Open Access. Powered by Scholars. Published by Universities.®
- Institution
-
- Singapore Management University (4424)
- TÜBİTAK (1233)
- Purdue University (1090)
- Old Dominion University (946)
- Missouri University of Science and Technology (917)
-
- University of Nebraska - Lincoln (874)
- Air Force Institute of Technology (847)
- Wright State University (817)
- San Jose State University (616)
- Dartmouth College (581)
- Edith Cowan University (567)
- Brigham Young University (506)
- California Polytechnic State University, San Luis Obispo (448)
- City University of New York (CUNY) (445)
- University of Texas at El Paso (387)
- Technological University Dublin (367)
- Portland State University (356)
- New Jersey Institute of Technology (342)
- University of Central Florida (334)
- China Simulation Federation (315)
- Washington University in St. Louis (314)
- Embry-Riddle Aeronautical University (312)
- Nova Southeastern University (295)
- Kennesaw State University (284)
- University of Arkansas, Fayetteville (267)
- Utah State University (259)
- Zayed University (254)
- University of South Florida (252)
- University of Texas at Arlington (252)
- Syracuse University (245)
- Keyword
-
- Machine learning (1030)
- Deep learning (592)
- Machine Learning (530)
- Artificial intelligence (505)
- Security (366)
-
- Computer Science (342)
- Deep Learning (305)
- Artificial Intelligence (268)
- Classification (263)
- Cybersecurity (245)
- Computer vision (230)
- Privacy (214)
- Department of Computer Science and Engineering (211)
- Neural networks (205)
- Algorithms (202)
- Data mining (193)
- Natural language processing (185)
- Computer science (182)
- Applied sciences (171)
- College for Professional Studies (165)
- Software engineering (163)
- School of Computer & Information Science (150)
- Simulation (145)
- AI (142)
- Natural Language Processing (142)
- Optimization (138)
- Blockchain (134)
- Image processing (133)
- Clustering (128)
- Education (123)
- Publication Year
- Publication
-
- Research Collection School Of Computing and Information Systems (4218)
- Theses and Dissertations (1687)
- Turkish Journal of Electrical Engineering and Computer Sciences (1233)
- Department of Computer Science Technical Reports (906)
- Master's Projects (565)
-
- Computer Science Faculty Publications (435)
- Computer Science Technical Reports (404)
- Computer Science Faculty Research & Creative Works (385)
- Electronic Theses and Dissertations (385)
- Dissertations (362)
- The R Journal (353)
- Faculty Publications (348)
- Journal of System Simulation (315)
- CCAC Theses and Dissertations (277)
- Browse all Theses and Dissertations (262)
- All Works (254)
- Theses (253)
- Computer Science Faculty Publications and Presentations (230)
- USF Tampa Graduate Theses and Dissertations (230)
- Chulalongkorn University Theses and Dissertations (Chula ETD) (217)
- Masters Theses (216)
- Computer Science & Engineering Syllabi (214)
- Departmental Technical Reports (CS) (212)
- Walden Dissertations and Doctoral Studies (210)
- Master's Theses (209)
- All Computer Science and Engineering Research (205)
- Kno.e.sis Publications (196)
- Computer Science and Software Engineering (180)
- Open Access Theses & Dissertations (171)
- Regis University Student Publications (comprehensive collection) (166)
- Publication Type
Articles 2101 - 2130 of 27744
Full-Text Articles in Entire DC Network
Ai-Based Accessibility Widget (Aibaw) Shortcomings For Blind Web Users, Joshua A. Rovira
Ai-Based Accessibility Widget (Aibaw) Shortcomings For Blind Web Users, Joshua A. Rovira
LSU Master's Theses
With legal, ethical, and financial motivations to make their websites more accessible, many businesses and sites have begun to employ the use of artificial intelligence (AI) based widgets to automatically make necessary modifications to their sites' pages and subdomains. In this work, we conduct a qualitative case study to determine the efficacy of these AI tools as they pertain specifically to blind users. We analyze pages from twelve websites using accessiBe's AI accessibility widget and provide a taxonomy of their violations against the Web Content Accessibility Guidelines (WCAG) 2.1 level AA compliance. We found each website to bear numerous violations …
Looking Good: The Math Behind Computer Vision*, Corbin Weiss
Looking Good: The Math Behind Computer Vision*, Corbin Weiss
Campus Research Month
Exploring the mathematical foundations of a Multilayer Perceptron (MLP), a foundational approach to computer vision. Then expanding this understanding to create a visualization of the representation of reality in the MLP.
