Theory and Algorithms Commons^™

Towards Understanding Why Lookahead Generalizes Better Than Sgd And Beyond, Pan Zhou, Hanshu Yan, Xiaotong Yuan, Jiashi Feng, Shuicheng Yan

Research Collection School Of Computing and Information Systems

To train networks, lookahead algorithm [1] updates its fast weights k times via an inner-loop optimizer before updating its slow weights once by using the latest fast weights. Any optimizer, e.g. SGD, can serve as the inner-loop optimizer, and the derived lookahead generally enjoys remarkable test performance improvement over the vanilla optimizer. But theoretical understandings on the test performance improvement of lookahead remain absent yet. To solve this issue, we theoretically justify the advantages of lookahead in terms of the excess risk error which measures the test performance. Specifically, we prove that lookahead using SGD as its inner-loop optimizer can …

Fair And Diverse Group Formation Based On Multidimensional Features, Mohammed Saad A Alqahtani

Graduate Theses and Dissertations

The goal of group formation is to build a team to accomplish a specific task. Algorithms are being developed to improve the team's effectiveness so formed and the efficiency of the group selection process. However, there is concern that team formation algorithms could be biased against minorities due to the algorithms themselves or the data on which they are trained. Hence, it is essential to build fair team formation systems that incorporate demographic information into the process of building the group. Although there has been extensive work on modeling individuals’ expertise for expert recommendation and/or team formation, there has been …

Russian Logics And The Culture Of Impossible: Part 1. Recovering Intelligentsia Logics, Ksenia Tatarchenko, Anya Yermakova, Liesbeth De Mol

Research Collection College of Integrative Studies

This article reinterprets algorithmic rationality by looking at the interaction between mathematical logic, mechanized reasoning, and, later, computing in the Russian Imperial and Soviet contexts to offer a history of the algorithm as a mathematical object bridging the inner and outer worlds, a humanistic vision that we, following logician Vladimir Uspensky, call the “culture of the impossible.” We unfold the deep roots of this vision as embodied in scientific intelligentsia. In Part I, we examine continuities between the turn-of-the-twentieth-century discussions of poznaniye—an epistemic orientation towards the process of knowledge acquisition—and the postwar rise of the Soviet school of mathematical logic. …

Rmix: Learning Risk-Sensitive Policies For Cooperative Reinforcement Learning Agents, Wei Qiu, Xinrun Wang, Runsheng Yu, Xu He, Rundong Wang, Bo An, Svetlana Obraztsova, Zinovi Rabinovich

Research Collection School Of Computing and Information Systems

Current value-based multi-agent reinforcement learning methods optimize individual Q values to guide individuals' behaviours via centralized training with decentralized execution (CTDE). However, such expected, i.e., risk-neutral, Q value is not sufficient even with CTDE due to the randomness of rewards and the uncertainty in environments, which causes the failure of these methods to train coordinating agents in complex environments. To address these issues, we propose RMIX, a novel cooperative MARL method with the Conditional Value at Risk (CVaR) measure over the learned distributions of individuals' Q values. Specifically, we first learn the return distributions of individuals to analytically calculate CVaR …

Beyond Smoothness : Incorporating Low-Rank Analysis Into Nonparametric Density Estimation, Rob Vandermeulen, Antoine Ledent

Research Collection School Of Computing and Information Systems

The construction and theoretical analysis of the most popular universally consistent nonparametric density estimators hinge on one functional property: smoothness. In this paper we investigate the theoretical implications of incorporating a multi-view latent variable model, a type of low-rank model, into nonparametric density estimation. To do this we perform extensive analysis on histogram-style estimators that integrate a multi-view model. Our analysis culminates in showing that there exists a universally consistent histogram-style estimator that converges to any multi-view model with a finite number of Lipschitz continuous components at a rate of ˜O(1/3√n) in L1 error. In contrast, the standard histogram estimator …

Ggnb: Graph-Based Gaussian Naive Bayes Intrusion Detection System For Can Bus, Riadul Islam, Maloy K. Devnath, Manar D. Samad, Syed Md Jaffrey Al Kadry

Computer Science Faculty Research

The national highway traffic safety administration (NHTSA) identified cybersecurity of the automobile systems are more critical than the security of other information systems. Researchers already demonstrated remote attacks on critical vehicular electronic control units (ECUs) using controller area network (CAN). Besides, existing intrusion detection systems (IDSs) often propose to tackle a specific type of attack, which may leave a system vulnerable to numerous other types of attacks. A generalizable IDS that can identify a wide range of attacks within the shortest possible time has more practical value than attack-specific IDSs, which is not a trivial task to accomplish. In this …

Three-Dimensional Graph Matching To Identify Secondary Structure Correspondence Of Medium-Resolution Cryo-Em Density Maps, Bahareh Behkamal, Mahmoud Naghibzadeh, Mohammad Reza Saberi, Zeinab Amiri Tehranizadeh, Andrea Pagnani, Kamal Al Nasr

