Digital Commons Network^™

Scaling Up Cooperative Multi-Agent Reinforcement Learning Systems, Minghong Geng

Research Collection School Of Computing and Information Systems

Cooperative multi-agent reinforcement learning methods aim to learn effective collaborative behaviours of multiple agents performing complex tasks. However, existing MARL methods are commonly proposed for fairly small-scale multi-agent benchmark problems, wherein both the number of agents and the length of the time horizons are typically restricted. My initial work investigates hierarchical controls of multi-agent systems, where a unified overarching framework coordinates multiple smaller multi-agent subsystems, tackling complex, long-horizon tasks that involve multiple objectives. Addressing another critical need in the field, my research introduces a comprehensive benchmark for evaluating MARL methods in long-horizon, multi-agent, and multi-objective scenarios. This benchmark aims to …

Explaining Sequences Of Actions In Multi-Agent Deep Reinforcement Learning Models, Phyo Wai Khaing, Minghong Geng, Shubham Pateria, Budhitama Subagdja, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

This paper introduces a method to explain MADRL agents’ behaviors by abstracting their actions into high-level strategies. Particularly, a spatio-temporal neural network model is applied to encode the agents’ sequences of actions as memory episodes wherein an aggregating memory retrieval can generalize them into a concise abstract representation of collective strategies. To assess the effectiveness of our method, we applied it to explain the actions of QMIX MADRL agents playing a StarCraft Multi-agent Challenge (SMAC) video game. A user study on the perceived explainability of the extracted strategies indicates that our method can provide comprehensible explanations at various levels of …

Benchmarking Marl On Long Horizon Sequential Multi-Objective Tasks, Minghong Geng, Shubham Pateria, Budhitama Subagdja, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

Current MARL benchmarks fall short in simulating realistic scenarios, particularly those involving long action sequences with sequential tasks and multiple conflicting objectives. Addressing this gap, we introduce Multi-Objective SMAC (MOSMAC), a novel MARL benchmark tailored to assess MARL methods on tasks with varying time horizons and multiple objectives. Each MOSMAC task contains one or multiple sequential subtasks. Agents are required to simultaneously balance between two objectives - combat and navigation - to successfully complete each subtask. Our evaluation of nine state-of-the-art MARL algorithms reveals that MOSMAC presents substantial challenges to many state-of-the-art MARL methods and effectively fills a critical gap …

Learning Nighttime Semantic Segmentation The Hard Way, Wenxi Liu, Jiaxin Cai, Qi Li, Chenyang Liao, Jingjing Cao, Shengfeng He, Yuanlong Yu

Research Collection School Of Computing and Information Systems

Nighttime semantic segmentation is an important but challenging research problem for autonomous driving. The major challenges lie in the small objects or regions from the under-/over-exposed areas or suffer from motion blur caused by the camera deployed on moving vehicles. To resolve this, we propose a novel hard- class-aware module that bridges the main network for full-class segmentation and the hard-class network for segmenting aforementioned hard-class objects. In specific, it exploits the shared focus of hard-class objects from the dual-stream network, enabling the contextual information flow to guide the model to concentrate on the pixels that are hard to classify. …

Cornac-Ab : An Open-Source Recommendation Framework With Native A/B Testing Integration, Rong Sheng Ong, Quoc Tuan Truong, Hady Wirawan Lauw

Research Collection School Of Computing and Information Systems

Recommender systems significantly impact user experience across diverse domains, yet existing frameworks often prioritize offline evaluation metrics, neglecting the crucial integration of A/B testing for forward-looking assessments. In response, this paper introduces a new framework seamlessly incorporating A/B testing into the Cornac recommendation library. Leveraging a diverse collection of model implementations in Cornac, our framework enables effortless A/B testing experiment setup from offline trained models. We introduce a carefully designed dashboard and a robust backend for efficient logging and analysis of user feedback. This not only streamlines the A/B testing process but also enhances the evaluation of recommendation models in …

Term Importance For Transformer-Based Qa Retrieval : A Case Study Of Stackexchange, Bryan Zhi Yang Tan, Hady W. Lauw

Research Collection School Of Computing and Information Systems

Question-answering (QA) retrieval is the task of retrieving the most relevant answer to a given question from a collection of answers. Various approaches to QA retrieval have been developed recently. One successful and popular model is Contextualized Late Interaction over BERT (ColBERT), a transformer-based approach that adopts a query-document scoring mechanism that retains the granularity of transformer matching, whilst improving on efficiency. However, one key limitation is that it requires further fine-tuning for new query or collection types. In this work, we explore and propose several non-parametric retrieval augmentation methods based on explicit signals of term importance that improve over …

