Artificial Intelligence and Robotics Commons^™

Towards Efficient Computation Of Quality Bounded Solutions In Pomdps: Expected Value Approximation And Dynamic Disjunctive Beliefs, Pradeep Reddy Varakantham, Rajiv Maheswaran, Tapana Gupta, Milind Tambe

Research Collection School Of Computing and Information Systems

While POMDPs (partially observable markov decision problems) are a popular computational model with wide-ranging applications, the computational cost for optimal policy generation is prohibitive. Researchers are investigating ever-more efficient algorithms, yet many applications demand such algorithms bound any loss in policy quality when chasing efficiency. To address this challenge, we present two new techniques. The first approximates in the value space to obtain solutions efficiently for a pre-specified error bound. Unlike existing techniques, our technique guarantees the resulting policy will meet this bound. Furthermore, it does not require costly computations to determine the quality loss of the policy. Our second …

Real-Time Supply Chain Control Via Multi-Agent Adjustable Autonomy, Hoong Chuin Lau, Lucas Agussurja, Ramesh Thangarajoo

Research Collection School Of Computing and Information Systems

Real-time supply chain management in a rapidly changing environment requires reactive and dynamic collaboration among participating entities. In this work, we model supply chain as a multi-agent system where agents are subject to an adjustable autonomy. The autonomy of an agent refers to its capability to make and influence decisions within a multi-agent system. Adjustable autonomy means changing the autonomy of the agents during runtime as a response to changes in the environment. In the context of a supply chain, different entities will have different autonomy levels and objective functions as the environment changes, and the goal is to design …

Plans As Products Of Learning, Samin Karim, Budhitama Subagdja, Liz Sonenberg

Research Collection School Of Computing and Information Systems

This paper presents motivations and current related work in the field of plan learning. Additionally, two approaches that achieve plan learning are presented. The two presented approaches are centred on the BDI framework of agency and have particular focus on plans, which, alongside goals, are the means to fulfil intentions in most pragmatic and theoretical realisations of the BDI framework. The first approach is a hybrid architecture that combines a BDI plan extractor and executor with a generic low-level learner. The second approach uses hypotheses to suggest incremental refinements of a priori plans. Both approaches achieve plan generation that is …

A Fuzzy Logic Controller For Autonomous Wheeled Vehicles, Mohamed Trabia, Linda Z. Shi, Neil Eugene Hodge

Mechanical Engineering Faculty Research

Autonomous vehicles have potential applications in many fields, such as replacing humans in hazardous environments, conducting military missions, and performing routine tasks for industry. Driving ground vehicles is an area where human performance has proven to be reliable. Drivers typically respond quickly to sudden changes in their environment. While other control techniques may be used to control a vehicle, fuzzy logic has certain advantages in this area; one of them is its ability to incorporate human knowledge and experience, via language, into relationships among the given quantities. Fuzzy logic controllers for autonomous vehicles have been successfully applied to address various …

Cosign: A Parallel Algorithm For Coordinated Traffic Signal Control, Shih-Fen Cheng, Marina A. Epelman, Robert L. Smith

Research Collection School Of Computing and Information Systems

The problem of finding optimal coordinated signal timing plans for a large number of traffic signals is a challenging problem because of the exponential growth in the number of joint timing plans that need to be explored as the network size grows. In this paper, the game-theoretic paradigm of fictitious play to iteratively search for a coordinated signal timing plan is employed, which improves a system-wide performance criterion for a traffic network. The algorithm is robustly scalable to realistic-size networks modeled with high-fidelity simulations. Results of a case study for the city of Troy, MI, where there are 75 signalized …

Dynamic Multi-Linked Negotiations In Multi-Echelon Production Scheduling Networks, Hoong Chuin Lau, Guan Li Soh, Wee Chong Wan

Research Collection School Of Computing and Information Systems

In this paper, we are concerned with scheduling resources in a multi-tier production/logistics system for multi-indenture goods. Unlike classical production scheduling problems, the problem we study is concerned with local utilities which are private. We present an agent model and investigate an efficient scheme for handling multi-linked agent negotiations. With this scheme we attempt to overcome the drawbacks of sequential negotiations and negotiation parameter settings. Our approach is based on embedding a credit-based negotiation protocol within a local search scheduling algorithm. We demonstrate the computational efficiency and effectiveness of the approach in solving a real-life dynamic production scheduling problem which …

Robust Controllability In Temporal Constraint Networks Under Uncertainty, Hoong Chuin Lau, Jia Li, Roland H. C. Yap

Research Collection School Of Computing and Information Systems

Temporal constraint networks are embedded in many planning and scheduling problems. In dynamic problems, a fundamental challenge is to decide whether such a network can be executed as uncertainty is revealed over time. Very little work in this domain has been done in the probabilistic context. In this paper, we propose a Temporal Constraint Network (TCN) model where durations of uncertain activities are represented by random variables. We wish to know whether such a network is robust controllable, i.e. can be executed dynamically within a given failure probability, and if so, how one might find a feasible schedule as the …