The Fear Of Replacement: How Ai Panic In Journalism Mirrors Existential Crisis In Industry, James Hutson
The Fear Of Replacement: How Ai Panic In Journalism Mirrors Existential Crisis In Industry, James Hutson
Faculty Scholarship
This study systematically examines the portrayal of artificial intelligence (AI) errors, such as hallucinations and deepfakes, in journalistic contexts, evaluating whether these narratives reflect a broader existential anxiety about AI's role in reshaping journalism. Using a systematic literature review combined with a qualitative content analysis of recent AI-focused news reports, this study identifies recurring themes in media coverage to assess the accuracy and context of reported AI errors relative to actual technological limitations and affordances. Findings suggest that while AI errors are comparatively rare, they receive amplified coverage, often fueling public mistrust in AI technologies. Nevertheless, a balanced examination reveals …
47 - Enhancing Patient Experience: A Secure Self-Service Kiosk For Prescription Status In Pharmacies, Raneem Alarian
47 - Enhancing Patient Experience: A Secure Self-Service Kiosk For Prescription Status In Pharmacies, Raneem Alarian
Undergraduate Research Symposium
Creating secure and dependable pharmacy kiosk systems is essential for improving both patient satisfaction and the overall efficiency of pharmacy operations. Patients often experience frustration due to long wait times at the pharmacy counter, especially when prescriptions are not ready. As a licensed pharmacy technician, I have witnessed this issue firsthand, which led me to develop a software solution developed to improve pharmacy workflow. This project focuses on the design and execution of a client-server application for a self-service pharmacy kiosk, enabling patients to securely check the status of their prescriptions before they reach the counter. Developed using Visual Studio …
Evaluating Wrist Placement And Signal Processing Techniques For Real-World Hrv Monitoring Using Ppg, Andrew Murphy
Evaluating Wrist Placement And Signal Processing Techniques For Real-World Hrv Monitoring Using Ppg, Andrew Murphy
College of Computing and Digital Media Dissertations
This thesis investigates trade-offs between signal quality and data coverage in photoplethysmographic (PPG) heart rate variability (HRV) monitoring using wrist-worn devices. The goal was to evaluate whether wrist placement and signal processing techniques can improve measurement reliability in real-world conditions. Data was collected from healthy participants wearing smartwatches on both wrists during rest and a structured math task introducing natural wrist movement. Three distinct processing methodologies were compared, including a proposed Rolling-Standardized Derivative (RSD) approach. Results showed that while HRV signals from both wrists were highly correlated at rest, motion caused a measurable drop in signal quality and inter-wrist agreement, …
Simplifying 3d Printing Using Natural Language Processing, Jared E. Rosenberger
Simplifying 3d Printing Using Natural Language Processing, Jared E. Rosenberger
Undergraduate Theses
3D printing is a crucial technology with many applications in different fields. To be able to use this technology to its full extent, expertise in computer aided design (CAD) technology and 3D modeling is required. Many people interested in 3D printing do not have this expertise and thus cannot build custom models, and are consequently forced to buy them instead. Natural language processing (NLP) is one tool that can vastly simplify 3D modeling for those lacking CAD experience. Using NLP, someone can simply dictate what they want to be able to print, and a computer can then build a 3D …
