Physical Sciences and Mathematics Commons^™

Unsupervised Learning Trojan, Arturo Geigel

CCE Theses and Dissertations

This work presents a proof of concept of an Unsupervised Learning Trojan. The Unsupervised Learning Trojan presents new challenges over previous work on the Neural network Trojan, since the attacker does not control most of the environment. The current work will presented an analysis of how the attack can be successful by proposing new assumptions under which the attack can become a viable one. A general analysis of how the compromise can be theoretically supported is presented, providing enough background for practical implementation development. The analysis was carried out using 3 selected algorithms that can cover a wide variety of …

Traccs: Trajectory-Aware Coordinated Urban Crowd-Sourcing, Cen Chen, Shih-Fen Cheng, Aldy Gunawan, Archan Misra, Koustuv Dasgupta, Deepthi Chander

Research Collection School Of Computing and Information Systems

We investigate the problem of large-scale mobile crowd-tasking, where a large pool of citizen crowd-workers are used to perform a variety of location-specific urban logistics tasks. Current approaches to such mobile crowd-tasking are very decentralized: a crowd-tasking platform usually provides each worker a set of available tasks close to the worker's current location; each worker then independently chooses which tasks she wants to accept and perform. In contrast, we propose TRACCS, a more coordinated task assignment approach, where the crowd-tasking platform assigns a sequence of tasks to each worker, taking into account their expected location trajectory over a wider time …

Understanding Intelligent Systems, Tamara Kneese

Media Studies

Science fiction has long imagined a workforce reshaped by robots, but the increasingly common instantiation of intelligent systems in business is much more mundane. Beyond the utopian and dystopian hype of increased efficiencies and job displacement, how do we understand what disruptions intelligent systems will have on the workforce?

Context-Aware Spelling Corrector For Sentiment Analysis, Dr. Muhammad Zubair Asghar, Fazal Masud Kundi

Dr. Muhammad Zubair Asghar

One of the most thrived features of the Web 2.0 era is the fastest growing of user-generated content in the shape of blogs and reviews, with unmatched speed and size. These reviews contain poor, text quality and structure which results spelling mistakes as well as out-of-vocabulary words. This paper presents a Context-Aware Spelling Corrector for Sentiment Analysis based on similarity measures and statistical language model. The paper also presents some compelling statistics about spelling errors. The comparative results show that the proposed framework outperforms the related systems, features wise and in accuracy.

Machine Learning Nuclear Detonation Features, Daniel T. Schmitt, Gilbert L. Peterson

Faculty Publications

Nuclear explosion yield estimation equations based on a 3D model of the explosion volume will have a lower uncertainty than radius based estimation. To accurately collect data for a volume model of atmospheric explosions requires building a 3D representation from 2D images. The majority of 3D reconstruction algorithms use the SIFT (scale-invariant feature transform) feature detection algorithm which works best on feature-rich objects with continuous angular collections. These assumptions are different from the archive of nuclear explosions that have only 3 points of view. This paper reduces 300 dimensions derived from an image based on Fourier analysis and five edge …

Hippi Care Hospital: Towards Proactive Business Processes In Emergency Room Services, Kar Way Tan, Venky Shankaraman

Research Collection School Of Computing and Information Systems

It was 2.35 am on a Saturday morning. Wiki Lim, process specialist from the Process Innovation Centre (PIC) of Hippi Care Hospital (HCH), desperately doodling on her notepad for ideas to improve service delivery at HCH’s Emergency Department (ED). HCH has committed to the public that its ED would meet the service quality criterion of serving 90% of A3 and A4 patients, non-emergency patients with moderate to mild symptoms, within 90 minutes of their arrival at the ED. The ED was not able to meet this performance goal and Dr. Edward Kim, the head of the ED at HCH, had …

Partisan Sharing: Facebook Evidence And Societal Consequences, Jisun An, Daniele Quercia, Jon Crowcroft

