Physical Sciences and Mathematics Commons^™

A Comparison Of Clustering Techniques For Malware Analysis, Swathi Pai

Master's Projects

In this research, we apply clustering techniques to the malware detection problem. Our goal is to classify malware as part of a fully automated detection strategy. We compute clusters using the well-known �-means and EM clustering algorithms, with scores obtained from Hidden Markov Models (HMM). The previous work in this area consists of using HMM and �-means clustering technique to achieve the same. The current effort aims to extend it to use EM clustering technique for detection and also compare this technique with the �-means clustering.

Clustering Versus Svm For Malware Detection, Usha Narra

Master's Projects

Previous work has shown that we can effectively cluster certain classes of mal- ware into their respective families. In this research, we extend this previous work to the problem of developing an automated malware detection system. We first compute clusters for a collection of malware families. Then we analyze the effectiveness of clas- sifying new samples based on these existing clusters. We compare results obtained using �-means and Expectation Maximization (EM) clustering to those obtained us- ing Support Vector Machines (SVM). Using clustering, we are able to detect some malware families with an accuracy comparable to that of SVMs. One …

Using Neural Networks For Image Classification, Tim Kang

Master's Projects

This paper will focus on applying neural network machine learning methods to images for the purpose of automatic detection and classification. The main advantage of using neural network methods in this project is its adeptness at fitting non­linear data and its ability to work as an unsupervised algorithm. The algorithms will be run on common, publically available datasets, namely the MNIST and CIFAR­10, so that our results will be easily reproducible.

Video Event Understanding With Pattern Theory, Fillipe Souza, Sudeep Sarkar, Anuj Srivastava, Jingyong Su

MODVIS Workshop

We propose a combinatorial approach built on Grenander’s pattern theory to generate semantic interpretations of video events of human activities. The basic units of representations, termed generators, are linked with each other using pairwise connections, termed bonds, that satisfy predefined relations. Different generators are specified for different levels, from (image) features at the bottom level to (human) actions at the highest, providing a rich representation of items in a scene. The resulting configurations of connected generators provide scene interpretations; the inference goal is to parse given video data and generate high-probability configurations. The probabilistic structures are imposed using energies that …

Two Correspondence Problems Easier Than One, Aaron Michaux, Zygmunt Pizlo

MODVIS Workshop

Computer vision research rarely makes use of symmetry in stereo reconstruction despite its established importance in perceptual psychology. Such stereo reconstructions produce visually satisfying figures with precisely located points and lines, even when input images have low or moderate resolution. However, because few invariants exist, there are no known general approaches to solving symmetry correspondence on real images. The problem is significantly easier when combined with the binocular correspondence problem, because each correspondence problem provides strong non-overlapping constraints on the solution space. We demonstrate a system that leverages these constraints to produce accurate stereo models from pairs of binocular images …

Formal Aspects Of Non-Rigid-Shape-From-Motion Perception, Vicky Froyen, Qasim Zaidi

MODVIS Workshop

Our world is full of objects that deform over time, for example animals, trees and clouds. Yet, the human visual system seems to readily disentangle object motions from non-rigid deformations, in order to categorize objects, recognize the nature of actions such as running or jumping, and even to infer intentions. A large body of experimental work has been devoted to extracting rigid structure from motion, but there is little experimental work on the perception of non-rigid 3-D shapes from motion (e.g. Jain, 2011). Similarly, until recently, almost all formal work had concentrated on the rigid case. In the last fifteen …

Object Recognition And Visual Search With A Physiologically Grounded Model Of Visual Attention, Frederik Beuth, Fred H. Hamker

MODVIS Workshop

Visual attention models can explain a rich set of physiological data (Reynolds & Heeger, 2009, Neuron), but can rarely link these findings to real-world tasks. Here, we would like to narrow this gap with a novel, physiologically grounded model of visual attention by demonstrating its objects recognition abilities in noisy scenes.

To base the model on physiological data, we used a recently developed microcircuit model of visual attention (Beuth & Hamker, in revision, Vision Res) which explains a large set of attention experiments, e.g. biased competition, modulation of contrast response functions, tuning curves, and surround suppression. Objects are represented by …

Modeling Visual Features To Recognize Biological Motion: A Developmental Approach, Giulio Sandini, Nicoletta Noceti, Alessia Vignolo, Alessandra Sciutti, Francesco Rea, Alessandro Verri, Francesca Odone

MODVIS Workshop

In this work we deal with the problem of designing and developing computational vision models – comparable to the early stages of the human development – using coarse low-level information.

