Theory and Algorithms Commons^™

Solving Multiple Inference In Graphical Models, Cong Chen

Dissertations, Theses, and Capstone Projects

For inference problems in graphical models, much effort has been directed at algorithms for obtaining one single optimal prediction. In practice, the data is often noisy or incomplete, which makes one single optimal solution unreliable. To address this problem, multiple Inference is proposed to find several best solutions, M-Best, where multiple hypotheses are preferred for advanced reasoning. People use oracle accuracy as an evaluation criterion expecting one of the solutions has high accuracy with the ground truth. It has been shown that it is beneficial for the top solutions to be diverse. Approaches for solving diverse multiple inference are proposed …

Teaching Machine Learning For The Physical Sciences: A Summary Of Lessons Learned And Challenges, Viviana Acquaviva

Publications and Research

This paper summarizes some challenges encountered and best practices established in several years of teaching Machine Learning for the Physical Sciences at the undergraduate and graduate level. I discuss motivations for teaching ML to physicists, desirable properties of pedagogical materials, such as accessibility, relevance, and likeness to real-world research problems, and give examples of components of teaching units.

Solving Large-Scale Extensive-Form Network Security Games Via Neural Fictitious Self-Play, Wanqi Xue, Youzhi Zhang, Shuxin Li, Xinrun Wang, Bo An, Chai Kiat Yeo

Research Collection School Of Computing and Information Systems

Securing networked infrastructures is important in the real world. The problem of deploying security resources to protect against an attacker in networked domains can be modeled as Network Security Games (NSGs). Unfortunately, existing approaches, including the deep learning-based approaches, are inefficient to solve large-scale extensive-form NSGs. In this paper, we propose a novel learning paradigm, NSG-NFSP, to solve large-scale extensive-form NSGs based on Neural Fictitious Self-Play (NFSP). Our main contributions include: i) reforming the best response (BR) policy network in NFSP to be a mapping from action-state pair to action-value, to make the calculation of BR possible in NSGs; ii) …

Neural Regret-Matching For Distributed Constraint Optimization Problems, Yanchen Deng, Runshen Yu, Xinrun Wang, Bo An

Research Collection School Of Computing and Information Systems

Distributed constraint optimization problems (DCOPs) are a powerful model for multi-agent coordination and optimization, where information and controls are distributed among multiple agents by nature. Sampling-based algorithms are important incomplete techniques for solving medium-scale DCOPs. However, they use tables to exactly store all the information (e.g., costs, confidence bounds) to facilitate sampling, which limits their scalability. This paper tackles the limitation by incorporating deep neural networks in solving DCOPs for the first time and presents a neural-based sampling scheme built upon regret-matching. In the algorithm, each agent trains a neural network to approximate the regret related to its local problem …

Desktop Application For The Puzzle Board Game “Rush Hour”, Huanqing Nong

Electronic Theses, Projects, and Dissertations

Rush Hour is a sliding block puzzle board game. This game comes with a board of 6 x 6 grid simulating a parking lot with an exit at the right end of the third row and some vehicle models of size 1 x 2 or 1 x 3 which can slide along the grooves of the grid forward or backward. The goal of the game is to clear the path by moving the vehicles on the board in a certain way for the target car, which lies on the third row of the grid, to merge out the “parking lot” …

Reproducibility Companion Paper: Knowledge Enhanced Neural Fashion Trend Forecasting, Yunshan Ma, Yujuan Ding, Xun Yang, Lizi Liao, Wai Keung Wong, Tat-Seng Chua, Jinyoung Moon, Hong-Han Shuai

Research Collection School Of Computing and Information Systems

This companion paper supports the replication of the fashion trend forecasting experiments with the KERN (Knowledge Enhanced Recurrent Network) method that we presented in the ICMR 2020. We provide an artifact that allows the replication of the experiments using a Python implementation. The artifact is easy to deploy with simple installation, training and evaluation. We reproduce the experiments conducted in the original paper and obtain similar performance as previously reported. The replication results of the experiments support the main claims in the original paper.