Research Collection School Of Computing and Information Systems

The hypothesis of selective exposure assumes that people seek out information that supports their views and eschew information that conflicts with their beliefs, and that has negative consequences on our society. Few researchers have recently found counter evidence of selective exposure in social media: users are exposed to politically diverse articles. No work has looked at what happens after exposure, particularly how individuals react to such exposure, though. Users might well be exposed to diverse articles but share only the partisan ones. To test this, we study partisan sharing on Facebook: the tendency for users to predominantly share like-minded news …

Multi-Agent Orienteering Problem With Time-Dependent Capacity Constraints, Cen Chen, Shih-Fen Cheng, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

In this paper, we formulate and study the Multi-agent Orienteering Problem with Time-dependent Capacity Constraints (MOPTCC). MOPTCC is similar to the classical orienteering problem at single-agent level: given a limited time budget, an agent travels around the network and collects rewards by visiting different nodes, with the objective of maximizing the sum of his collected rewards. The most important feature we introduce in MOPTCC is the inclusion of multiple competing agents. All agents in MOPTCC are assumed to be self-interested, and they interact with each other when arrive at certain nodes simultaneously. As all nodes are capacitated, if a particular …

Camera Viewpoint Invariant Vehicular Traffic Segmentation And Classification, Jeffrey B. Flora

Electrical & Computer Engineering Theses & Dissertations

The goal of this intelligent transportation systems work is to develop a computer vision method that is view angle independent for segmenting and classifying vehicular traffic on highway systems. In order to achieve this goal, this work implements an algorithm for vehicle segmentation, feature extraction, and classification using the existing Virginia Department of Transportation (VDOT) infrastructure on networked traffic cameras. The VDOT traffic video is analyzed for vehicle detection and segmentation using an adaptive Gaussian mixture model algorithm. Speed estimation is performed using a single camera calibration. Size and shape features from morphological properties and texture features from histogram of …

Auction With Rolling Horizon For Urban Consolidation Centre, Chen Wang, Stephanus Daniel Handoko, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

A number of cities around the world have adopted urban consolidation centres (UCCs) to address some challenges of their last-mile deliveries. At the UCC, goods are consolidated based on their destinations prior to their deliveries into the city centre. In many examples, the UCC owns a fleet of eco-friendly vehicles to carry out the deliveries. A carrier/shipper who buys the UCC’s service hence no longer needs to enter the city centre in which time-window and vehicle-type restrictions may apply. As a result, it becomes possible to retain the use of large trucks for the economies of scale outside the city …

Energy Based Multi-Model Fitting And Matching Problems, Hossam N. Isack

Electronic Thesis and Dissertation Repository

Feature matching and model fitting are fundamental problems in multi-view geometry. They are chicken-&-egg problems: if models are known it is easier to find matches and vice versa. Standard multi-view geometry techniques sequentially solve feature matching and model fitting as two independent problems after making fairly restrictive assumptions. For example, matching methods rely on strong discriminative power of feature descriptors, which fail for stereo images with repetitive textures or wide baseline. Also, model fitting methods assume given feature matches, which are not known a priori. Moreover, when data supports multiple models the fitting problem becomes challenging even with known matches …

Adapting Monte Carlo Localization To Utilize Floor And Wall Texture Data, Stephanie Krapil

Master's Theses

Monte Carlo Localization (MCL) is an algorithm that allows a robot to determine its location when provided a map of its surroundings. Particles, consisting of a location and an orientation, represent possible positions where the robot could be on the map. The probability of the robot being at each particle is calculated based on sensor input.

Traditionally, MCL only utilizes the position of objects for localization. This thesis explores using wall and floor surface textures to help the algorithm determine locations more accurately. Wall textures are captured by using a laser range finder to detect patterns in the surface. Floor …

Using Gist Features To Constrain Search In Object Detection, Joanna Browne Solmon

Dissertations and Theses

This thesis investigates the application of GIST features [13] to the problem of object detection in images. Object detection refers to locating instances of a given object category in an image. It is contrasted with object recognition, which simply decides whether an image contains an object, regardless of the object's location in the image.