Pedagogical Possibilities For The N-Puzzle Problem, Zdravko Markov, Ingrid Russell, Todd W. Neller, Neli Zlatareva

Computer Science Faculty Publications

In this paper we present work on a project funded by the National Science Foundation with a goal of unifying the Artificial Intelligence (AI) course around the theme of machine learning. Our work involves the development and testing of an adaptable framework for the presentation of core AI topics that emphasizes the relationship between AI and computer science. Several hands-on laboratory projects that can be closely integrated into an introductory AI course have been developed. We present an overview of one of the projects and describe the associated curricular materials that have been developed. The project uses machine learning as …

Viz: A Visual Analysis Suite For Explaining Local Search Behavior, Steven Halim, Roland H. C. Yap, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

NP-hard combinatorial optimization problems are common in real life. Due to their intractability, local search algorithms are often used to solve such problems. Since these algorithms are heuristic-based, it is hard to understand how to improve or tune them. We propose an interactive visualization tool, VIZ, meant for understanding the behavior of local search. VIZ uses animation of abstract search trajectories with other visualizations which are also animated in a VCR-like fashion to graphically playback the algorithm behavior. It combines generic visualizations applicable on arbitrary algorithms with algorithm and problem specific visualizations. We use a variety of techniques such as …

Two-Instant Reallocation In Two-Echelon Spare Parts Inventory Systems, Huawei Song, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

In this paper, we study the problem of deciding when and how to perform reallocation of existing spare parts in a multi-echelon reparable item inventory system. We present a mathematical model that solves the problem when there are two reallocation instants, in response to the open challenge post by Cao and Silver(2005) to consider two or more possible reallocations within a replenishment cycle.

Multi-Learner Based Recursive Supervised Training, Laxmi R. Iyer, Kiruthika Ramanathan, Sheng-Uei Guan

Research Collection School Of Computing and Information Systems

In this paper, we propose the multi-learner based recursive supervised training (MLRT) algorithm, which uses the existing framework of recursive task decomposition, by training the entire dataset, picking out the best learnt patterns, and then repeating the process with the remaining patterns. Instead of having a single learner to classify all datasets during each recursion, an appropriate learner is chosen from a set of three learners, based on the subset of data being trained, thereby avoiding the time overhead associated with the genetic algorithm learner utilized in previous approaches. In this way MLRT seeks to identify the inherent characteristics of …

Learning As A Nonlinear Line Of Attraction For Pattern Association, Classification And Recognition, Ming-Jung Seow

Electrical & Computer Engineering Theses & Dissertations

Development of a mathematical model for learning a nonlinear line of attraction is presented in this dissertation, in contrast to the conventional recurrent neural network model in which the memory is stored in an attractive fixed point at discrete location in state space. A nonlinear line of attraction is the encapsulation of attractive fixed points scattered in state space as an attractive nonlinear line, describing patterns with similar characteristics as a family of patterns.

It is usually of prime imperative to guarantee the convergence of the dynamics of the recurrent network for associative learning and recall. We propose to alter …

Socio-Economic Impacts Of Computer Viruses In Tanzania, M Victor

Tanzania Journal of Engineering and Technology (TJET)

This paper reports on a research project conducted with an objective of identifying and assessing various approaches used by different computer users (Management, System Administrators and end users) in Tanzania to combat computer viruses (CVs), and to assess users' awareness level on CVs. Specifically, the study aimed at assessing the awareness level on CVs to the Tanzanian business community; analyze the socio -economic impact caused by CVs in Tanzania and; assess existing methods, capacity and limitations on controlling CVs in Tanzania. Data was collected using both questionnaires and interview from financial institutions such as NBC and BOT, and telecommunications sector …

Fuzzy Cognitive Goal Net For Interactive Storytelling Plot Design, Yundong Cai, Chunyan Miao, Ah-Hwee Tan, Zhiqi Shen

Research Collection School Of Computing and Information Systems

Interactive storytelling attracts a lot of research interests among the interactive entertainments in recent years. Designing story plot for interactive storytelling is currently one of the most critical problems of interactive storytelling. Some traditional AI planning methods, such as Hierarchical Task Network, Heuristic Searching Method are widely used as the planning tool for the story plot design. This paper proposes a model called Fuzzy Cognitive Goal Net as the story plot planning tool for interactive storytelling, which combines the planning capability of Goal net and reasoning ability of Fuzzy Cognitive Maps. Compared to conventional methods, the proposed model shows a …

Fuzzy State Aggregation And Off-Policy Reinforcement Learning For Stochastic Environments, Dean C. Wardell, Gilbert L. Peterson