Cfr-Mix: Solving Imperfect Information Extensive-Form Games With Combinatorial Action Space, Shuxin Li, Youzhi Zhang, Xinrun Wang, Wanqi Xue, Bo An

Research Collection School Of Computing and Information Systems

In many real-world scenarios, a team of agents must coordinate with each other to compete against an opponent. The challenge of solving this type of game is that the team's joint action space grows exponentially with the number of agents, which results in the inefficiency of the existing algorithms, e.g., Counterfactual Regret Minimization (CFR). To address this problem, we propose a new framework of CFR: CFR-MIX. Firstly, we propose a new strategy representation that represents a joint action strategy using individual strategies of all agents and a consistency relationship to maintain the cooperation between agents. To compute the equilibrium with …

On The Use Of Minimum Penalties In Statistical Learning, Ben Sherwood, Bradley S. Price

Faculty & Staff Scholarship

Modern multivariate machine learning and statistical methodologies estimate parameters of interest while leveraging prior knowledge of the association between outcome variables. The methods that do allow for estimation of relationships do so typically through an error covariance matrix in multivariate regression which does not scale to other types of models. In this article we proposed the MinPEN framework to simultaneously estimate regression coefficients associated with the multivariate regression model and the relationships between outcome variables using mild assumptions. The MinPen framework utilizes a novel penalty based on the minimum function to exploit detected relationships between responses. An iterative algorithm that …

Design And Development Of Techniques To Ensure Integrity In Fog Computing Based Databases, Abdulwahab Fahad S. Alazeb

Graduate Theses and Dissertations

The advancement of information technology in coming years will bring significant changes to the way sensitive data is processed. But the volume of generated data is rapidly growing worldwide. Technologies such as cloud computing, fog computing, and the Internet of things (IoT) will offer business service providers and consumers opportunities to obtain effective and efficient services as well as enhance their experiences and services; increased availability and higher-quality services via real-time data processing augment the potential for technology to add value to everyday experiences. This improves human life quality and easiness. As promising as these technological innovations, they are prone …

Mmconv: An Environment For Multimodal Conversational Search Across Multiple Domains, Lizi Liao, Le Hong Long, Zheng Zhang, Minlie Huang, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Although conversational search has become a hot topic in both dialogue research and IR community, the real breakthrough has been limited by the scale and quality of datasets available. To address this fundamental obstacle, we introduce the Multimodal Multi-domain Conversational dataset (MMConv), a fully annotated collection of human-to-human role-playing dialogues spanning over multiple domains and tasks. The contribution is two-fold. First, beyond the task-oriented multimodal dialogues among user and agent pairs, dialogues are fully annotated with dialogue belief states and dialogue acts. More importantly, we create a relatively comprehensive environment for conducting multimodal conversational search with real user settings, structured …

An Adaptive Large Neighborhood Search For The Green Mixed Fleet Vehicle Routing Problem With Realistic Energy Consumption And Partial Recharges, Vincent F. Yu, Panca Jodiawan, Aldy Gunawan

Research Collection School Of Computing and Information Systems

This study addresses a variant of the Electric Vehicle Routing Problem with Mixed Fleet, named as the Green Mixed Fleet Vehicle Routing Problem with Realistic Energy Consumption and Partial Recharges. This problem contains three important characteristics — realistic energy consumption, partial recharging policy, and carbon emissions. An adaptive Large Neighborhood Search heuristic is developed for the problem. Experimental results show that the proposed ALNS finds optimal solutions for most small-scale benchmark instances in a significantly faster computational time compared to the performance of CPLEX solver. Moreover, it obtains high quality solutions for all medium- and large-scale instances under a reasonable …

Task Similarity Aware Meta Learning: Theory-Inspired Improvement On Maml, Pan Zhou, Yingtian Zpu, Xiaotong Yuan, Jiashi Feng, Caiming Xiong, Steven C. H. Hoi

Research Collection School Of Computing and Information Systems

Few-shot learning ability is heavily desired for machine intelligence. By meta-learning a model initialization from training tasks with fast adaptation ability to new tasks, model-agnostic meta-learning (MAML) has achieved remarkable success in a number of few-shot learning applications. However, theoretical understandings on the learning ability of MAML remain absent yet, hindering developing new and more advanced meta learning methods in a principled way. In this work, we solve this problem by theoretically justifying the fast adaptation capability of MAML when applied to new tasks. Specifically, we prove that the learnt meta-initialization can benefit the fast adaptation to new tasks with …