From Adversarial Attacks To Robust Classifiers - A Study In Social Media Spam Detection - Black Box & White Box, Jonathan Jose Penaloza Rumie
From Adversarial Attacks To Robust Classifiers - A Study In Social Media Spam Detection - Black Box & White Box, Jonathan Jose Penaloza Rumie
Undergraduate Theses
Adversarial attacks pose a significant threat to the reliability of machine learning-based spam detection systems in social media. This undergraduate thesis, "From Adversarial Attacks to Robust Classifiers: A Study in Social Media Spam Detection – Black Box & White Box," systematically examines the impact of both black-box and white-box adversarial attacks on a range of spam classifiers, including Logistic Regression, Decision Trees, Random Forests, K-Nearest Neighbors, Bagging, Gradient Boosting, and Support Vector Machines. Leveraging a novel dataset derived from Twitter spam messages and enhanced with adversarial perturbations such as synonym replacement and character-level modifications, this study evaluates classifier performance under …
32 - Nested Two Level Decomposition For Quantum Computing, Andrew Maciejunes, John Stenger, Dan Gunlycke, Nikos Chrisochoides
32 - Nested Two Level Decomposition For Quantum Computing, Andrew Maciejunes, John Stenger, Dan Gunlycke, Nikos Chrisochoides
Undergraduate Research Symposium
Abstract—We present a two-level decomposition strategy for solving the Vehicle Routing Problem (VRP) using the Quantum Approximate Optimization Algorithm (QAOA). A Problem-Level Decomposition (PLD) partitions a 9-node (72-qubit) VRP into smaller Traveling Salesman Problem (TSP) instances. Each TSP is then further simplified via Circuit-Level Decomposition (CLD), enabling execution on near-term quantum devices. Our approach achieves up to 90% reductions in circuit depth and qubit count. These results demonstrate the feasibility of solving VRPs previously too complex for quantum simulators and provide early evidence of potential quantum utility.
Enhancing State-Of-The-Art Motor Imagery Classification With Reinforcement Learning, Anton Shepelev
Enhancing State-Of-The-Art Motor Imagery Classification With Reinforcement Learning, Anton Shepelev
USF Tampa Graduate Theses and Dissertations
One of the key obstacles to the rapid adoption of non-invasive Brain-Computer Interfaces (BCIs) for Motor Imagery (MI) is the low signal-to-noise ratio, and the substantial data requirements which can be mentally taxing for users. EEGNet, a compact Convolutional Neural Network (CNN), has long been considered the state-of-the-art (SOTA) for MI classification, demonstrating strong performance even with limited data. However, recent studies advocate for integrating Deep Reinforcement Learning (RL) to further enhance classification accuracy by dynamically optimizing feature extraction and decision-making processes. Despite this potential, practical implementations remain scarce due to challenges in stabilizing RL training and adapting it to …
Unsupervised Contrastive Learning Based Clustering For Robust Emotion Classification, Nozaer Omar, Sanjay Kumar Madria
Unsupervised Contrastive Learning Based Clustering For Robust Emotion Classification, Nozaer Omar, Sanjay Kumar Madria
Miners Solving for Tomorrow Research Conference
No abstract provided.
On The Robustness Of Adaptive Resonance Theory Neural Networks, Shane Cairns, Leonardo Enzo Brito Da Silva, Sasha Petrenko, Donald C. Wunsch
On The Robustness Of Adaptive Resonance Theory Neural Networks, Shane Cairns, Leonardo Enzo Brito Da Silva, Sasha Petrenko, Donald C. Wunsch
Miners Solving for Tomorrow Research Conference
No abstract provided.
