Digital Commons Network^™

End-To-End Task-Oriented Dialogue: A Survey Of Tasks, Methods, And Future Directions, Libo Qin, Wenbo Pan, Qiguang Chen, Lizi Liao, Zhou Yu, Yue Zhang, Wanxiang Che, Min Li

Research Collection School Of Computing and Information Systems

End-to-end task-oriented dialogue (EToD) can directly generate responses in an end-to-end fashion without modular training, which attracts escalating popularity. The advancement of deep neural networks, especially the successful use of large pre-trained models, has further led to significant progress in EToD research in recent years. In this paper, we present a thorough review and provide a unified perspective to summarize existing approaches as well as recent trends to advance the development of EToD research. The contributions of this paper can be summarized: (1) First survey: to our knowledge, we take the first step to present a thorough survey of this …

Generative Modelling Of Stochastic Actions With Arbitrary Constraints In Reinforcement Learning, Changyu Chen, Ramesha Karunasena, Thanh Hong Nguyen, Arunesh Sinha, Pradeep Varakantham

Research Collection School Of Computing and Information Systems

Many problems in Reinforcement Learning (RL) seek an optimal policy with large discrete multidimensional yet unordered action spaces; these include problems in randomized allocation of resources such as placements of multiple security resources and emergency response units, etc. A challenge in this setting is that the underlying action space is categorical (discrete and unordered) and large, for which existing RL methods do not perform well. Moreover, these problems require validity of the realized action (allocation); this validity constraint is often difficult to express compactly in a closed mathematical form. The allocation nature of the problem also prefers stochastic optimal policies, …

A Reliable And Secure Mobile Cyber-Physical Digital Microfluidic Biochip For Intelligent Healthcare, Yinan Yao, Decheng Qiu, Huangda Liu, Zhongliao Yang, Ximeng Liu, Yang Yang, Chen Dong

Research Collection School Of Computing and Information Systems

Digital microfluidic, as an emerging and potential technology, diversifies the biochemical applications platform, such as protein dilution sewage detection. At present, a vast majority of universal cyberphysical digital microfluidic biochips (DMFBs) transmit data through wires via personal computers and microcontrollers (like Arduino), consequently, susceptible to various security threats and with the popularity of wireless devices, losing competitiveness gradually. On the premise that security be ensured first and foremost, calls for wireless portable, safe, and economical DMFBs are imperative to expand their application fields, engage more users, and cater to the trend of future wireless communication. To this end, a new …

Comparison And Evaluation On Static Application Security Testing (Sast) Tools For Java, Kaixuan Li, Sen Chen, Lingling Fan, Ruitao Feng, Han Liu, Chengwei Liu, Yang Liu, Yixiang Chen

Research Collection School Of Computing and Information Systems

Static application security testing (SAST) takes a significant role in the software development life cycle (SDLC). However, it is challenging to comprehensively evaluate the effectiveness of SAST tools to determine which is the better one for detecting vulnerabilities. In this paper, based on well-defined criteria, we first selected seven free or open-source SAST tools from 161 existing tools for further evaluation. Owing to the synthetic and newly-constructed real-world benchmarks, we evaluated and compared these SAST tools from different and comprehensive perspectives such as effectiveness, consistency, and performance. While SAST tools perform well on synthetic benchmarks, our results indicate that only …

Scalelong: Towards More Stable Training Of Diffusion Model Via Scaling Network Long Skip Connection, Zhongzhan Huang, Pan Zhou, Shuicheng Yan, Liang Lin

Research Collection School Of Computing and Information Systems

In diffusion models, UNet is the most popular network backbone, since its long skip connects (LSCs) to connect distant network blocks can aggregate long-distant information and alleviate vanishing gradient. Unfortunately, UNet often suffers from unstable training in diffusion models which can be alleviated by scaling its LSC coefficients smaller. However, theoretical understandings of the instability of UNet in diffusion models and also the performance improvement of LSC scaling remain absent yet. To solve this issue, we theoretically show that the coefficients of LSCs in UNet have big effects on the stableness of the forward and backward propagation and robustness of …

