carlo inference algorithms that rely on execution suspension. State-of-the-art solutions enable execution suspension either through (i) continuation-passing style (CPS) transformations or (ii) efficient, but comparatively complex, low-level solutions that are often not available in high-level languages. CPS transformations introduce overhead due to unnecessary closure allocations—a problem the PPL community has generally overlooked. To reduce overhead, we develop a new efficient selective CPS approach for PPLs. Specifically, we design a novel static suspension analysis technique that determines parts of programs that require suspension, given a particular inference algorithm. The analysis allows selectively CPS transforming the program only where necessary. We formally prove the correctness of the analysis and implement the analysis and transformation in the Miking CorePPL compiler. We evaluate the implementation for a large number of Monte Carlo inference algorithms on real-world models from phylogenetics, epidemiology, and topic modeling. The evaluation results demonstrate significant improvements across all models and inference algorithms. Copyright © 2023, The Authors. All rights reserved.

关键词： Static analysis

Principles and Components of Federated Learning Architectures

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Al Nasim, Abdullah Soshi, Fatema Tuj Johura Biswas, Parag Ferdous, A.S.M. Anas Rashid, Abdur Biswas, Angona Gupta, Kishor Datta Research and Development Department Pioneer Alpha Dhaka Bangladesh Msc in Data Science and Analytics University of Hertfordshire Hatfield United Kingdom Department of Biomedical Engineering Bangladesh University of Engineering and Technology Dhaka Bangladesh MSEM Department Westcliff university CA United States Department of Computer and Information Science Clark Atlanta University Georgia United States

Federated Learning (FL) is a machine learning framework where multiple clients, from mobiles to enterprises, collaboratively construct a model under the orchestration of a central server but still retain the decentralized nature of the training data. This decentralized training of models offers numerous advantages, including cost savings, enhanced privacy, improved security, and compliance with legal requirements. However, for all its apparent advantages, FL is not immune to the limitations of conventional machine learning methodologies. This article provides an elaborate explanation of the inherent concepts and features found within federated learning architecture, addressing five key domains: system heterogeneity, data partitioning, machine learning models, communication protocols, and privacy techniques. This article also highlights the limitations in this domain and proposes avenues for future work. Besides, we provide a set of architectural patterns for federated learning systems, which are derived from the systematic survey of the literature. The main elements of FL, the fundamentals of Federated Learning, and a few architectural specifics will all be better understood with the aid of this research. © 2025, CC0.

关键词： Federated learning

Designing an Energy Tracker and Smart Home Solutions Using Design Thinking Methodology

学校读者我要写书评

暂无评论

Designing an Energy Tracker and Smart Home Solutions Using D...

International Conference on ICT For Smart Society (ICISS)

作者： Siti Elda Hiererra Victor Chandra Jane Ivana Jessica Lynn Wibowo Faisal Asadi Information Systems Department School of Information Systems Bina Nusantara University Jakarta Indonesia Business Analytics Department School of Information Systems Bina Nusantara University Jakarta Indonesia Computer Science Department School of Computer Science Bina Nusantara University Jakarta Indonesia Bioinformatics and Data Science Research Center School of Computer Science Bina Nusantara University Jakarta Indonesia

As energy consumption rises, finding ways to reduce energy usage in homes is crucial. This paper presents a case study using design thinking to create a daily energy tracker that helps users reduce energy consumption. We conduct user research, define the problem, brainstorm solutions (mobile app, smart thermostat, smart plug, physical tracker), and test prototypes. A mobile app that tracks energy usage in real-time and offers personalized tips emerges as the most promising solution. We develop and launch the app, partnering with an energy company. This case study showcases the effectiveness of design thinking in creating a user-centered solution for reducing energy consumption.

关键词：

A Novel Multi-Scale Deep Learning Approach for the Early Detection of Alzheimer’s Disease Using fMRI

学校读者我要写书评

暂无评论

A Novel Multi-Scale Deep Learning Approach for the Early Det...

computer, Big data and Artificial Intelligence (ICCBD+AI), International Conference on

作者： Ateeq Ur Rehman Butt Isma Hamid Qamar Nawaz Tariq Mahmood Xiang Zhang Muhammad Yaqub Department of Computer Science National Textile University Faisalabad Pakistan Department of Computer Science University of Agriculture Faisalabad Pakistan Artificial Intelligence and Data Analytics Lab CCIS Prince Sultan University Riyadh Kingdom of Saudi Arabia Institute of Biophysics Chinese Academy of Sciences Beijing China School of Biomedical Sciences Hunan University Changsha China

ISBN: (数字)9798331533991

ISBN: (纸本)9798331534004

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that remains challenging to diagnose in its early stages. This study presents a novel 3D multi-scale Convolutional Neural Network (CNN) model that utilizes resting-state functional MRI (rs-fMRI) data to classify different stages of Alzheimer’s disease. The proposed model integrates multi-scale feature extraction from the hippocampal region, optimizing the detection of fine-grained and global patterns indicative of AD progression. The model was evaluated on a balanced dataset of 300 subjects, applying cross-validation to ensure robustness across multiple binary classification tasks. Our experimental results demonstrate a classification accuracy of up to 93% for distinguishing Alzheimer’s Disease from Cognitively Normal subjects. High sensitivity, specificity, and precision were observed across all classification tasks, indicating the model’s effectiveness in early-stage Alzheimer’s detection. These findings suggest the proposed approach could enhance clinical diagnosis and facilitate timely interventions for AD patients.