Reframing Information Seeking In The Age Of Generative Ai: A Critical And Humanistic Approach, Joseph Kevin Sebastian
Reframing Information Seeking In The Age Of Generative Ai: A Critical And Humanistic Approach, Joseph Kevin Sebastian
Library Faculty Research
Information-seeking has long been the subject of theoretical modeling, often drawing from cognitive, behavioral, computational, and even evolutionary perspectives to explain how individuals navigate, filter, and utilize information. Several dominant frameworks—Carol Kuhlthau’s Information Search Process, Marcia Bates’ Berrypicking Model, Peter Pirolli & Stuart Card’s Information Foraging Theory, Kiyohiko Nakamura’s Information Criteria framework, and Ian Ruthven’s Information Shaping Theory —have provided structured ways of understanding how people interact with information environments. However, while these frameworks offer valuable insights, they often operate within mechanistic or efficiency-driven paradigms, which risk overlooking the complex, embodied, and socioculturally situated nature of human information behaviors. These …
Why Um And U*Log(U) Are The Most Effective Nonlinear Functions In Fuzzy Clustering: Theoretical Explanation Of The Empirical Fact, Olga Kosheleva, Vladik Kreinovich, Yuchi Kanzawa
Why Um And U*Log(U) Are The Most Effective Nonlinear Functions In Fuzzy Clustering: Theoretical Explanation Of The Empirical Fact, Olga Kosheleva, Vladik Kreinovich, Yuchi Kanzawa
Departmental Technical Reports (CS)
In fuzzy clustering, we need to have non-linear functions of the membership degrees. Different nonlinear functions have been tried. Empirical evidence shows that for fuzzy clustering, the most effective nonlinear functions are um and u*log(u). In this paper, we provide a theoretical explanation for this empirical fact.
Why Interval-Valued (And Type-2) Fuzzy Methods Are Often More Effective, Christian Servin, Olga Kosheleva, Vladik Kreinovich
Why Interval-Valued (And Type-2) Fuzzy Methods Are Often More Effective, Christian Servin, Olga Kosheleva, Vladik Kreinovich
Departmental Technical Reports (CS)
Interval-valued and type-2 fuzzy techniques were designed to provide a more adequate representation of expert knowledge than the traditional (type-1) fuzzy techniques. Somewhat unexpectedly, they also often turn out to be more effective even when there is no expert knowledge at all -- when we are simply using fuzzy rules to fit experimental data. In precise terms, for the same number of parameters, interval-valued and type-2 systems often provide a better fit for the data and/or better quality control than traditional (type-1) fuzzy techniques. In this paper, we provide a theoretical explanation for this surprising phenomenon.
Grounding Ai Use In Learning Science: A Conversation With Steven Miller, Steven Miller, Lieven Demeester
Grounding Ai Use In Learning Science: A Conversation With Steven Miller, Steven Miller, Lieven Demeester
CASTLe: Collection of Articles on Scholarship for Teaching and Learning
In this insightful interview, SMU Associate Provost (Teaching and Learning Innovation) Lieven Demeester and Professor Emeritus of Information Systems Steven Miller discuss the integration of artificial intelligence (AI) in teaching and learning, emphasising the importance of grounding AI use in the fundamentals of learning science. They explore the evolving role of education in the context of AI advancements, highlighting the need for educators to focus on the cognitive aspects of learning, such as goal-directed practice and feedback. They also address the potential of AI as a collaborative agent in group projects and the importance of maintaining accountability and quality control …
Impact Of Data Snooping On Deep Learning Models For Locating Vulnerabilities In Lifted Code, Gary Mccully, John Hastings, Shengjie Xu
Impact Of Data Snooping On Deep Learning Models For Locating Vulnerabilities In Lifted Code, Gary Mccully, John Hastings, Shengjie Xu
Research & Publications
This study examines the impact of data snooping on neural networks used to detect vulnerabilities in lifted code, and builds on previous research that used word2vec and unidirectional and bidirectional transformer-based embeddings. The research specifically focuses on how model performance is affected when embedding models are trained with datasets, which include samples used for neural network training and validation. The results show that introducing data snooping did not significantly alter model performance, suggesting that data snooping had a minimal impact or that samples randomly dropped as part of the methodology contained hidden features critical to achieving optimal performance. In addition, …