Offline Rl With Discrete Proxy Representations For Generalizability In Pomdps, Pengjie Gu, Xinyu Cai, Dong Xing, Xinrun Wang, Mengchen Zhao, Bo An

Research Collection School Of Computing and Information Systems

Offline Reinforcement Learning (RL) has demonstrated promising results in various applications by learning policies from previously collected datasets, reducing the need for online exploration and interactions. However, real-world scenarios usually involve partial observability, which brings crucial challenges of the deployment of offline RL methods: i) the policy trained on data with full observability is not robust against the masked observations during execution, and ii) the information of which parts of observations are masked is usually unknown during training. In order to address these challenges, we present Offline RL with DiscrEte pRoxy representations (ORDER), a probabilistic framework which leverages novel state …

Beyond Factuality: A Comprehensive Evaluation Of Large Language Models As Knowledge Generators, Liang Chen, Yang Deng, Yatao Bian, Zeyu Qin, Bingzhe Wu, Tat-Seng Chua, Kam-Fai Wong

Research Collection School Of Computing and Information Systems

Large language models (LLMs) outperform information retrieval techniques for downstream knowledge-intensive tasks when being prompted to generate world knowledge. Yet, community concerns abound regarding the factuality and potential implications of using this uncensored knowledge. In light of this, we introduce CONNER, a COmpreheNsive kNowledge Evaluation fRamework, designed to systematically and automatically evaluate generated knowledge from six important perspectives - Factuality, Relevance, Coherence, Informativeness, Helpfulness and Validity. We conduct an extensive empirical analysis of the generated knowledge from three different types of LLMs on two widely-studied knowledge-intensive tasks, i.e., open-domain question answering and knowledge-grounded dialogue. Surprisingly, our study reveals that the …

Cue-Cot: Chain-Of-Thought Prompting For Responding To In-Depth Dialogue Questions With Llms, Hongru Wang, Rui Wang, Fei Mi, Yang Deng, Zezhong Wang, Bin Liang, Ruifeng Xu, Kam-Fai Wong

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs), such as ChatGPT, greatly empower dialogue systems with strong language understanding and generation capabilities. However, most of the previous works prompt the LLMs to directly generate a response based on the dialogue context, overlooking the underlying linguistic cues about the user status exhibited in the context. Such in-depth dialogue scenarios are challenging for existing LLMs to figure out the user’s hidden needs and respond satisfactorily through a single-step inference. To this end, we propose a novel linguistic cue-based chain-of-thoughts (Cue-CoT), which enhances the LLMs inference with an intermediate reasoning step to find cues exhibited in the …

Controlling Type Confounding In Ad Hoc Teamwork With Instance-Wise Teammate Feedback Rectification, Dong Xing, Pengjie Gu, Qian Zheng, Xinrun Wang, Shanqi Liu, Longtao Zheng, Bo An, Gang Pan

Research Collection School Of Computing and Information Systems

Ad hoc teamwork requires an agent to cooperate with unknown teammates without prior coordination. Many works propose to abstract teammate instances into high-level representation of types and then pre-train the best response for each type. However, most of them do not consider the distribution of teammate instances within a type. This could expose the agent to the hidden risk of type confounding. In the worst case, the best response for an abstract teammate type could be the worst response for all specific instances of that type. This work addresses the issue from the lens of causal inference. We first theoretically …

Post-War Restitution In The Republic Of Croatia: Awareness Of The Komza And Reconciliation Of Moveable Property Through Public Sculpture, Anita Govic

Theses

This project is a public sculpture prototype focused on the need for post-war art restitution in the Republic of Croatia: it will have an impact by enhancing public knowledge of the KOMZA list, which offers legal proof for families to retrieve heirlooms taken during WWII and the wars of secession from Yugoslavia.