Faculty Publications

Reinforcement learning is one of the more attractive machine learning technologies, due to its unsupervised learning structure and ability to continually learn even as the environment it is operating in changes. This ability to learn in an unsupervised manner in a changing environment is applicable in complex domains through the use of function approximation of the domain’s policy. The function approximation presented here is that of fuzzy state aggregation. This article presents the use of fuzzy state aggregation with the current policy hill climbing methods of Win or Lose Fast (WoLF) and policy-dynamics based WoLF (PD-WoLF), exceeding the learning rate …

Winning Back The Cup For Distributed Pomdps: Planning Over Continuous Belief Spaces, Pradeep Varakantham, Ranjit Nair, Milind Tambe, Makoto Yokoo

Research Collection School Of Computing and Information Systems

Distributed Partially Observable Markov Decision Problems (Distributed POMDPs) are evolving as a popular approach for modeling multiagent systems, and many different algorithms have been proposed to obtain locally or globally optimal policies. Unfortunately, most of these algorithms have either been explicitly designed or experimentally evaluated assuming knowledge of a starting belief point, an assumption that often does not hold in complex, uncertain domains. Instead, in such domains, it is important for agents to explicitly plan over continuous belief spaces. This paper provides a novel algorithm to explicitly compute finite horizon policies over continuous belief spaces, without restricting the space of …

Mobius: An Omnidirectional Robotic Platform And Software Architecture For Network Teleoperation, Samuel Aaron Miller

Electrical & Computer Engineering Theses & Dissertations

The following thesis presents the results of a project to develop and test an omnidirectional robotic system (hardware and software) at NASA Langley Research Center's Robotics and Intelligent Machines Lab. The impetus for the project was the unique capabilities of omnidirectional systems. Some of the many potential benefits these systems have include improved material-handling capabilities in constrained environments (such as might be found in extraterrestrial manned habitats), efficient camera-based vehicle teleoperation, and simplified route planning for autonomous robot operations.

The project's focus was to design, build, and test a system that used Mecanum wheels to achieve omnidirectional motion. In addition …

Evaluation Of Time-Varying Availability In Multi-Echelon Spare Parts Systems With Passivation, Hoong Chuin Lau, Huawei Song, Chuen Teck See, Siew Yen Cheng

Research Collection School Of Computing and Information Systems

The popular models for repairable item inventory, both in the literature as well as practical applications, assume that the demands for items are independent of the number of working systems. However this assumption can introduce a serious underestimation of availability when the number of working systems is small, the failure rate is high or the repair time is long. In this paper, we study a multi-echelon repairable item inventory system under the phenomenon of passivation, i.e. serviceable items are passivated (“switched off”) upon system failure. This work is motivated by corrective maintenance of high-cost technical equipment in the miltary. We …

Application Of Fuzzy State Aggregation And Policy Hill Climbing To Multi-Agent Systems In Stochastic Environments, Dean C. Wardell

Theses and Dissertations

Reinforcement learning is one of the more attractive machine learning technologies, due to its unsupervised learning structure and ability to continually even as the operating environment changes. Applying this learning to multiple cooperative software agents (a multi-agent system) not only allows each individual agent to learn from its own experience, but also opens up the opportunity for the individual agents to learn from the other agents in the system, thus accelerating the rate of learning. This research presents the novel use of fuzzy state aggregation, as the means of function approximation, combined with the policy hill climbing methods of Win …

A Monocular Vision Based Approach To Flocking, Brian Kirchner

Theses and Dissertations

Flocking is seen in nature as a means for self protection, more efficient foraging, and other search behaviors. Although much research has been done regarding the application of this principle to autonomous vehicles, the majority of the research has relied on GPS information, broadcast communication, an omniscient central controller, or some other form of "global" knowledge. This approach, while effective, has serious drawbacks, especially regarding stealth, reliability, and biological grounding. This research effort uses three Pioneer P2-AT8 robots to achieve flocking behavior without the use of global knowledge. The sensory inputs are limited to two cameras, offset such that the …

Proximity In Context: An Empirically Grounded Computational Model Of Proximity For Processing Topological Spatial Expression., John D. Kelleher, Geert-Jan Kruijff, Fintan Costello

Conference papers

The paper presents a new model for context-dependent interpretation of linguistic expressions about spatial proximity between objects in a natural scene. The paper discusses novel psycholinguistic experimental data that tests and verifies the model. The model has been implemented, and enables a conversational robot to identify objects in a scene through topological spatial relations (e.g. ''X near Y''). The model can help motivate the choice between topological and projective prepositions.

Incremental Generation Of Spatial Referring Expressions In Situated Dialogue, John D. Kelleher, Geert-Jan Kruijff

Conference papers

This paper presents an approach to incrementally generating locative expressions. It addresses the issue of combinatorial explosion inherent in the construction of relational context models by: (a) contextually defining the set of objects in the context that may function as a landmark, and (b) sequencing the order in which spatial relations are considered using a cognitively motivated hierarchy of relations, and visual and discourse salience.