Baffle : Hiding Backdoors In Offline Reinforcement Learning Datasets, Chen Gong, Zhou Yang, Yunpeng Bai, Junda He, Jieke Shi, Kecen Li, Arunesh Sinha, Bowen Xu, Xinwen Hou, David Lo, Tianhao Wang

Research Collection School Of Computing and Information Systems

Reinforcement learning (RL) makes an agent learn from trial-and-error experiences gathered during the interaction with the environment. Recently, offline RL has become a popular RL paradigm because it saves the interactions with environments. In offline RL, data providers share large pre-collected datasets, and others can train high-quality agents without interacting with the environments. This paradigm has demonstrated effectiveness in critical tasks like robot control, autonomous driving, etc. However, less attention is paid to investigating the security threats to the offline RL system. This paper focuses on backdoor attacks, where some perturbations are added to the data (observations) such that given …

Rule-Guided Counterfactual Explainable Recommendation, Yinwei Wei, Xiaoyang Qu, Xiang Wang, Yunshan Ma, Liqiang Nie, Tat‑Seng Chua

Research Collection School Of Computing and Information Systems

To empower the trust of current recommender systems, the counterfactual explanation (CE) method is adopted to generate the counterfactual instance for each input and take their changes causing the different outcomes as the explanation. Although promising results have been achieved by existing CE-based methods, we propose to generate the attribute-oriented counterfactual explanation. Different from them, we aim to generate the counterfactual instance by performing the intervention on the attributes, and then build an attribute-oriented counterfactual explainable recommender system. Considering the correlation and categorical values of attributes, how to efficiently generate the reliable counterfactual instances on the attributes challenges us. To …

Learning To Generate Explainable Stock Predictions Using Self‑Reflective Large Language Models, Kelvin J.L. Koa, Yunshan Ma, Ritchie Ng, Tat‑Seng Chua

Research Collection School Of Computing and Information Systems

Explaining stock predictions is generally a difficult task for traditional non-generative deep learning models, where explanations are limited to visualizing the attention weights on important texts. Today, Large Language Models (LLMs) present a solution to this problem, given their known capabilities to generate human-readable explanations for their decision-making process. However, the task of stock prediction remains challenging for LLMs, as it requires the ability to weigh the varying impacts of chaotic social texts on stock prices. The problem gets progressively harder with the introduction of the explanation component, which requires LLMs to explain verbally why certain factors are more important …

Do Emotion Dysregulation And Internalized Shame Mediate The Relationship Between Borderline Personality Disorder Symptomology And Criminal Involvement?, Madison Smith

Electronic Theses and Dissertations

Borderline personality disorder (BPD) is a severe and debilitating mental disorder that is over-represented within the criminal legal system (CLS), with 15-29.5% of individuals in the CLS meeting criteria. However, little research has examined mechanisms leading to criminal involvement within this population. Potential mechanisms include positive/negative emotion dysregulation and internalized shame, which (aside from positive emotion dysregulation) have been linked to criminal involvement in other populations. Research has yet to examine these constructs as mechanisms explaining criminal involvement among individuals with BPD. This study examined positive/negative emotion dysregulation and internalized shame as mechanisms of the relationship between BPD and criminal …

Comparative Analysis Of Surrogate Models For The Dissolution Of Spent Nuclear Fuel, Dayo Awe

Electronic Theses and Dissertations

This thesis presents a comparative analysis of surrogate models for the dissolution of spent nuclear fuel, with a focus on the use of deep learning techniques. The study explores the accuracy and efficiency of different machine learning methods in predicting the dissolution behavior of nuclear waste, and compares them to traditional modeling approaches. The results show that deep learning models can achieve high accuracy in predicting the dissolution rate, while also being computationally efficient. The study also discusses the potential applications of surrogate modeling in the field of nuclear waste management, including the optimization of waste disposal strategies and the …

Reviving “Cult”: A Qualitative Analysis Of A Female-Led Cultic Sect, Olivia Summers

Electronic Theses and Dissertations

This research analyzes the Word of Faith Fellowship (WOF), a female-led Evangelical church that social scientists have not yet studied. My thesis explores how a female leader operates within a patriarchal space and why WOF owns a Holocaust Museum. I conducted content analysis of the church and museum websites, Google reviews, and visited the museum in person. My research highlights the limitations of current taxonomies of religion in sociology. I address this oversight, argue for the re-introduction of “cult” as an analytical term, and propose a rubric for cult identification. I suggest that WOF is a sectarian cult with similar …