Computer Science Faculty Research

Cryo-electron microscopy (cryo-EM) is a structural technique that has played a significant role in protein structure determination in recent years. Compared to the traditional methods of X-ray crystallography and NMR spectroscopy, cryo-EM is capable of producing images of much larger protein complexes. However, cryo-EM reconstructions are limited to medium-resolution (~4–10 Å) for some cases. At this resolution range, a cryo-EM density map can hardly be used to directly determine the structure of proteins at atomic level resolutions, or even at their amino acid residue backbones. At such a resolution, only the position and orientation of secondary structure elements (SSEs) such …

Comparison Of Multiple Imputation Algorithms And Verification Using Whole-Genome Sequencing In The Cmuh Genetic Biobank, Ting-Yuan Liu, Chih-Fan Lin, Hsing-Tsung Wu, Ya-Lun Wu, Yu-Chia Chen, Chi-Chou Liao, Yu-Pao Chou, Dysan Chao, Hsing-Fang Lu, Ya-Sian Chang, Jan-Gowth Chang, Kai-Cheng Hsu, Fuu‑Jen Tsai

BioMedicine

A genome-wide association study (GWAS) can be conducted to systematically analyze the contributions of genetic factors to a wide variety of complex diseases. Nevertheless, existing GWASs have provided highly ethnic specific data. Accordingly, to provide data specific to Taiwan, we established a large-scale genetic database in a single medical institution at the China Medical University Hospital. With current technological limitations, microarray analysis can detect only a limited number of single-nucleotide polymorphisms (SNPs) with a minor allele frequency of >1%. Nevertheless, imputation represents a useful alternative means of expanding data. In this study, we compared four imputation algorithms in terms of …

Transfer-Learned Pruned Deep Convolutional Neural Networks For Efficient Plant Classification In Resource-Constrained Environments, Martinson Ofori

Masters Theses & Doctoral Dissertations

Traditional means of on-farm weed control mostly rely on manual labor. This process is time-consuming, costly, and contributes to major yield losses. Further, the conventional application of chemical weed control can be economically and environmentally inefficient. Site-specific weed management (SSWM) counteracts this by reducing the amount of chemical application with localized spraying of weed species. To solve this using computer vision, precision agriculture researchers have used remote sensing weed maps, but this has been largely ineffective for early season weed control due to problems such as solar reflectance and cloud cover in satellite imagery. With the current advances in artificial …

Information Extraction And Classification On Journal Papers, Lei Yu

School of Computing: Dissertations, Theses, and Student Research

The importance of journals for diffusing the results of scientific research has increased considerably. In the digital era, Portable Document Format (PDF) became the established format of electronic journal articles. This structured form, combined with a regular and wide dissemination, spread scientific advancements easily and quickly. However, the rapidly increasing numbers of published scientific articles requires more time and effort on systematic literature reviews, searches and screens. The comprehension and extraction of useful information from the digital documents is also a challenging task, due to the complex structure of PDF.

To help a soil science team from the United States …

Span-Level Emotion Cause Analysis With Neural Sequence Tagging, Xiangju Li, Wei Gao, Shi Feng, Daling Wang, Shafiq Joty

Research Collection School Of Computing and Information Systems

This paper addresses the task of span-level emotion cause analysis (SECA). It is a finer-grained emotion cause analysis (ECA) task, which aims to identify the specific emotion cause span(s) behind certain emotions in text. In this paper, we formalize SECA as a sequence tagging task for which several variants of neural network-based sequence tagging models to extract specific emotion cause span(s) in the given context. These models combine different types of encoding and decoding approaches. Furthermore, to make our models more "emotionally sensitive'', we utilize the multi-head attention mechanism to enhance the representation of context. Experimental evaluations conducted on two …

Analytical Models For Traffic Congestion And Accident Analysis, Hongrui Liu, Rahul Ramachandra Shetty

Mineta Transportation Institute

In the US, over 38,000 people die in road crashes each year, and 2.35 million are injured or disabled, according to the statistics report from the Association for Safe International Road Travel (ASIRT) in 2020. In addition, traffic congestion keeping Americans stuck on the road wastes millions of hours and billions of dollars each year. Using statistical techniques and machine learning algorithms, this research developed accurate predictive models for traffic congestion and road accidents to increase understanding of the complex causes of these challenging issues. The research used US Accidents data consisting of 49 variables describing 4.2 million accident records …

Expediting The Accuracy-Improving Process Of Svms For Class Imbalance Learning, Bin Cao, Yuqi Liu, Chenyu Hou, Jing Fan, Baihua Zheng, Jianwei Jin