From Data To Decisions: Safeguarding Athletes In The Age Of Ai, Nathan Elmer
From Data To Decisions: Safeguarding Athletes In The Age Of Ai, Nathan Elmer
SLU Law Journal Online
Artificial intelligence (AI) and data analytics are transforming professional sports by enhancing player performance, injury prevention, and scouting. However, the rapid adoption of AI raises significant concerns about data privacy, ownership, and decision-making biases that affect athletes. While collective bargaining agreements in major sports leagues provide some protections, they fail to address the complexities of AI-driven data collection and processing. The United States should adopt a regulatory framework similar to the European Union’s General Data Protection Regulation (GDPR) to safeguard athletes’ personal data. Implementing explicit consent requirements, addressing power imbalances, and ensuring transparency in AI decision-making would protect athletes while …
Development And Evaluation Of The Da Vinci Ai Tutor: Enhancing Accessibility And Personalized Learning In Art History Education, James Hutson, Tiffani Barner
Development And Evaluation Of The Da Vinci Ai Tutor: Enhancing Accessibility And Personalized Learning In Art History Education, James Hutson, Tiffani Barner
Faculty Scholarship
This study examines the implementation of the Da Vinci AI Tutor, an innovative artificial intelligence (AI)-based tutoring platform designed specifically for enhancing personalized and accessible learning in art history within higher education. Launched in Fall 2024 at a private liberal arts institution in the Midwest, the system integrates a conversational AI avatar modeled after Leonardo da Vinci, incorporating immersive virtual reality environments and multimodal interaction capabilities to engage students across undergraduate survey courses, advanced Renaissance classes, and graduate comprehensive exam preparations. Addressing significant gaps in existing humanities education research, the current study explores two primary research questions: (i) How AI-driven …
Neural Network-Based Low-Level 3d Point Cloud Processing, Pingping Cai
Neural Network-Based Low-Level 3d Point Cloud Processing, Pingping Cai
Theses and Dissertations
3D computer vision is a promising research field with the potential to revolutionize future lifestyles. Among various 3D representation formats, point clouds stand out for their efficiency in depicting 3D objects using a set of coordinates, enabling advancements in fields such as autonomous driving, virtual reality, and robotics. Due to the limitations of sensor fields of view and scanning trajectories, the collected point clouds are usually sparse, noisy, and incomplete, impeding the performance of many downstream applications. Thus, the tasks of low-level point cloud processing are proposed to refine and generate dense, clean, and complete point clouds. To accomplish these …
Augmenting Deep Learning For Efficient Nextg Wireless Communication And Sensing Systems, Hem Kanta Regmi
Augmenting Deep Learning For Efficient Nextg Wireless Communication And Sensing Systems, Hem Kanta Regmi
Theses and Dissertations
Wireless networks have become an integral aspect of our daily lives. Over the years, earlier generations of wireless networks have enabled some innovative applications, such as wireless gaming, fast internet browsing, and home automation, which were previously unattainable. However, to support emerging technologies like autonomous driving, virtual reality, telemedicine, and intelligent manufacturing, which require high data throughput and low latency, there is a need for advanced wireless networks. Legacy networks like WiFi/LTE, which operate below 6 GHz, have limited bandwidth and are insufficient to fulfill the high data throughput demands of several applications. Millimeter-wave (mmWave) networks, operating between 30 GHz …
Code Of Faith: Programming Spirituality In The Digital Age, Caleb Martin
Code Of Faith: Programming Spirituality In The Digital Age, Caleb Martin
Senior Honors Theses
With the increasing pervasiveness of technology in our daily lives, it is critical to consider the potential effects on an individual's religious practices and beliefs from software applications developed within a religious framework. This thesis delves into the relationship between software development and religious experiences, examining the manner in which the creation, operation, and application of religious technology can mold and impact an individual's spiritual development. This thesis aims to inform the implementation of technologies with a nonsecular application, ensuring that they are developed with a deep respect for the nuances of religious experience. The findings of this study will …