Object Manipulation With Modular Planar Tensegrity Robots, Maxine Perroni-Scharf

Dartmouth College Undergraduate Theses

This thesis explores the creation of a novel two-dimensional tensegrity-based mod- ular system. When individual planar modules are linked together, they form a larger tensegrity robot that can be used to achieve non-prehensile manipulation. The first half of this dissertation focuses on the study of preexisting types of tensegrity mod- ules and proposes different possible structures and arrangements of modules. The second half describes the construction and actuation of a modular 2D robot com- posed of planar three-bar tensegrity structures. We conclude that tensegrity modules are suitably adapted to object manipulation and propose a future extension of the modular 2D …

A Hybrid Stochastic-Deterministic Minibatch Proximal Gradient Method For Efficient Optimization And Generalization, Pan Zhou, Xiao-Tong Yuan, Lin Zhouchen, Steven C. H. Hoi

Research Collection School Of Computing and Information Systems

Despite the success of stochastic variance-reduced gradient (SVRG) algorithms in solving large-scale problems, their stochastic gradient complexity often scales linearly with data size and is expensive for huge data. Accordingly, we propose a hybrid stochastic-deterministic minibatch proximal gradient (HSDMPG) algorithm for strongly convex problems with linear prediction structure, e.g. least squares and logistic/softmax regression. HSDMPG enjoys improved computational complexity that is data-size-independent for large-scale problems. It iteratively samples an evolving minibatch of individual losses to estimate the original problem, and can efficiently minimize the sampled subproblems. For strongly convex loss of n components, HSDMPG attains an -optimization-error within O κ …

Set Team Orienteering Problem With Time Windows, Aldy Gunawan, Vincent F. Yu, Andros Nicas Sutanto, Panca Jodiawan

Research Collection School Of Computing and Information Systems

This research introduces an extension of the Orienteering Problem (OP), known as Set Team Orienteering Problem with Time Windows (STOPTW), in which customers are first grouped into clusters. Each cluster is associated with a profit that will be collected if at least one customer within the cluster is visited. The objective is to find the best route that maximizes the total collected profit without violating time windows and time budget constraints. We propose an adaptive large neighborhood search algorithm to solve newly introduced benchmark instances. The preliminary results show the capability of the proposed algorithm to obtain good solutions within …

Improving Additional Adversarial Robustness For Classification, Michael Guo

McKelvey School of Engineering Graduate Student Theses & Dissertations

Although neural networks have achieved remarkable success on classification, adversarial robustness is still a significant concern. There are now a series of approaches for designing adversarial examples and methods to defending against them. This paper consists of two projects. In our first work, we propose an approach by leveraging cognitive salience to enhance additional robustness on top of these methods. Specifically, for image classification, we split an image into the foreground (salient region) and background (the rest) and allow significantly larger adversarial perturbations in the background to produce stronger attacks. Furthermore, we show that adversarial training with dual-perturbation attacks yield …

Reinforcement Learning For Realistic Robotic Training: A Survey, Andres Jaramillo

Honors Scholar Theses

Reinforcement learning is a widely popular topic that has resulted in a plethora of

research papers and interest from academia and industry. When applied with robotics,

the field has showed some promising signs that robots can achieve levels of complex

cognitive abilities rivaling humans, but the goal of creating sapient robots is far from

a reality due to many challenges involved with training robots in a real world setting.

This paper will provide a survey regarding the keys towards realistic robotic training by

detailing the challenges and overviewing the reinforcement learning solutions involved

in getting a robot to think like …

Semi-Supervised Spatial-Temporal Feature Learning On Anomaly-Based Network Intrusion Detection, Huy Mai

Computer Science and Computer Engineering Undergraduate Honors Theses

Due to a rapid increase in network traffic, it is growing more imperative to have systems that detect attacks that are both known and unknown to networks. Anomaly-based detection methods utilize deep learning techniques, including semi-supervised learning, in order to effectively detect these attacks. Semi-supervision is advantageous as it doesn't fully depend on the labelling of network traffic data points, which may be a daunting task especially considering the amount of traffic data collected. Even though deep learning models such as the convolutional neural network have been integrated into a number of proposed network intrusion detection systems in recent years, …