In much of computer vision literature, object detection uses a "sliding window" approach to finding objects in an image. This requires moving various sizes of windows across an image and running a trained classifier on the visual features of each window. This brute force method can …

Identification Of Informativeness In Text Using Natural Language Stylometry, Rushdi Shams

Electronic Thesis and Dissertation Repository

In this age of information overload, one experiences a rapidly growing over-abundance of written text. To assist with handling this bounty, this plethora of texts is now widely used to develop and optimize statistical natural language processing (NLP) systems. Surprisingly, the use of more fragments of text to train these statistical NLP systems may not necessarily lead to improved performance. We hypothesize that those fragments that help the most with training are those that contain the desired information. Therefore, determining informativeness in text has become a central issue in our view of NLP. Recent developments in this field have spawned …

Automated Image Interpretation For Science Autonomy In Robotic Planetary Exploration, Raymond Francis

Electronic Thesis and Dissertation Repository

Advances in the capabilities of robotic planetary exploration missions have increased the wealth of scientific data they produce, presenting challenges for mission science and operations imposed by the limits of interplanetary radio communications. These data budget pressures can be relieved by increased robotic autonomy, both for onboard operations tasks and for decision- making in response to science data.

This thesis presents new techniques in automated image interpretation for natural scenes of relevance to planetary science and exploration, and elaborates autonomy scenarios under which they could be used to extend the reach and performance of exploration missions on planetary surfaces.

Two …

Improved Microrobotic Control Through Image Processing And Automated Hardware Interfacing, Archit R. Aggarwal, Wuming Jing, David J. Cappelleri

The Summer Undergraduate Research Fellowship (SURF) Symposium

Untethered submilliliter-sized robots (microrobots) are showing potential use in different industrial, manufacturing and medical applications. A particular type of these microrobots, magnetic robots, have shown improved performance in power and control capabilities compared to the other thermal and electrostatic based robots. However, the magnetic robot designs have not been assessed in a robust manner to understand the degree of control in different environments and their application feasibility. This research project seeks to develop a custom control software interface to provide a holistic tool for researchers to evaluate the microrobotic performance through advance control features. The software deliverable involved two main …

3d Robotic Sensing Of People: Human Perception, Representation And Activity Recognition, Hao Zhang

Doctoral Dissertations

The robots are coming. Their presence will eventually bridge the digital-physical divide and dramatically impact human life by taking over tasks where our current society has shortcomings (e.g., search and rescue, elderly care, and child education). Human-centered robotics (HCR) is a vision to address how robots can coexist with humans and help people live safer, simpler and more independent lives.

As humans, we have a remarkable ability to perceive the world around us, perceive people, and interpret their behaviors. Endowing robots with these critical capabilities in highly dynamic human social environments is a significant but very challenging problem in practical …

Convergence Of A Reinforcement Learning Algorithm In Continuous Domains, Stephen Carden

All Dissertations

In the field of Reinforcement Learning, Markov Decision Processes with a finite number of states and actions have been well studied, and there exist algorithms capable of producing a sequence of policies which converge to an optimal policy with probability one. Convergence guarantees for problems with continuous states also exist. Until recently, no online algorithm for continuous states and continuous actions has been proven to produce optimal policies. This Dissertation contains the results of research into reinforcement learning algorithms for problems in which both the state and action spaces are continuous. The problems to be solved are introduced formally as …

Multi-Valued Argumentation Frameworks, Pierpaolo Dondio

Conference papers

In this paper we explore how the seminal Dung’s abstract argumentation framework can be extended to handle arguments containing gradual concepts. We allow arguments to have a degree of truth associated with them and we investigate the degree of truth to which each argument can be considered accepted, rejected and undecided by an abstract argumentation semantics. We propose a truth-compositional recursive computation, and we discuss examples using the major multi-valued logics such as Godel’s, Zadeh’s and Łukasiewicz's logic. The findings are a contribution in the field of non-monotonic approximate reasoning and they also represent a well-grounded proposal towards the introduction …