More specifically, we consider a binary classification setting to characterize biological movements with respect to non-biological dynamic events. To this purpose, our model builds on top of the optical flow estimation, and abstract the representation to simulate the limited amount of visual information available at birth. We take inspiration from known biological motion regularities explained by the Two-Thirds Power Law, and design a motion representation that includes different low-level features, …

Comparative Analysis Of Particle Swarm Optimization Algorithms For Text Feature Selection, Shuang Wu

Master's Projects

With the rapid growth of Internet, more and more natural language text documents are available in electronic format, making automated text categorization a must in most fields. Due to the high dimensionality of text categorization tasks, feature selection is needed before executing document classification. There are basically two kinds of feature selection approaches: the filter approach and the wrapper approach. For the wrapper approach, a search algorithm for feature subsets and an evaluation algorithm for assessing the fitness of the selected feature subset are required. In this work, I focus on the comparison between two wrapper approaches. These two approaches …

Using Hidden Markov Models To Detect Dna Motifs, Santrupti Nerli

Master's Projects

During the process of gene expression in eukaryotes, mRNA splicing is one of the key processes carried out by a complex called spliceosome. Spliceosome guarantees proper removal of introns and joining of exons before the translation process. Precise splicing is essential for the production of functional proteins. Spliceosome detects specific sequence motifs within an mRNA sequence called splice sites. Two of the splice sites are the 5’ and 3’ sites that border all the introns. Normal splicing process if disrupted by mutation may lead to fatal diseases. In this work, we predict splice sites in a human genome using hidden …

Using Probabilistic Graphical Models To Solve Np-Complete Puzzle Problems, Fengjiao Wu

Master's Projects

Probabilistic Graphical Models (PGMs) are commonly used in machine learning to solve problems stemming from medicine, meteorology, speech recognition, image processing, intelligent tutoring, gambling, games, and biology. PGMs are applicable for both directed graph and undirected graph. In this work, I focus on the undirected graphical model. The objective of this work is to study how PGMs can be applied to find solutions to two puzzle problems, sudoku and jigsaw puzzles. First, both puzzle problems are represented as undirected graphs, and then I map the relations of nodes to PGMs and Belief Propagation (BP). This work represents the puzzle grid …

Neuroscience-Inspired Dynamic Architectures, Catherine Dorothy Schuman

Doctoral Dissertations

Biological brains are some of the most powerful computational devices on Earth. Computer scientists have long drawn inspiration from neuroscience to produce computational tools. This work introduces neuroscience-inspired dynamic architectures (NIDA), spiking neural networks embedded in a geometric space that exhibit dynamic behavior. A neuromorphic hardware implementation based on NIDA networks, Dynamic Adaptive Neural Network Array (DANNA), is discussed. Neuromorphic implementations are one alternative/complement to traditional von Neumann computation. A method for designing/training NIDA networks, based on evolutionary optimization, is introduced. We demonstrate the utility of NIDA networks on classification tasks, a control task, and an anomaly detection task. There …

Using Genetic Learning In Weight-Based Game Ai, Dylan Anthony Kordsmeier

Computer Science and Computer Engineering Undergraduate Honors Theses

Human beings have been playing games for centuries, and over time, mankind has learned how to excel at these fun competitions. With the ever-growing interest in the field of Machine Learning and Artificial Intelligence (AI), developers have been finding ways to let the game compete against the player much like another human would. While there are many approaches to humanlike learning in machines, this article will focus on using Evolutionary Optimization as a method to develop different levels of pseudo-thinking inan AI used for ato effectively play the Connect Four game.

Near-Optimal Decentralized Power Supply Restoration In Smart Grids, Pritee Agrawal, Akshat Kumar, Pradeep Varakantham

Research Collection School Of Computing and Information Systems

Next generation of smart grids face a number of challenges including co-generation from intermittent renewable power sources, a shift away from monolithic control due to increased market deregulation, and robust operation in the face of disasters. Such heterogeneous nature and high operational readiness requirement of smart grids necessitates decentralized control for critical tasks such as power supply restoration (PSR) after line failures. We present a novel multiagent system based approach for PSR using Lagrangian dual decomposition. Our approach works on general graphs, provides provable quality-bounds and requires only local message-passing among different connected sub-regions of a smart grid, enabling decentralized …