Cogprog: Utilizing Large Language Models To Forecast In-The-Moment Health Assessment, Gina Sprint, Maureen Schmitter-Edgecombe, Raven Weaver, Lisa Wiese, Diane Cook
Cogprog: Utilizing Large Language Models To Forecast In-The-Moment Health Assessment, Gina Sprint, Maureen Schmitter-Edgecombe, Raven Weaver, Lisa Wiese, Diane Cook
Computer Science Faculty Scholarship
Forecasting future health status is beneficial for understanding health patterns and providing anticipatory support for cognitive and physical health difficulties. In recent years, generative Large Language Models (LLMs) have shown promise as forecasters. Though not traditionally considered strong candidates for numeric tasks, LLMs demonstrate emerging abilities to address various forecasting problems. They also provide the ability to incorporate unstructured information and explain their reasoning process. In this article, we explore whether LLMs can effectively forecast future self-reported health state. To do this, we utilized in-the-moment assessments of mental sharpness, fatigue, and stress from multiple studies, utilizing daily responses (N = …
The Use Of Call Graphs And Deep Learning To Improve Software Testing, Ziad A. Al-Sharif, Hemanth G. Chintala, Safwan Omari
The Use Of Call Graphs And Deep Learning To Improve Software Testing, Ziad A. Al-Sharif, Hemanth G. Chintala, Safwan Omari
Engineering, Computing and Mathematical Sciences Faculty Conferences
Software testing is a critical part of software development, it is essential for preventing failures and enhancing software quality attributes. However, the testing process can be costly and time-consuming, often involving a large number of test cases. Over time, the accumulation of redundant and overlapping test cases can complicate and lengthen the testing time. To address these challenges, this paper utilizes graph similarity and deep learning techniques to optimize test suites. It uses call graphs from test cases to identify redundant and similar test cases. A machine learning model is used to calculate and predict the similarity scores between these …
Nothing Is Off Limits, But Not Everything Is Done Right: Exploring Ineffability, Alejandro Maciel
Nothing Is Off Limits, But Not Everything Is Done Right: Exploring Ineffability, Alejandro Maciel
ART 108: Introduction to Games Studies
Games are capable of addressing absolutely anything. No topic is too big, too controversial, or too emotional. but not everything is done well; the problem isn't the subject, it's always the delivery. When the developers approach heavy topics without care, they risk doing harm. bwhen they do their research and consult communities and write with purpose, the result can be transformative. Representation, when handled with attention, isn't political correctness; it's simply good storytelling. it's how games become more than entertainment. They become tools for empathy, education, and more valuable connections
Does Chatgpt-Permitted Assessments Help Students Generate Better Answers And Learn More?, Michelle L. F. Cheong, Yun-Chen Chen
Does Chatgpt-Permitted Assessments Help Students Generate Better Answers And Learn More?, Michelle L. F. Cheong, Yun-Chen Chen
Research Collection School Of Computing and Information Systems
We discuss our methodology and implementation of ChatGPT-permitted assessments for a university-level spreadsheets modelling module. Through our quantitative data analysis, our students rated ChatGPT’s answers to be incorrect on average and thus will not help them generate better answers directly, representing low “Perceived usefulness” (PU), while they rated ChatGPT 3.5 with relatively high “Perceived ease of use” (PE). They gave a good “Behavioural intention” (BI) rating indicating that they were motivated to use it in future as they could still learn more about this module by using ChatGPT 3.5. We found that both PU and PE affected BI positively, with …
Chatcrs: Incorporating External Knowledge And Goal Guidance For Llm-Based Conversational Recommender Systems, Chuang Li, Yang Deng, Hengchang Hu, Min-Yen Kan, Haizhou Li
Chatcrs: Incorporating External Knowledge And Goal Guidance For Llm-Based Conversational Recommender Systems, Chuang Li, Yang Deng, Hengchang Hu, Min-Yen Kan, Haizhou Li
Research Collection School Of Computing and Information Systems