关键词： Neuroimaging Solid modeling Accuracy Three-dimensional displays Sensitivity Functional magnetic resonance imaging data models Robustness Convolutional neural networks Alzheimer's disease

Modeling lens potentials with continuous neural fields in galaxy-scale strong lenses

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Biggio, Luca Vernardos, Georgios Galan, Aymeric Peel, Austin Data Analytics Lab Institute of Machine Learning Department of Computer Science ETHZ Switzerland Observatoire de Sauverny Versoix1290 Switzerland

Strong gravitational lensing is a unique observational tool for studying the dark and luminous mass distribution both within and between galaxies. Given the presence of substructures, current strong lensing observations demand more complex mass models than smooth analytical profiles, such as power-law ellipsoids. In this work, we introduce a continuous neural field to predict the lensing potential at any position throughout the image plane, allowing for a nearly model-independent description of the lensing mass. We apply our method on simulated Hubble Space Telescope imaging data containing different types of perturbations to a smooth mass distribution: a localized dark subhalo, a population of subhalos, and an external shear perturbation. Assuming knowledge of the source surface brightness, we use the continuous neural field to model either the perturbations alone or the full lensing potential. In both cases, the resulting model is able to fit the imaging data, and we are able to accurately recover the properties of both the smooth potential and of the perturbations. Unlike many other deep learning methods, ours explicitly retains lensing physics (i.e., the lens equation) and introduces high flexibility in the model only where required, namely, in the lens potential. Moreover, the neural network does not require pre-training on large sets of labelled data and predicts the potential from the single observed lensing image. Our model is implemented in the fully differentiable lens modeling code Herculens. © 2022, CC BY.

关键词： Gravitation

Early Fault Detection in Rolling Element Bearings Using Signal Processing Techniques

学校读者我要写书评

暂无评论

Early Fault Detection in Rolling Element Bearings Using Sign...

International Symposium on High Capacity Optical Networks and Enabling Technologies, HONET

作者： Ali Saeed M. Usman Akram Muazzam A. Khan M. Belal Afsar Department of Computer Science ICESCO Chair Big Data Analytics and Edge Computing Quaid-i-Azam University Islamabad Pakistan Department of Computer and Software Engineering National University of Sciences & Technology Islamabad Pakistan Faculty of Electrical Engineering and Information Technology Otto-von-Guericke University Maddeburg DE

ISBN: (数字)9798350378078

ISBN: (纸本)9798350378085

Many studies show that bearings are the most vulnerable components in low-voltage motors. While advanced bearing diagnostic systems exist, their cost can be a barrier for non-critical machinery due to the potential wait time to see a return on investment. This research explores the feasibility of using basic signal processing techniques on vibration data for bearing condition monitoring. The primary goal is to analyze benchmark data from the Case Western Reserve University (CWRU) dataset and establish a baseline performance for this data-driven approach. The results suggest that, with appropriate signal processing, it's possible to achieve early detection of bearing faults, leading to more efficient time-to-fault identification with higher detection accuracies ranging from 83 % to 100 % for each method.

关键词： Vibrations Fault diagnosis Low voltage Smart cities Rolling bearings Signal processing Motors Robots Monitoring Investment

Unsupervised Alignment of Distributional Word Embeddings

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Diallo, Aïssatou Fürnkranz, Johannes Department of Computer Science University College London United Kingdom Computational Data Analytics FAW Johannes Kepler University Linz Austria

Cross-domain alignment plays a key role in tasks ranging from image-text retrieval to machine translation. The main objective is to associate related entities across different domains. Recently, purely unsupervised methods operating on monolingual embeddings have successfully been used to infer a bilingual lexicon without relying on supervision. However, current state-of-the art methods only focus on point vectors although distributional embeddings have proven to embed richer semantic information when representing words. This paper investigates a novel stochastic optimization approach for aligning word distributional embeddings. Our method builds upon techniques in optimal transport to resolve the cross-domain matching problem in a principled manner. We evaluate our method on the problem of unsupervised word translation, by aligning word embeddings trained on monolingual data. We present empirical evidence to demonstrate the validity of our approach to the bilingual lexicon induction task across several language pairs. © 2022, CC BY.

关键词： Embeddings

GausSetExpander: A Simple Approach for Entity Set Expansion

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Diallo, Aïssatou Fürnkranz, Johannes Department of Computer Science University College London United Kingdom Computational Data Analytics FAW Johannes Kepler University Linz Austria

Entity Set Expansion (ESE) is an important task in Natural Language Processing that aims at expanding a small set of entities into a larger one with items from a large pool of candidates. The problem implicitly requires the definition of the notion of similarity between the given entities and the candidates. In this paper, we propose GausSetExpander, an unsupervised approach for the task of ESE based on optimal transport techniques. We propose to re-frame the problem as choosing the entity that best completes the input set. For this, we interpret a set as an elliptical distribution with a centroid which represents the mean and a dispersion that serves as the spread of the variance. The best candidate entity is the one that increases the spread of the set the least. We analyze the strength and the weaknesses of the proposed solution in order to assess the validity of our proposed approach. © 2022, CC BY.

关键词： Semantics