Oscar: Online Selection Of Algorithm Portfolios With Case Study On Memetic Algorithms, Mustafa Misir, Stephanus Daniel Handoko, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

This paper introduces an automated approach called OSCAR that combines algorithm portfolios and online algorithm selection. The goal of algorithm portfolios is to construct a subset of algorithms with diverse problem solving capabilities. The portfolio is then used to select algorithms from for solving a particular (set of) instance(s). Traditionally, algorithm selection is usually performed in an offline manner and requires the need of domain knowledge about the target problem; while online algorithm selection techniques tend not to pay much attention to a careful construction of algorithm portfolios. By combining algorithm portfolios and online selection, our hope is to design …

Direct: A Scalable Approach For Route Guidance In Selfish Orienteering Problems, Pradeep Varakantham, Hala Mostafa, Na Fu, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

We address the problem of crowd congestion at venues like theme parks, museums and world expos by providing route guidance to multiple selfish users (with budget constraints) moving through the venue simultaneously. To represent these settings, we introduce the Selfish Orienteering Problem (SeOP) that combines two well studied problems from literature, namely Orienteering Problem (OP) and Selfish Routing (SR). OP is a single agent routing problem where the goal is to minimize latency (or maximize reward) in traversing a subset of nodes while respecting budget constraints. SR is a game between selfish agents looking for minimum latency routes from source …

Adviser: A Web-Based Algorithm Portfolio Deviser, Mustafa Misir, Stephanus Daniel Handoko, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

The basic idea of algorithm portfolio [1] is to create a mixture of diverse algorithms that complement each other’s strength so as to solve a diverse set of problem instances. Algorithm portfolios have taken on a new and practical meaning today with the wide availability of multi-core processors: from an enterprise perspective, the interest is to make best use of parallel machines within the organization by running different algorithms simultaneously on different cores to solve a given problem instance. Parallel execution of a portfolio of algorithms as suggested by [2, 3] a number of years …

Predicting Bundles Of Spatial Locations From Learning Revealed Preference Data, Truc Viet Le, Siyuan Liu, Hoong Chuin Lau, Ramayya Krishnan

Research Collection School Of Computing and Information Systems

We propose the problem of predicting a bundle of goods, where the goods considered is a set of spatial locations that an agent wishes to visit. This typically arises in the tourism setting where attractions can often be bundled and sold as a package to visitors. While the problem of predicting future locations given the current and past trajectories is well-established, we take a radical approach by looking at it from an economic point of view. We view an agent's past trajectories as revealed preference (RP) data, where the choice of locations is a solution to an optimisation problem according …

An Approach To Artificial Society Generation For Video Games, Bryan Sarlo

Electronic Thesis and Dissertation Repository

Since their inception in the 1940s, video games have always had a need for non-player characters (NPCs) driven by some form of artificial intelligence (AI). More recently, researchers and developers have attempted to create believable, or human-like, agents by modeling them after humans by borrowing concepts from the social sciences. This thesis explores an approach to generating a society of such believable agents with human-like attributes and social connections. This approach allows agents to form various kinds of relationships with other agents in the society, and even provides an introductory form of shared or influenced attributes based on their spouse …

A Behavior-Reactive Autonomous System To Identify Pokémon Characters, Xu Cao, Bohan Zhang, Jeremy Straub, Eunjin Kim

Jeremy Straub

Pokémon is an entertainment franchise with a large fan base. This project uses well-known Pokémon characters to demonstrate the operations of a question selection system. Presented in the form of a game where the computer attempts to guess the user-selected character, the system attempts to minimize the number of questions required for this purpose by identifying questions that most constrain the decision space. The decision making process is refined based on actual user behavior.

Supervisory Control And Data Acquisition (Scada) Control Optimization, Garrett Johnson, Jeremy Straub, Eunjin Kim

Jeremy Straub

SCADA systems are generally used to monitor and control multiple systems of the same type to allow them to be remotely controlled and monitored. Water plants, for example, could be controlled and monitored by a SCADA system. This project seeks to optimize a SCADA system using Artificial Intelligence. A constraint satisfaction / optimization algorithm is used to maximize performance relative to weighted system goals.

Scheduling Algorithm Development For An Open Source Software And Open Hardware Spacecraft, Calvin Bina, Jeremy Straub, Ronald Marsh

Jeremy Straub

The efficacy of each type of scheduler is assessed rela-tive to the goal of having a time and resource efficient scheduling algorithm. The scheduler must ensure suc-cessful spacecraft operations and maximize the perfor-mance of tasks relative to performance constraints and their respective due dates.