This paper aims to efficiently enable large language models (LLMs) to use external knowledge and goal guidance in conversational recommender system (CRS) tasks. Advanced LLMs (e.g., ChatGPT) are limited in domain-specific CRS tasks for 1) generating grounded responses with recommendation-oriented knowledge, or 2) proactively leading the conversations through different dialogue goals. In this work, we first analyze those limitations through a comprehensive evaluation, showing the necessity of external knowledge and goal guidance which contribute significantly to the recommendation accuracy and language quality. In light of this finding, we propose a novel ChatCRS framework to decompose the complex CRS task into …
Comadice: Offline Cooperative Multi-Agent Reinforcement Learning With Stationary Distribution Shift Regularization, The Viet Bui, Tien Mai, Hong Thanh Nguyen
Comadice: Offline Cooperative Multi-Agent Reinforcement Learning With Stationary Distribution Shift Regularization, The Viet Bui, Tien Mai, Hong Thanh Nguyen
Research Collection School Of Computing and Information Systems
Offline reinforcement learning (RL) has garnered significant attention for its ability to learn effective policies from pre-collected datasets without the need for further environmental interactions. While promising results have been demonstrated in single-agent settings, offline multi-agent reinforcement learning (MARL) presents additional challenges due to the large joint state-action space and the complexity of multi-agent behaviors. A key issue in offline RL is the distributional shift, which arises when the target policy being optimized deviates from the behavior policy that generated the data. This problem is exacerbated in MARL due to the interdependence between agents' local policies and the expansive joint …
Agentstudio: A Toolkit For Building General Virtual Agents, Longtao Zheng, Zhiyuan Huang, Zhenghai Xue, Xinrun Wang, Bo An, Shuicheng Yan
Agentstudio: A Toolkit For Building General Virtual Agents, Longtao Zheng, Zhiyuan Huang, Zhenghai Xue, Xinrun Wang, Bo An, Shuicheng Yan
Research Collection School Of Computing and Information Systems
General virtual agents need to handle multimodal observations, master complex action spaces, and self-improve in dynamic, open-domain environments. However, existing environments are often domain-specific and require complex setups, which limits agent development and evaluation in real-world settings. As a result, current evaluations lack in-depth analyses that decompose fundamental agent capabilities. We introduce AgentStudio, a trinity of environments, tools, and benchmarks to address these issues. AgentStudio provides a lightweight, interactive environment with highly generic observation and action spaces, e.g., video observations and GUI/API actions. It integrates tools for creating online benchmark tasks, annotating GUI elements, and labeling actions in videos. Based …
Semantic Loss-Guided Data-Efficient Supervised Fine-Tuning For Safe Responses In Llms, Yuxiao Lu, Pradeep Varakantham, Pradeep Varakantham
Semantic Loss-Guided Data-Efficient Supervised Fine-Tuning For Safe Responses In Llms, Yuxiao Lu, Pradeep Varakantham, Pradeep Varakantham
Research Collection School Of Computing and Information Systems
Large Language Models (LLMs) generating unsafe responses to toxic prompts is a significant issue in their applications. While various efforts aim to address this safety concern, previous approaches often demand substantial human data collection or rely on the less dependable option of using another LLM to generate corrective data. In this paper, we aim to take this problem and overcome limitations of requiring significant high-quality human data. Our method requires only a small set of unsafe responses to toxic prompts, easily obtained from the unsafe LLM itself. By employing a semantic cost combined with a negative Earth Mover Distance (EMD) …
Prioritizing Speech Test Cases, Zhou Yang, Jieke Shi, Muhammad Hilmi Asyrofi, Bowen Xu, Xin Zhou, Donggyun Han, David Lo
Prioritizing Speech Test Cases, Zhou Yang, Jieke Shi, Muhammad Hilmi Asyrofi, Bowen Xu, Xin Zhou, Donggyun Han, David Lo
Research Collection School Of Computing and Information Systems
As Automated Speech Recognition (ASR) systems gain widespread acceptance, there is a pressing need to rigorously test and enhance their performance. Nonetheless, the process of collecting and executing speech test cases is typically both costly and time-consuming. This presents a compelling case for the strategic prioritization of speech test cases, which consist of a piece of audio and the corresponding reference text. The central question we address is: In what sequence should speech test cases be collected and executed to identify the maximum number of errors at the earliest stage? In this study, we introduce PRiOritizing sPeecH tEsT …