The sculpture design includes three pillars, two etched with the 1529 names from the KOMZA list and a third that will be progressively built, brick-by-brick, as artworks are restituted. The nature of viewers’ interaction includes access to QR codes that offer links to the KOMZA list, information about …

Customer Churn Prediction Using Composite Deep Learning Technique, Asad Khattak, Zartashia Mehak, Hussain Ahmad, Muhammad Usama Asghar, Muhammad Zubair Asghar, Aurangzeb Khan

All Works

Customer churn, a phenomenon that causes large financial losses when customers leave a business, makes it difficult for modern organizations to retain customers. When dissatisfied customers find their present company's services inadequate, they frequently migrate to another service provider. Machine learning and deep learning (ML/DL) approaches have already been used to successfully identify customer churn. In some circumstances, however, ML/DL-based algorithms lacks in delivering promising results for detecting client churn. Previous research on estimating customer churn revealed unexpected forecasts when utilizing machine learning classifiers and traditional feature encoding methodologies. Deep neural networks were also used in these efforts to extract …

Use Of Ai To Recreate And Repatriate Lost, Destroyed Or Stolen Paintings: The 1785 Parisian Salon Case Study, Charles E. O'Brien, James Hutson, Trent Olsen, Jay Ratican

Faculty Scholarship

This study investigates the efficacy of artificial intelligence (AI) in the field of artwork restoration, focusing on lost, stolen, or destroyed artworks. Employing a dual approach that combines traditional manual restoration techniques with advanced generative AI tools, the research centers on a case study of the 1785 Parisian Salon. It specifically examines the reconstitution of Antoine François Callet's painting, Achilles Dragging the Body of Hector, unveiled alongside Jacques-Louis David's Oath of the Horatii. The study utilizes Easy Diffusion and Stable Diffusion 2.1 technologies for inpainting and colorization processes. These AI tools are employed in concert with manual restoration practices to …

Demystifying Artificial Intelligence (Ai) For Early Childhood And Elementary Education: A Case Study Of Perceptions Of Ai Of State Of Missouri Educators, Kathryn Arnone, James Hutson, Karen Woodruff

Faculty Scholarship

Artificial intelligence (AI) and its impact on society have received a great deal of attention in the past five years since the first Stanford AI100 report. AI already globally impacts individuals in critical and personal ways, and many industries will continue to experience disruptions as the full algorithmic effects are understood. However, with regard to education, adopting in disciplines remains limited largely to Computer Science and Information Technology in postsecondary education. Recent advances with technology are especially promising for their potential to create and scale personalized learning for students, to optimize strategies for learning outcomes, and to increase access to …

Essence As Algorithm: Public Perceptions Of Ai-Powered Avatars Of Real People, James Hutson, Jay Ratican, Colleen Biri

Faculty Scholarship

This paper investigates the intersection of generative AI, Large Language Models (LLM), and robotics. Exemplified by systems like ChatGPT and technological marvels such as Ameca the Robot, the combination of technologies will allow humans to transcend the limitations of death. Through digital necromancy, a practice encompassing the technological resurrection of deceased individuals, the ability to not only passively see recordings of loved ones but to interact with them is made possible, leading to ethical and psychological considerations. Therefore, examining these trends extends into the motives underlying engagement with both incorporeal and corporeal reproductions of individuals, with reasons ranging from memory …

Experimental And Ansys Investigation On The Aerogel Envelopes For Energy-Efficient Building, Syed Mohammed Tareq

Masters Theses and Doctoral Dissertations

Energy consumption, specifically in the building sector is always increasing. One potential way to reduce energy consumption, is to reduce the heat loss in buildings or homes. Silica aerogels have grown in popularity as an insulating material due to their extremely low thermal conductivity. Very often the benefits of transparency as well as strength is ignored due to its brittleness. An effort has been taken to increase the strength using vacuum seal while maintaining their transparency. To understand the benefits of using silica aerogels as a replacement for glass window or wall in buildings, experimental validation was carried out in …

Temperature Gradient Effect On Solid-Liqid Interface Properties Of Al-Cu Alloy: A Molecular Dynamics Study, Prashant Kumar Jha

All Theses

Aluminum-copper (Al-Cu) alloys are widely used in the aerospace industry due to their favorable strength-to-weight ratio, good fatigue resistance, and corrosion resistance. These properties make Al-Cu alloys an excellent choice for aircraft structural components that require high strength and low weight. Additive manufacturing (AM), also known as 3D printing, has emerged as a promising processing method for Al-Cu alloys in aerospace manufacturing. AM enables the production of lightweight optimized geometries difficult to manufacture through conventional subtractive methods. AM also reduces material waste by only depositing material where needed in the part geometry. The rapid solidification conditions in AM processes motivate …