A Computational Model Of The Referential Semantics Of Projective Prepositions, John D. Kelleher, Josef Van Genabith

Conference papers

In this paper we present a framework for interpreting locative expressions containing the prepositions in front of and behind. These prepositions have different semantics in the viewer-centred and intrinsic frames of reference (Vandeloise, 1991). We define a model of their semantics in each frame of reference. The basis of these models is a novel parameterized continuum function that creates a 3-D spatial template. In the intrinsic frame of reference the origin used by the continuum function is assumed to be known a priori and object occlusion does not impact on the applicability rating of a point in the spatial template. …

Visualization For Analyzing Trajectory-Based Metaheuristic Search Algorithms, Steven Halim, Roland H. C. Yap, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

No abstract provided.

The History Of Computer Games, Jill Cirasella, Danny Kopec

Publications and Research

This handout presents milestones in the history of computer backgammon, computer bridge, computer checkers, computer chess, computer Go, computer Othello, and computer poker.

Multiagent Teamwork: Hybrid Approaches, Praveen Paruchuri, Emma Bowring, Ranjit Nair, Jonathan Pearce, Nathan Schurr, Milind Tambe, Pradeep Varakantham

Research Collection School Of Computing and Information Systems

Today within the multiagent community, we see at least four competing methods to building multiagent systems: beliefdesireintention (BDI), distributed constraint optimization (DCOP), distributed POMDPs, and auctions or game-theoretic methods. While there is exciting progress within each approach, there is a lack of cross-cutting research. This article highlights the various hybrid techniques for multiagent teamwork developed by the teamcore group. In particular, for the past decade, the TEAMCORE research group has focused on building agent teams in complex, dynamic domains. While our early work was inspired by BDI, we will present an overview of recent research that uses DCOPs and distributed …

Transition Rule Elicitation Methods For Urban Cellular Automata Models, Junfeng Jiao, Luc Boerboom

Research Collection College of Integrative Studies

In this chapter, transition rules used in urban CA models are reviewed and classified into two categories: transition potential rules and conflict resolution rules. Then, four widely used rule elicitation methods: Regression analysis, Artificial Neural network (ANN), Visual calibration, and Analytical Hierarchy Processing – Multi Criteria Evaluation (AHP-MCE) are discussed. Most of these methods are data driven methods and can be used to elicit the transition potential rules in the urban CA models. In the following, three possible rule elicitation methods: Interview, Document analysis, and Card sorting are explained and demonstrated. These three methods are driven by knowledge and can …

Hybrid Committee Classifier For A Computerized Colonic Polyp Detection System, Jiang Li, Jianhua Yao, Nicholas Petrick, Ronald M. Summers, Amy K. Hara, Joseph M. Reinhardt (Ed.), Josien P.W. Pluim (Ed.)

Electrical & Computer Engineering Faculty Publications

We present a hybrid committee classifier for computer-aided detection (CAD) of colonic polyps in CT colonography (CTC). The classifier involved an ensemble of support vector machines (SVM) and neural networks (NN) for classification, a progressive search algorithm for selecting a set of features used by the SVMs and a floating search algorithm for selecting features used by the NNs. A total of 102 quantitative features were calculated for each polyp candidate found by a prototype CAD system. 3 features were selected for each of 7 SVM classifiers which were then combined to form a committee of SVMs classifier. Similarly, features …

A Generic Object-Oriented Tabu Search Framework, Hoong Chuin Lau, Xiaomin Jia, Wee Chong Wan

Research Collection School Of Computing and Information Systems

Presently, most tabu search designers devise their applications without considering the potential of design and code reuse, which consequently prolong the development of subsequent applications. In this paper, we propose a software solution known as Tabu Search Framework (TSF), which is a generic C++ software framework for tabu search implementation. The framework excels in code recycling through the use of a well- designed set of generic abstract classes that clearly define their collaborative roles in the algorithm. Additionally, the framework incorporates a centralized process and control mechanism that enhances the search with intelligence. This results in a generic framework that …

Integration Of Probabilistic Graphic Models For Decision Support, Jiang C., Poh K., Tze-Yun Leong

Research Collection School Of Computing and Information Systems

It is a frequently encountered problem that new knowledge arrived when making decisions in a dynamic world. Usually, domain experts cannot afford enough time and knowledge to effectively assess and combine both qualitative and quantitative information in these models. Existing approaches can solve only one of two tasks instead of both. We propose a four-step algorithm to integrate multiple probabilistic graphic models, which can effectively update existing models with newly acquired models. In this algorithm, the qualitative part of model integration is performed first, followed by the quantitative combination. We illustrate our method with an example of combining three models. …