Community, Race, And National Socialism: The Evolution Of The Ideology Of Volksgemeinschaft, 1807-1945, Robert B. Anderson

Electronic Theses and Dissertations

Historiography of the National Socialist Volksgemeinschaft, or people’s community, has traditionally been divided between historians surmising its construction under the Third Reich as a genuine undertaking meant to uplift German society, and those who view the project as a propaganda effort which assisted the Nazi Party (NSDAP) in retaining legitimacy. Utilizing the plethora of works written on the topic, and a handful of primary sources from pre-Nazified Germany, NSDAP officials, and average citizens alike, this work will demonstrate that, as early as 1807, German philosophers, statesmen, and eventually a large majority of the population yearned for the national unity …

Does My Voice Matter? Understanding The Lived Experiences And Turnover Intentions Of Professional Academic Advisors, Keith Kanard Douglas

Electronic Theses and Dissertations

Academic advising has been a vital aspect of the higher education setting for at least a century. Academic advising has evolved from the vocational development of young men to a full-time profession, especially after the creation of the National Academic Advising Association in 1979. Although the academic profession has developed through research, master’s programs, and courses designed to expand the field of academic advising, researchers still question whether full-time academic advising is a profession. Questions regarding the legitimacy of the work of academic advisors have led to miscommunication between upper administrators within the university. Universities have experienced a trend of …

Disability Services As Related To Higher Education Leadership: An Exploratory Study, Tara Wade Franklin

Electronic Theses and Dissertations

Over the past three decades, the number of students with disabilities attending postsecondary institutions has increased, yet their completion rates remain lower than those of students without disabilities. This mixed methods exploratory study investigates the role of provost leadership in supporting disability services and understanding the needs of students with disabilities in higher education. The study focuses on Texas higher education institutions, examining the relationships between provosts and directors of disability services. The researcher collected data through esurveys and open-ended questions of 16 participants, primarily directors of disability services. Quantitative analysis of Likert scale items explored domains such as leadership, …

A Comparison Of The Whytry Curriculum And Aggression Replacement Training In Adolescents At Risk For Family Breakdown, Grace Elizabeth Taylor

Electronic Theses and Dissertations

The Prevention and Early Intervention (PEI) division of the Texas Department of Family and Protective Services seeks to provide evidence-based programs to prevent family breakdown. Evidence-based programs are defined as programs that have qualitative and quantitative support of positive outcomes by peer-reviewed articles. Family and Youth Success (FAYS) is a PEI program that seeks to promote resilience in families and youth by offering one-on-one skills-based counseling and group-based learning for children and parents. FAYS utilizes the WhyTry Program and Aggression Replacement Training (ART) as social skills training programs that seek to decrease social problems and aggression and increase social skills. …

Standing Posterior Trunk Rotation With Adaptive Proprioceptive Toss, Jennifer Elvir

Electronic Theses and Dissertations

Parkinson’s disease (PD) is a remarkably individualized disease. Nearly every person’s trajectory is unique. One person may experience tremor at an early stage of the disease, while another may experience tremor at a later stage or even not at all. For all PD patients, for certain can expect increasing difficulty as the disease progresses. Parkinson’s disease is responsible for the loss of dopamine which displays itself most notably through motor and cognitive symptom disruption. Although Parkinson’s is a chronic and progressive condition that is irremediable, significant strides have been facilitated to help control and manage the trajectory of the disease. …

Ein Bericht Aus Einer Akademie: Writing About Germany's Lost Jewish Scholars, 80 Years On, J.G. Allen

Research Collection Yong Pung How School Of Law

This volume explores the life and work of F.A. Mann—a German-born and (mostly) German-educated jurist who spent the operative parts of his professional life in the United Kingdom (UK). The project on which the volume is based set out to understand Mann’s life as a lawyer and describe his legacy—in English law, and also in German, European, and international law.