Material World, Nicole Weldy

All Theses

My thesis, Material World, delves into the use of the crochet unit as a construction technique for building forms. Through this, I aim to organize different materials in a way that responds to the challenges posed by the physical world. My artistic process is centered around honoring the inherent qualities of thread and uses these qualities to create form, lightness, and linearity. At the same time, I remain receptive to transformative processes such as combining three-dimensional (3D) printed lines and lace stiffener to push the boundaries of what thread can do. By combining manual craftsmanship with technology materialized as …

Dynamic Translation Of Drone View To Vehicle View For Autonomous Driving, Grayson Byrd

All Theses

As the field of computer vision continues to advance, the use of autonomous vehicles in military applications has increased and the tasks associated with these systems have grown in scope and complexity. These vehicles tend to operate in combat situations, presenting significant risk in the form of sensor damage. Since these autonomy algorithms rely on sensor information to navigate their environment, any threat to sensor functionality negatively impacts the reliability of the system. As a potential solution, I propose Dynamic Diffusion-based View Translation (DDVT), a novel computer vision algorithm capable of restoring image sensor function in ground vehicles through aerial …

Development Of A User-Friendly Shelf-Life Model To Evaluate The Suitability Of Sustainable Materials In Roasted And Ground Coffee Fractional Packs, Matthew Baxley

All Theses

Roasted and ground coffee is a shelf stable product yet quite sensitive to oxidative staling. A consumer acceptance-based shelf-life modeling system was proposed with intent for the rapid determination of suitable coffee packages. This model requires as input the oxygen consumption rate (OCR) of the coffee, barrier values of packages, and the size of the packaging. Within the time period tested, it was shown that this model accurately predicted the oxygen uptake of coffee over time. Four bio-based packaging systems with barrier layers including mPLA, mPE, mcellophane, and paper were compared against a control (mPET). These materials displayed a range …

Phenology Of Asian Longhorned Beetle (Anoplophora Glabripennis, Coleoptera: Cerambycidae) In South Carolina, U.S.A., Lena Schmitt

All Theses

A new population of the Asian longhorned beetle (Anoplophora glabripennis Motschulsky, hereafter ALB), an invasive species in the United States since 1996, was discovered in Charleston County, South Carolina in 2020. This population is the furthest south ALB has been found in North America and provides an opportunity to examine the phenology of this beetle in a novel climate. I collected 1009 eggs, larvae, pupae, and adult beetles from the federal quarantine zone in South Carolina and used larval head capsule widths to determine the rate of development and voltinism of ALB in South Carolina’s coastal climate. With this …

Use Of Graphene Oxide Nanoplatelets As A Barrier Coating For Food Grade Packaging, Michaela Gibson

All Theses

The development and evaluation of graphene oxide (GO) coatings as a replacement for the aluminum foil layer in retort grade pouches is an important area of interest as we look to send humans to Mars. The packaging needs to not only restrict the permeation of oxygen and water vapor molecules, but also needs to be capable of thermally degrading. Compared to aluminum foil, which is responsible for the hermetic barrier properties of retort grade pouches, a GO coating is a more flexible barrier material and allows for ash-free incineration. Water-based and ethanol-based GO coatings were applied to polyethylene terephthalate (PET) …

Generalized Vulnerability Measures Of Graphs, Julia Vanlandingham

All Theses

Several measures of vulnerability of a graph look at how easy it is to disrupt the network by removing/disabling vertices. As graph-theoretical parameters, they treat all vertices alike: each vertex is equally important. For example, the integrity parameter considers the number of vertices removed and the maximum number of vertices in a component that remains. We consider the generalization of these measures of vulnerability to weighted vertices in order to better model real-world applications. In particular, we investigate bounds on the weighted versions of connectivity and integrity, when polynomial algorithms for computation exist, and other characteristics of the generalized measures.