Scada System Security: Accounting For Operator Error And Malicious Intent, Ryan Kilbride, Jeremy Straub, Eunjin Kim

Jeremy Straub

Supervisory control and data acquisition (SCADA) systems are becoming more and more com-monplace in many industries today. Industries are making better use of software and large scale control systems to run efficiently, without the need for large amounts of oversight. Security is a particularly large issue with such systems, however. A human must still be involved to ensure smooth operation in the event of catastrophic system error, or unusual circumstanc-es. Human involvement presents problems: operators could make mistakes, configure the system to operate sub-optimally or take malicious actions. This imple-mentation of SCADA security aims to combat these problems.

Artificial Intelligence Animal Recognition System, Bohan Zhang, Xu Cao, Jeremy Straub, Eunjin Kim

Jeremy Straub

Artificial Intelligence Animal recognition system can be widely used for chil-dren education, zoology database implement. It is known that existing animal classification system can be used to distinguish an animal fairly fast. For example mammals, reptiles, amphibians. Our goal is to implement an AI system which could eliminate the possibility by half or more for which animal the user is thinking by means of asking the player "the best" question for an animal property.

Medical Procedure Expert System, Timothy Whitney, Jeremy Straub, Eunjin Kim

Jeremy Straub

Artificial intelligence is an area of research in computer science that allows for computer systems to make logical decisions based on its environment. This basic ability can be applied to many problems, including detecting abnormalities in a system, such as commands issued by malicious users, or erroneous commands from valid users. The ability to detect such commands is particularly useful in high risk application where such a command could result in harm to people. This research focuses on one such area by applying the decision making power of AI towards a medical use by developing an expert system to detect …

Pattern Recognition And Expert Systems For Microwave Wireless Power Transmission Failure Prevention, Cameron Kerbaugh, Allen Mcdermott, Jeremy Straub, Eunjin Kim

Jeremy Straub

Wireless power transfer (WPT) can be used to deliver space-generated power to ground stations through the use of microwave beams. WPT satellite power delivery systems have two major failure states: misdi-recting a beam and failing to send power to a station. This project has implemented an expert system to perform pattern recognition in an effort to prevent failures by analyzing the system state and predicting potential failures before they happen in support of space-based testing [1] and deployment [2].

Course Outcome Prediction Using An Expert System, Michael Kuehn, Jared Estad, Jeremy Straub, Eunjin Kim

Jeremy Straub

Determining how well a student will perform in a course based on their prior knowledge of the course material and other factors may help determine student placement and the need for remedial instruction. This poster presents work on the creation of an expert system that attempts to predict a student’s performance based on a pre-evaluation test and responses to background preparation questions. This work utilizes data from prior students to train and test the system.

Pattern Recognition For Detecting Failures In Space Solar Power Systems, Allen Mcdermott, Cameron Kerbaugh, Jeremy Straub, Eunjin Kim

Jeremy Straub

This poster covers work relating to the use of expert systems and pattern recognition to attempt to identify, detect and prospectively stop patterns of activity that could potentially lead to failure of a space solar power (SSP) system. A database-based expert system has is presented to identify patterns, which can be used to determine whether a power beam could hit a unintend- ed target and potentially cause a calamity. This has been implemented via a facts-rule network via which supplied and collected facts and a rule set is used to de- termine whether the system is operating correctly (from a …

Geological Object Recognition In Extraterrestrial Environments, Gregory M. Elfers

Electronic Thesis and Dissertation Repository

On July 4 1997, the landing of NASA’s Pathnder probe and its rover Sojourner marked the beginning of a new era in space exploration; robots with the ability to move have made up the vanguard of human extraterrestrial exploration ever since. With Sojourners landing, for the rst time, a ground traversing robot was at a distance too far from earth to make direct human control practical. This has given rise to the development of autonomous systems to improve the e?ciency of these robots,in both their ability to move,and their ability to make decisions regarding their environment. Computer Vision comprises a …

Prepositional Phrase Attachment Problem Revisited: How Verbnet Can Help, Dan Bailey, Yuliya Lierler, Benjamin Susman

Computer Science Faculty Proceedings & Presentations

Resolving attachment ambiguities is a pervasive problem in syntactic analysis. We propose and investigate an approach to resolving prepositional phrase attachment that centers around the ways of incorporating semantic knowledge derived from the lexico-semantic ontologies such as VERBNET and WORDNET.