A Comparative Study Of Underwater Robot Path Planning Algorithms For Adaptive Sampling In A Network Of Sensors, Sreeja Banerjee

Department of Computer Science and Engineering: Dissertations, Theses, and Student Research

Monitoring lakes, rivers, and oceans is critical to improving our understanding of complex large-scale ecosystems. We introduce a method of underwater monitoring using semi-mobile underwater sensor networks and mobile underwater robots in this thesis. The underwater robots can move freely in all dimension while the sensor nodes are anchored to the bottom of the water column and can move only up and down along the depth of the water column. We develop three different algorithms to optimize the path of the underwater robot and the positions of the sensors to improve the overall quality of sensing of an area of …

Integrating Motivated Learning And K-Winner-Take-All To Coordinate Multi-Agent Reinforcement Learning, Teck-Hou Teng, Ah-Hwee Tan, Janusz Starzyk, Yuan-Sin Tan, Loo-Nin Teow

Research Collection School Of Computing and Information Systems

This work addresses the coordination issue in distributed optimization problem (DOP) where multiple distinct and time-critical tasks are performed to satisfy a global objective function. The performance of these tasks has to be coordinated due to the sharing of consumable resources and the dependency on non-consumable resources. Knowing that it can be sub-optimal to predefine the performance of the tasks for large DOPs, the multi-agent reinforcement learning (MARL) framework is adopted wherein an agent is used to learn the performance of each distinct task using reinforcement learning. To coordinate MARL, we propose a novel coordination strategy integrating Motivated Learning (ML) …

A Mathematical Model And Metaheuristics For Time Dependent Orienteering Problem, Aldy Gunawan, Zhi Yuan, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

This paper presents a generalization of the Orienteering Problem, the Time-Dependent Orienteering Problem (TDOP) which is based on the real-life application of providing automatic tour guidance to a large leisure facility such as a theme park. In this problem, the travel time between two nodes depends on the time when the trip starts. We formulate the problem as an integer linear programming (ILP) model. We then develop various heuristics in a step by step fashion: greedy construction, local search and variable neighborhood descent, and two versions of iterated local search. The proposed metaheuristics were tested on modified benchmark instances, randomly …

Diversity-Oriented Bi-Objective Hyper-Heuristics For Patrol Scheduling, Mustafa Misir, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

The patrol scheduling problem is concerned with assigning security teams to different stations for distinct time intervals while respecting a limited number of contractual constraints. The objective is to minimise the total distance travelled while maximising the coverage of the stations with respect to their security requirement levels. This paper introduces a hyper-heuristic strategy focusing on generating diverse solutions for a bi-objective patrol scheduling problem. While a variety of hyper-heuristics have been applied to a large suite of problem domains usually in the form of single-objective optimisation, we suggest an alternative approach for solving the patrol scheduling problem with two …

Unleashing Dec-Mdps In Security Games: Enabling Effective Defender Teamwork, Eric Shieh, Albert Jiang, Amulya Yadav, Pradeep Reddy Varakantham, Milind Tambe

Research Collection School Of Computing and Information Systems

Multiagent teamwork and defender-attacker security games are two areas that are currently receiving significant attention within multiagent systems research. Unfortunately, despite the need for effective teamwork among multiple defenders, little has been done to harness the teamwork research in security games. This paper is the first to remedy this situation by integrating the powerful teamwork mechanisms offered by Dec-MDPs into security games. We offer the following novel contributions in this paper: (i) New models of security games where a defender team’s pure strategy is defined as a DecMDP policy for addressing coordination under uncertainty; (ii) New algorithms based on column …

Interpretable Machine Learning And Sparse Coding For Computer Vision, Will Landecker