Dynamic Task Allocation In Partially Defined Environments Using A* With Bounded Costs, James Hendrickson

Doctoral Dissertations and Master's Theses

The sector of maritime robotics has seen a boom in operations in areas such as surveying and mapping, clean-up, inspections, search and rescue, law enforcement, and national defense. As this sector has continued to grow, there has been an increased need for single unmanned systems to be able to undertake more complex and greater numbers of tasks. As the maritime domain can be particularly difficult for autonomous vehicles to operate in due to the partially defined nature of the environment, it is crucial that a method exists which is capable of dynamically accomplishing tasks within this operational domain. By considering …

A Matheuristic Algorithm For The Vehicle Routing Problem With Cross-Docking, Aldy Gunawan, Audrey Tedja Widjaja, Pieter Vansteenwegen, Vincent F. Yu

Research Collection School Of Computing and Information Systems

This paper studies the integration of the vehicle routing problem with cross-docking (VRPCD). The aim is to find a set of routes to deliver products from a set of suppliers to a set of customers through a cross-dock facility, such that the operational and transportation costs are minimized, without violating the vehicle capacity and time horizon constraints. A two-phase matheuristic based on column generation is proposed. The first phase focuses on generating a set of feasible candidate routes in both pickup and delivery processes by implementing an adaptive large neighborhood search algorithm. A set of destroy and repair operators are …

Machine Learning Approaches To Dribble Hand-Off Action Classification With Sportvu Nba Player Coordinate Data, Dembe Stephanos

Electronic Theses and Dissertations

Recently, strategies of National Basketball Association teams have evolved with the skillsets of players and the emergence of advanced analytics. One of the most effective actions in dynamic offensive strategies in basketball is the dribble hand-off (DHO). This thesis proposes an architecture for a classification pipeline for detecting DHOs in an accurate and automated manner. This pipeline consists of a combination of player tracking data and event labels, a rule set to identify candidate actions, manually reviewing game recordings to label the candidates, and embedding player trajectories into hexbin cell paths before passing the completed training set to the classification …

Technological Tethereds: Potential Impact Of Untrustworthy Artificial Intelligence In Criminal Justice Risk Assessment Instruments, Sonia M. Gipson Rankin

Faculty Scholarship

Issues of racial inequality and violence are front and center in today’s society, as are issues surrounding artificial intelligence (AI). This Article, written by a law professor who is also a computer scientist, takes a deep dive into understanding how and why hacked and rogue AI creates unlawful and unfair outcomes, particularly for persons of color.

Black Americans are disproportionally featured in criminal justice, and their stories are obfuscated. The seemingly endless back-to-back murders of George Floyd, Breonna Taylor, and Ahmaud Arbery, and heartbreakingly countless others have finally shaken the United States from its slumbering journey towards intentional criminal justice …

Wg2An: Synthetic Wound Image Generation Using Generative Adversarial Network, Salih Sarp, Murat Kuzlu, Emmanuel Wilson, Ozgur Guler

Engineering Technology Faculty Publications

In part due to its ability to mimic any data distribution, Generative Adversarial Network (GAN) algorithms have been successfully applied to many applications, such as data augmentation, text-to-image translation, image-to-image translation, and image inpainting. Learning from data without crafting loss functions for each application provides broader applicability of the GAN algorithm. Medical image synthesis is also another field that the GAN algorithm has great potential to assist clinician training. This paper proposes a synthetic wound image generation model based on GAN architecture to increase the quality of clinical training. The proposed model is trained on chronic wound datasets with various …

Norm-Based Generalisation Bounds For Deep Multi-Class Convolutional Neural Networks, Antoine Ledent, Waleed Mustafa, Yunwen Lei, Marius Kloft

Research Collection School Of Computing and Information Systems

We show generalisation error bounds for deep learning with two main improvements over the state of the art. (1) Our bounds have no explicit dependence on the number of classes except for logarithmic factors. This holds even when formulating the bounds in terms of the Frobenius-norm of the weight matrices, where previous bounds exhibit at least a squareroot dependence on the number of classes. (2) We adapt the classic Rademacher analysis of DNNs to incorporate weight sharing—a task of fundamental theoretical importance which was previously attempted only under very restrictive assumptions. In our results, each convolutional filter contributes only once …