Research Collection School Of Computing and Information Systems

To improve the classification performance of support vector machines (SVMs) on imbalanced datasets, cost-sensitive learning methods have been proposed, e.g., DEC (Different Error Costs) and FSVM-CIL (Fuzzy SVM for Class Imbalance Learning). They relocate the hyperplane by adjusting the costs associated with misclassifying samples. However, the error costs are determined either empirically or by performing an exhaustive search in the parameter space. Both strategies can not guarantee effectiveness and efficiency simultaneously. In this paper, we propose ATEC, a solution that can efficiently find a preferable hyperplane by automatically tuning the error cost for between-class samples. ATEC distinguishes itself from all …

On A Multistage Discrete Stochastic Optimization Problem With Stochastic Constraints And Nested Sampling, Thuy Anh Ta, Tien Mai, Fabian Bastin, Pierre L'Ecuyer

Research Collection School Of Computing and Information Systems

We consider a multistage stochastic discrete program in which constraints on any stage might involve expectations that cannot be computed easily and are approximated by simulation. We study a sample average approximation (SAA) approach that uses nested sampling, in which at each stage, a number of scenarios are examined and a number of simulation replications are performed for each scenario to estimate the next-stage constraints. This approach provides an approximate solution to the multistage problem. To establish the consistency of the SAA approach, we first consider a two-stage problem and show that in the second-stage problem, given a scenario, the …

Adaptive Posterior Knowledge Selection For Improving Knowledge-Grounded Dialogue Generation, Weichao Wang, Wei Gao, Shi Feng, Ling Chen, Daling Wang

Research Collection School Of Computing and Information Systems

In open-domain dialogue systems, knowledge information such as unstructured persona profiles, text descriptions and structured knowledge graph can help incorporate abundant background facts for delivering more engaging and informative responses. Existing studies attempted to model a general posterior distribution over candidate knowledge by considering the entire response utterance as a whole at the beginning of decoding process for knowledge selection. However, a single smooth distribution could fail to model the variability of knowledge selection patterns over different decoding steps, and make the knowledge expression less consistent. To remedy this issue, we propose an adaptive posterior knowledge selection framework, which sequentially …

Span-Level Emotion Cause Analysis By Bert-Based Graph Attention Network, Xiangju Li, Wei Gao, Shi Feng, Daling Wang, Shafiq Joty

Research Collection School Of Computing and Information Systems

We study the task of span-level emotion cause analysis (SECA), which is focused on identifying the specific emotion cause span(s) triggering a certain emotion in the text. Compared to the popular clause-level emotion cause analysis (CECA), it is a finer-grained emotion cause analysis (ECA) task. In this paper, we design a BERT-based graph attention network for emotion cause span(s) identification. The proposed model takes advantage of the structure of BERT to capture the relationship information between emotion and text, and utilizes graph attention network to model the structure information of the text. Our SECA method can be easily used for …

Fuzzy Multi Criteria Decision Making Approach For Technology Selection For Emissions Reduction From Seaborne Transportation Under Uncertainty And Vagueness, Rachid Mouaici

World Maritime University Dissertations

No abstract provided.

Multi-Valued Solutions For The Equation Of Motion, Darcy-Jordan Model, As A Cauchy Problem: A Shocking Event, Chandler Shimp

Master's Theses

Shocks are physical phenomenon that occur quite often around us. In this thesis we examine the occurrence of shocks in finite amplitude acoustic waves from a numerical perspective. These waves, or jump discontinuities, yield ill-behaved solutions when solved numerically. This study takes on the challenge of finding both single- and multi-valued solutions.

The previously unsolved problem in this study is the representation of the Equation of Motion (EoM) in the form of the Darcy-Jordan model (DJM) and expressed as a dimensionless IVP Cauchy problem. Prior attempts to solve have resulted only in implicit solutions or explicit solutions with certain initial …

Introduction To Discrete Mathematics: An Oer For Ma-471, Mathieu Sassolas

Open Educational Resources

The first objective of this book is to define and discuss the meaning of truth in mathematics. We explore logics, both propositional and first-order , and the construction of proofs, both formally and human-targeted. Using the proof tools, this book then explores some very fundamental definitions of mathematics through set theory. This theory is then put in practice in several applications. The particular (but quite widespread) case of equivalence and order relations is studied with detail. Then we introduces sequences and proofs by induction, followed by number theory. Finally, a small introduction to combinatorics is …

Disentangling Hate In Online Memes, Ka Wei, Roy Lee, Rui Cao, Ziqing Fan, Jing Jiang, Wen Haw Chong

Research Collection School Of Computing and Information Systems

Hateful and offensive content detection has been extensively explored in a single modality such as text. However, such toxic information could also be communicated via multimodal content such as online memes. Therefore, detecting multimodal hateful content has recently garnered much attention in academic and industry research communities. This paper aims to contribute to this emerging research topic by proposing DisMultiHate, which is a novel framework that performed the classification of multimodal hateful content. Specifically, DisMultiHate is designed to disentangle target entities in multimodal memes to improve the hateful content classification and explainability. We conduct extensive experiments on two publicly available …