Dissertations and Theses

Machine learning offers many powerful tools for prediction. One of these tools, the binary classifier, is often considered a black box. Although its predictions may be accurate, we might never know why the classifier made a particular prediction. In the first half of this dissertation, I review the state of the art of interpretable methods (methods for explaining why); after noting where the existing methods fall short, I propose a new method for a particular type of black box called additive networks. I offer a proof of trustworthiness for this new method (meaning a proof that my method does not …

Element Detection In Japanese Comic Book Panels, Toshihiro Kuboi

Master's Theses

Comic books are a unique and increasingly popular form of entertainment combining visual and textual elements of communication. This work pertains to making comic books more accessible. Specifically, this paper explains how we detect elements such as speech bubbles present in Japanese comic book panels. Some applications of the work presented in this paper are automatic detection of text and its transformation into audio or into other languages. Automatic detection of elements can also allow reasoning and analysis at a deeper semantic level than what’s possible today. Our approach uses an expert system and a machine learning system. The expert …

An Empirical Study Of Off-Line Configuration And On-Line Adaptation In Operator Selection, Zhi Yuan, Stephanus Daniel Handoko, Duc Thien Nguyen, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

Automating the process of finding good parameter settings is important in the design of high-performing algorithms. These automatic processes can generally be categorized into off-line and on-line methods. Off-line configuration consists in learning and selecting the best setting in a training phase, and usually fixes it while solving an instance. On-line adaptation methods on the contrary vary the parameter setting adaptively during each algorithm run. In this work, we provide an empirical study of both approaches on the operator selection problem, explore the possibility of varying parameter value by a non-adaptive distribution tuned off-line, and incorporate the off-line with on-line …

Hybrid Metaheuristics For Solving The Quadratic Assignment Problem And The Generalized Quadratic Assignment Problem, Aldy Gunawan, Kien Ming Ng, Kim Leng Poh, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

This paper presents a hybrid metaheuristic for solving the Quadratic Assignment Problem (QAP). The proposed algorithm involves using the Greedy Randomized Adaptive Search Procedure (GRASP) to construct an initial solution, and then using a hybrid Simulated Annealing and Tabu Search (SA-TS) algorithm to further improve the solution. Experimental results show that the hybrid metaheuristic is able to obtain good quality solutions for QAPLIB test problems within reasonable computation time. The proposed algorithm is extended to solve the Generalized Quadratic Assignment Problem (GQAP), with an emphasis on modelling and solving a practical problem, namely an examination timetabling problem. We found that …

Integrating Motivated Learning And K-Winner-Take-All To Coordinate Multi-Agent Reinforcement Learning, Teck-Hou Teng, Ah-Hwee Tan, Janusz Starzyk, Yuan-Sin Tan, Loo-Nin Teow

Research Collection School Of Computing and Information Systems

This work addresses the coordination issue in distributed optimization problem (DOP) where multiple distinct and time-critical tasks are performed to satisfy a global objective function. The performance of these tasks has to be coordinated due to the sharing of consumable resources and the dependency on non-consumable resources. Knowing that it can be sub-optimal to predefine the performance of the tasks for large DOPs, the multi-agent reinforcement learning (MARL) framework is adopted wherein an agent is used to learn the performance of each distinct task using reinforcement learning. To coordinate MARL, we propose a novel coordination strategy integrating Motivated Learning (ML) …

Lexicon-Based Sentiment Analysis In The Social Web, Fazal Masud Kundi, Dr. Muhammad Zubair Asghar

Dr. Muhammad Zubair Asghar

Sentiment analysis is a compelling issue for both information producers and consumers. We are living in the “age of customer”, where customer knowledge and perception is a key for running successful business. The goal of sentiment analysis is to recognize and express emotions digitally. This paper presents the lexicon-based framework for sentiment classification, which classifies tweets as a positive, negative, or neutral. The proposed framework also detects and scores the slangs used in the tweets. The comparative results show that the proposed system outperforms the existing systems. It achieves 92% accuracy in binary classification and 87% in multi-class classification.