Fine-Grained Generalization Analysis Of Vector-Valued Learning, Liang Wu, Antoine Ledent, Yunwen Lei, Marius Kloft

Research Collection School Of Computing and Information Systems

Many fundamental machine learning tasks can be formulated as a problem of learning with vector-valued functions, where we learn multiple scalar-valued functions together. Although there is some generalization analysis on different specific algorithms under the empirical risk minimization principle, a unifying analysis of vector-valued learning under a regularization framework is still lacking. In this paper, we initiate the generalization analysis of regularized vector-valued learning algorithms by presenting bounds with a mild dependency on the output dimension and a fast rate on the sample size. Our discussions relax the existing assumptions on the restrictive constraint of hypothesis spaces, smoothness of loss …

Using Torchattacks To Improve The Robustness Of Models With Adversarial Training, William S. Matos Díaz

Cybersecurity: Deep Learning Driven Cybersecurity Research in a Multidisciplinary Environment

Adversarial training has proven to be one of the most successful ways to defend models against adversarial examples. This process consists of training a model with an adversarial example to improve the robustness of the model. In this experiment, Torchattacks, a Pytorch library made for importing adversarial examples more easily, was used to determine which attack was the strongest. Later on, the strongest attack was used to train the model and make it more robust against adversarial examples. The datasets used to perform the experiments were MNIST and CIFAR-10. Both datasets were put to the test using PGD, FGSM, and …

Representing And Learning Preferences Over Combinatorial Domains, Michael Huelsman

Theses and Dissertations--Computer Science

Agents make decisions based on their preferences. Thus, to predict their decisions one has to learn the agent's preferences. A key step in the learning process is selecting a model to represent those preferences. We studied this problem by borrowing techniques from the algorithm selection problem to analyze preference example sets and select the most appropriate preference representation for learning. We approached this problem in multiple steps.

First, we determined which representations to consider. For this problem we developed the notion of preference representation language subsumption, which compares representations based on their expressive power. Subsumption creates a hierarchy of preference …

Learning From Multi-Class Imbalanced Big Data With Apache Spark, William C. Sleeman Iv

Theses and Dissertations

With data becoming a new form of currency, its analysis has become a top priority in both academia and industry, furthering advancements in high-performance computing and machine learning. However, these large, real-world datasets come with additional complications such as noise and class overlap. Problems are magnified when with multi-class data is presented, especially since many of the popular algorithms were originally designed for binary data. Another challenge arises when the number of examples are not evenly distributed across all classes in a dataset. This often causes classifiers to favor the majority class over the minority classes, leading to undesirable results …

Scaling Up Exact Neural Network Compression By Relu Stability, Thiago Serra, Xin Yu, Abhinav Kumar, Srikumar Ramalingam

Faculty Conference Papers and Presentations

We can compress a rectifier network while exactly preserving its underlying functionality with respect to a given input domain if some of its neurons are stable. However, current approaches to determine the stability of neurons with Rectified Linear Unit (ReLU) activations require solving or finding a good approximation to multiple discrete optimization problems. In this work, we introduce an algorithm based on solving a single optimization problem to identify all stable neurons. Our approach is on median 183 times faster than the state-of-art method on CIFAR-10, which allows us to explore exact compression on deeper (5 x 100) and wider …

Ssentiaa: A Self-Supervised Sentiment Analyzer For Classification From Unlabeled Data, Salim Sazzed, Sampath Jayarathna

Computer Science Faculty Publications

In recent years, supervised machine learning (ML) methods have realized remarkable performance gains for sentiment classification utilizing labeled data. However, labeled data are usually expensive to obtain, thus, not always achievable. When annotated data are unavailable, the unsupervised tools are exercised, which still lag behind the performance of supervised ML methods by a large margin. Therefore, in this work, we focus on improving the performance of sentiment classification from unlabeled data. We present a self-supervised hybrid methodology SSentiA (Self-supervised Sentiment Analyzer) that couples an ML classifier with a lexicon-based method for sentiment classification from unlabeled data. We first introduce LRSentiA …