How Gender Affirming Care Affects The Current Sex Estimation Standards In Forensic Anthropology: A Preliminary Study, Dakota Taylor

Department of Anthropology: Theses and Student Research

Current sex estimation standards in forensic anthropology are based on individuals whose gender matches their biological/osteological sex, also known as Cisgendered individuals. Recently, transgender individuals have started to become more common in the forensic context due to the increase in hate crimes and violence. This research builds upon past research done on how facial feminization surgery can affect both visual and metric methods, where it was found that forensic anthropologists should rely on the visual methods if they suspect someone to be transgender due to it being more accurate and being able to clearly state the scars left on the …

Numerical Variations In The Thoracic And Lumbar Vertebrae Within The John A. Williams Skeletal Collection, Western Carolina University, Cullowhee, Nc (Usa), Leanna Annette Sanford

Department of Anthropology: Theses and Student Research

This research is on how human variation can lead to the identification of remains based on skeletal variation. The data were collected by performing a morphoscopic trait study of the John A. Williams (J.A.W.) Documented Human Skeletal Collection at Western Carolina University in Cullowhee, NC (USA). Morphoscopic traits are nonmetric traits, visually identified using the knowledge of osteology. The study was performed to study variation of the vertebral column, specifically focusing on morphoscopic traits of the thoracic and lumbar vertebrae. The focus of this research is centralized on the presence of numerical variations in the vertebral column such as eleven …

The Osteobiography Of Human Remains From The Seaview And Indian Town Trail Archaeological Sites, Maggie M. Klemm

Department of Anthropology: Theses and Student Research

Extensive site surveys and excavations on the Island of Barbuda led by Sophia Perdikaris have identified over 62 sites spanning from the Archaic time period to Historic times. Over the last 18 years, these multidisciplinary teams have focused on mapping all sites and performing rescue excavations on sites threatened by sea level rise, erosion or development. Two such sites are the Saladoid site of Seaview (BA016) and the Troumassoid site of Indian Town Trail (BA01). The dunes surrounding the site of Seaview receive the brunt of storms and hurricanes. In 1998 hurricane Georges exposed skeletal material now part of the …

The Kruger Collection Reimagined: A Case Study In 3d Scanning And Interactive Exhibit Design, Annissa Davis

Department of Anthropology: Theses and Student Research

This thesis examines the use of 3D modeling in museum exhibition to create exploratory exhibits that facilitate unique relationships between the visitors and the collection beyond what is provided by the collection’s in person counterparts. Typical use of 3D modeling in museums is currently often representative rather than exploratory. By employing a Digital Humanities lens to approach the development of a digital exhibition utilizing 3D technology and interactive elements created in a video game engine (Unity), this thesis project evaluates these potential new relationships. Using the Eloise Kruger Collection of Miniatures as a case study, the following text details the …

A Manual On The Planning And Integration Of Archaeological Methods In To Outdoor Forensic Search Investigations, Jonathan Robert Garcia

Department of Anthropology: Theses and Student Research

As archaeology and its applications into forensic contexts develop into the growing discipline of forensic archaeology, an increasing amount of literature has resulted stemming from research on the integration of common archaeological methods. However, much of this literature is intended for professional archaeologists or forensic anthropologists who are well experienced in their respective disciplines. Emerging literature generally does not consider those who leads efforts at forensic scenes in the outdoors such as law enforcement officers who often lack a background in archaeology or forensic anthropology. This thesis seeks to resolve this dilemma by creating a new and accessible manual. The …

Error Estimates For A Mixed Finite Element Method For The Maxwell's Transmission Eigenvalue Problem, Chao Wang, Jintao Cui, Jiguang Sun

Michigan Tech Publications

In this paper, we analyze a numerical method combining the Ciarlet-Raviart mixed finite element formulation and an iterative algorithm for the Maxwell's transmission eigenvalue problem. The eigenvalue problem is first written as a nonlinear quad-curl eigenvalue problem. Then the real transmission eigenvalues are proved to be the roots of a non-linear function. They are the generalized eigenvalues of a related linear self-adjoint quad-curl eigenvalue problem. These generalized eigenvalues are computed by a mixed finite element method. We derive the error estimates using the spectral approximation of compact operators, the theory of mixed finite element method for quad-curl problems, and the …

Remnant Polarization And Structural Arrangement In P(Vdf-Trfe) Electrospun Fiber Meshes Affect Osteogenic Differentiation Of Human Mesenchymal Stromal Cells, Bahareh Azimi, Massimiliano Labardi, Mohammad Sajad Sorayani Bafqi, Teresa Macchi, Claudio Ricci, Veronica Carnicelli, Lorenzo Scarpelli, Istiak Hussain, Francesca Matino, Mohammed Uddin

Physics & Astronomy Faculty Publications

Highlights

• The type of solvent had noticeable effects on morphology and piezoelectric properties of P(VDF-TrFE) electrospun fibers.