Rapid Reaction Studies On The Chemistry Of Flavin Oxidation In Urocanate Reductase, Niusha Delavari

Masters Theses

Urocanate reductase (UrdA) is a bacterial flavin-dependent enzyme that reduces urocanate to imidazole propionate, enabling bacteria to use urocanate as an alternative respiratory electron acceptor. Elevated serum levels of imidazole propionate are associated with the development of type 2 diabetes and, since UrdA is only present in humans in gut bacteria, this enzyme has emerged as a significant factor linking the health of the gut microbiome and the metabolic disorder insulin resistance. Here we investigated the chemistry of flavin oxidation by urocanate in the isolated FAD domain of UrdA (UrdA’) using anaerobic stopped-flow experiments at 4 °C. This analysis unveiled …

Analyzing Functional Interactions Of Designed Peptides By Nmr Spectroscopy, Wonsuk Choi

Pharmaceutical Sciences (MS) Theses

The development of small peptide-based therapeutics can be accelerated by the knowledge of relationships between the peptide structure and its functional interactions. Here, we report the analysis of two groups of synthetic peptides designed for two applications – broad bactericidal action and inhibition of protein-protein interactions in human cells. Novel amphiphilic peptides designed for antibacterial application incorporated arginine as cationic amino acids and non-natural amino acids that have aromatic side chains with similar hydrophobic properties as tryptophan. The interaction of lead cyclic peptides and their linear analogs with a phospholipid bilayer mimicking a bacterial membrane was studied using nuclear magnetic …

Type 2 Diabetes Modifies The Association Of Cad Genomic Risk Variants With Subclinical Atherosclerosis, Natalie R Hasbani, Kenneth E Westerman, Soo Heon Kwak, Han Chen, Xihao Li, Daniel Di Corpo, Jennifer Wessel, Joshua C Bis, Chloè Sarnowski, Peitao Wu, Lawrence F Bielak, Xiuqing Guo, Nancy Heard-Costa, Gregory L Kinney, Michael C Mahaney, May E Montasser, Nicholette D Palmer, Laura M Raffield, James G Terry, Lisa R Yanek, Jessica Bon, Donald W Bowden, Jennifer A Brody, Ravindranath Duggirala, David R Jacobs, Rita R Kalyani, Leslie A Lange, Braxton D Mitchell, Jennifer A Smith, Kent D Taylor, April P Carson, Joanne E Curran, Myriam Fornage, Barry I Freedman, Stacey Gabriel, Richard A Gibbs, Namrata Gupta, Sharon L R Kardia, Brian G Kral, Zeineen Momin, Anne B Newman, Wendy S Post, Karine A Viaud-Martinez, Kendra A Young, Lewis C Becker, Alain G Bertoni, John Blangero, John J Carr, Katherine Pratte, Bruce M Psaty, Stephen S Rich, Joseph C Wu, Rajeev Malhotra, Patricia A Peyser, Alanna C Morrison, Ramachandran S Vasan, Xihong Lin, Jerome I Rotter, James B Meigs, Alisa K Manning, Paul S De Vries

Faculty, Staff and Student Publications

BACKGROUND: Individuals with type 2 diabetes (T2D) have an increased risk of coronary artery disease (CAD), but questions remain about the underlying pathology. Identifying which CAD loci are modified by T2D in the development of subclinical atherosclerosis (coronary artery calcification [CAC], carotid intima-media thickness, or carotid plaque) may improve our understanding of the mechanisms leading to the increased CAD in T2D.

METHODS: We compared the common and rare variant associations of known CAD loci from the literature on CAC, carotid intima-media thickness, and carotid plaque in up to 29 670 participants, including up to 24 157 normoglycemic controls and 5513 …

Exploring Non-Linear Dynamical Structure For Knee Kinematics Using Machine Learning, Liora Mayats-Alpay, Rahul Soangra

Physical Therapy Faculty Articles and Research

Human movement involves complex coordination between multiple limbs during execution. Human gait is cyclic, and the knee's movement inherently follows nonlinear dynamic behavior that linear models cannot adequately capture. In this study, advanced Machine Learning (ML) techniques were employed to combine the Sparse Identification of Nonlinear Dynamics (SINDy) algorithm using Python to reveal governing equations of knee movement during walking. We gathered a single subject's knee motion data using infrared markers during normal walking. We utilized the PySINDy library to determine the governing equations and calculated the coefficient of dynamical systems associated with knee kinematics. Our results emphasize governing equations …

The Economics Of Financial Scams: Evidence From Initial Coin Offerings, Kenny Phua, Bo Sang, Chi Shen Wei, Gloria Yang Yu

Research Collection Lee Kong Chian School Of Business

We examine the economics of financial scams by analyzing the market for initial coin offerings (ICOs). Using data snapshots of 5,873 ICOs, we find that irregularities in ICO characteristics across listing websites predict higher scam risk and are likely intentional. These patterns are consistent with a model where malicious issuers maximize profits by using irregularities to screen for naïve investors. Almost half of the ICOs in our sample may be scams, amounting to more than U.S. $6 billion in losses. Our results draw attention to the frequent use of screening mechanisms in financial scams.

Discovery Of Novel 2-Aminopyridine Derivatives As Ros1 And Alk Dual Inhibitors To Combat Drug-Resistant Mutants Including Ros1g2032r And Alkg1202r, Siming Liu, Chuan Huang, Chunhui Huang, Yaqi Huang, Yonghuan Yu, Guowu Wu, Fengqiu Guo, Ying Jiang, Shanhe Wan, Zhengguang Zhu, Yuanxin Tian, Jianghua Zhu, Jiajie Zhang

Faculty, Staff and Student Publications

Clinical treatment by FDA-approved ROS1/ALK inhibitor Crizotinib significantly improved the therapeutic outcomes. However, the emergence of drug resistance, especially driven by acquired mutations, have become an inevitable problem and worsened the clinical effects of Crizotinib. To combat drug resistance, some novel 2-aminopyridine derivatives were designed rationally based on molecular simulation, then synthesised and subjected to biological test. The preferred spiro derivative C01 exhibited remarkable activity against CD74-ROS1G2032R cell with an IC50 value of 42.3 nM, which was about 30-fold more potent than Crizotinib. Moreover, C01 also potently inhibited enzymatic activity against clinically Crizotinib-resistant ALKG1202R, harbouring a 10-fold potency superior to …

Wallaby Pre-Pilot Survey: Ultra-Diffuse Galaxies In The Eridanus Supergroup, B.-Q. For, K. Spekkens, L. Staveley-Smith, K. Bekki, A. Karunakaran, B. Catinella, Bärbel S. Koribalski, K. Lee-Waddell, Juan P. Madrid, C. Murugeshan

Physics & Astronomy Faculty Publications

We present a pilot study of the atomic neutral hydrogen gas (H I) content of ultra-diffuse galaxy (UDG) candidates. In this paper, we use the pre-pilot Eridanus field data from the Widefield ASKAP L-band Legacy All-sky Blind Survey to search for H I in UDG candidates found in the Systematically Measuring Ultra-diffuse Galaxies survey (SMUDGes). We narrow down to 78 SMUDGes UDG candidates within the maximum radial extents of the Eridanus subgroups for this study. Most SMUDGes UDGs candidates in this study have effective radii smaller than 1.5 kpc and thus fail to meet the defining size threshold. We …

Type 2 Diabetes Modifies The Association Of Cad Genomic Risk Variants With Subclinical Atherosclerosis, Natalie. R. Hasbani, Kenneth E. Westerman, Soo Heon Kwak, Han Chen, Xihao Li, Daniel Di Corpio, Michael C. Mahaney, Ravi Duggirala, Joanne E. Curran, John Blangero

School of Medicine Publications

Background: Individuals with type 2 diabetes (T2D) have an increased risk of coronary artery disease (CAD), but questions remain about the underlying pathology. Identifying which CAD loci are modified by T2D in the development of subclinical atherosclerosis (coronary artery calcification [CAC], carotid intima-media thickness, or carotid plaque) may improve our understanding of the mechanisms leading to the increased CAD in T2D.

Methods: We compared the common and rare variant associations of known CAD loci from the literature on CAC, carotid intima-media thickness, and carotid plaque in up to 29 670 participants, including up to 24 157 normoglycemic controls and 5513 …