• Using MEK as a solvent and specific environmental conditions led to the obtainment of surface nanopores.

• Uniaxially aligned fibers exhibited higher β phase and mechanical properties than random ones.

• Randomly oriented fibers had higher remnant piezoelectric properties (Vout, d31 piezoelectric coefficient) than aligned ones.

• Human mesenchymal stem cells cultured on randomly oriented fibers showed an accelerated osteogenic differentiation.

Abstract

Many tissues and cells are influenced by mechano-electric stimulation, thus the application of piezoelectric materials has recently received considerable attention in tissue engineering. …

Lattice Investigations Of The Chimera Baryon Spectrum In The Sp (4) Gauge Theory, Ed Bennett, Deog Ki Hong, Ho Hsiao, Jong Wan Lee, C. J.David Lin, Biagio Lucini, Maurizio Piai, Davide Vadacchino

School of Engineering, Computing and Mathematics

We report the results of lattice numerical studies of the Sp(4) gauge theory coupled to fermions (hyperquarks) transforming in the fundamental and two-index antisymmetric representations of the gauge group. This strongly coupled theory is the minimal candidate for the ultraviolet completion of composite Higgs models that facilitate the mechanism of partial compositeness for generating the top-quark mass. We measure the spectrum of the low-lying, half-integer spin, bound states composed of two fundamental and one antisymmetric hyperquarks, dubbed chimera baryons, in the quenched approximation. In this first systematic, nonperturbative study, we focus on the three lightest parity-even chimera-baryon states, in analogy …

(Non)Cognitive Dissonance? A Stakeholder-Based Exploration Of The Consideration Of Graduate Admissions Applicants' Personal Skills And Qualities, Reginald M. Gooch, Joseph H. Paris, Sara B. Haviland, Jose Sotelo

Journal of College Access

Prospective graduate students’ noncognitive attributes are commonly evaluated as a part of a holistic review of their admission applications. Yet it is difficult to determine which noncognitive attributes are considered by those who evaluate graduate admissions applications and what approaches they take to measure applicants’ noncognitive attributes. It is even less clear to what degree prospective graduate students understand how they are evaluated for graduate admissions and how the evaluation of their noncognitive attributes factor into admissions decisions. Drawing on surveys of graduate enrollment management (GEM) professionals and prospective graduate students in the United States, our study investigates the noncognitive …

Master Narrative Of College Access Belies Reality For Today’S Students, Jonathan S. Lewis, René A. Hernandez

Journal of College Access

Master and alternative narratives offer a useful framework through which to consider contemporary issues in college access. Implicit and ubiquitous, the master narrative of a linear progression from high school through a residential college toward a fulfilling career has long been dominant. Meanwhile, alternative narratives of fluid, dynamic, alternate pathways are ascendant, having received a boost from the Covid-19 pandemic and the stubborn lack of affordable postsecondary options. Observing a decline in a shared cultural narrative about college, the authors recommend that advisors help students to sort through possible storylines and then write their own.

A Novel Correction For The Multivariate Ljung-Box Test, Minhao Huang

Computational and Data Sciences (PhD) Dissertations

This research introduces an analytical improvement to the Multivariate Ljung-Box test that addresses significant deviations of the original test from the nominal Type I error rates under almost all scenarios. Prior attempts to mitigate this issue have been directed at modification of the test statistics or correction of the test distribution to achieve precise results in finite samples. In previous studies, focused on designing corrections to the univariate Ljung-Box, a method that specifically adjusts the test rejection region has been the most successful of attaining the best Type I error rates. We adopt the same approach for the more complex, …

Machine Learning And Geostatistical Approaches For Discovery Of Weather And Climate Events Related To El Niño Phenomena, Sachi Perera

Computational and Data Sciences (PhD) Dissertations

El Nino and La Nina are worldwide environmental phenomena brought about by repetitive changes in the water temperature of the Pacific Ocean. Even though the El-Nino impact focuses on a smaller area in the Pacific Ocean near the Equator, these developments have global repercussions, where temperature and precipitation are influenced across the globe, causing droughts and floods simultaneously. In this dissertation, we first derived a drought vulnerability index for the Nile basin, identifying regions with high and low drought risk under ENSO conditions. Next, we evaluated the coherence and periodicity of the ENSO signal to detect its implications on MENA …