Welcome To CIS Department of University Of Massachussetts
Computer Science Seminar Series - Spring 2008
Open dates: 3/28, 4/11
Cumulative Subgoal Fulfillment in Software
in Optical Networks
Defense: The study of Initial Commitment Decision Problem and
Crisis Management in Multi-Agent Virtual Organizations
Cooperative Salvo for Autonomous Underwater Vehicle
of Inter-organizational Loosely Coupled Workflows
Processing and Storage for Scientific and
Senior project / MS project presentations
Senior project / MS project presentations
Title: Cumulative Subgoal Fulfillment in Software
This paper introduces a new approach to programming that
facilitates the construction and verification of applications.
Called "Cumulative Subgoal Fulfillment" (CSF), it implements procedures
by fulfilling and accumulating a sufficient set of subgoals. Each
code block fulfilling a subgoal must keep invariant all subgoals
already fulfilled. Using well-known programming problems, the paper
demonstrates that CSF is new and that its simple form facilitates both
the creation and checking of code.
Eric Braude has a Ph. D. from Columbia University in
mathematics and Master's in Computer Science from the University of
Miami. He taught at CUNY and Penn State, followed by twelve years in
government and industry as a software engineer, scientist, and manager.
He has been an Associate Professor of Computer Science at Boston
University's Metropolitan College since 1990, holding the chairmanship
for five years and an acting associate deanship for one. His papers
deal with simulation, compilers, program correctness, and set theory.
Eric has written or edited five books, including "Software Engineering"
and "Software Design." 2/22/08
Master Thesis Presentation: Automatic Agent Generation and Coordination
in Multi-Agent systems
An open multi-agent system is a dynamic system in which agents can join
or leave an agent society at will and also take or release roles at
runtime. Traditional ways of developing multi-agent systems assume a
fixed number of agents in a multi-agent system and use role modeling
for system analysis to determine the number of agents, which is
unsuitable for developing open multi-agent systems. In our case, role
models are only used at conceptual level with no realizations in the
implementation system. Role organization and role space are specified
as containers of conceptual roles and role instances, respectively.
Agents are defined as members of an agent society and they can take or
release role instances from a role space dynamically. We present an
automatic agent generation process using the Java Agent Framework and
define coordination rules between agents. We describe the design of
agents and agent’s motivations within such framework. We introduce a
practical approach for modeling agent’s motivation and specifying
agent’s goals, where a role-agent mapping mechanism is developed based
on this design. Dynamic task allocation is achieved through the
creation of role instances and the mapping from role instances to
agents. We also introduce the RTÆMS language based on the
extension of TÆMS to model the plan tree for each goal. This
representation enables the reuse of general planning/scheduling and
collaboration/cooperation mechanisms developed in multi-agent system
research community. We have developed an automatic agent generation
interface, defined coordination rules and also implemented a simple
demo system in health care domain.
Computer Science Seminar Series - Fall 2007
The Next Generation of Content Repositories
Prof. Meinolf Sellmann
Combinatorial optimization and constraint
12:30 pm - 1:30 pm
On iterons of automata
of Technology, Poznan, Poland
Dr. Chris Jones
iRobot Corporation R&D Overview
Prof. Iren Valova
Olafctory bulb neurotransmitter studies and
Students - Research Projects
Vishen - Vision Based Docking System for Autonomous Mobile Robots
2. Sravan K.
Patti - Implementing Ontology-Based Dynamic Service Discovery in
Bates, Title: Rule-Based Specification of Flexible Bidding Strategies
in Agent-Based Online Auctions
2. Ankit Goel, Title: Evaluation of a Bayesian Network with Multi-State
Nodes for Shill Verification in Online Auction
3. Danielle Stack - Building Tutorial Agent With Knowledge Base
Title: The Next Generation of Content Repositories
The world is changing. Arguably our lives, both personally and
professionally, are invaded and surrounded by more digital data than
ever before. Whether you look at the entertainment sector, the
communications sector, the healthcare or finance sector, encompassing
digital information is directing our behaviors. Our digital information
is increasingly more valuable as physical counterparts are fading in
favor of online fixed content. And the devices that generate that
digital information are not limited to the traditional devices of
yesterday. Today, information may be captured and retrieved at diverse
locations; by convergent technologies. The converged cell
phone/PDA/MP3-4 player will become a common information access device.
And just as technologies converge, so do the information management
models of the consumer and enterprise world as well. We are at the cusp
of an age where a person will be completely connected to the
information that enterprises generate on the individual's behalf. For
example, the digital information contained in an MRI will be
effortlessly transferable to any specialist based on permission of the
This talk explores the implications of this convergence - one that
demands new categories of content repositories that guarantee content
authenticity, leverage metadata to classify data enabling its
preservation in a self-describing way and support standardization so
that this information can migrate between storage devices from
different vendors, across different technologies.
Michael Kilian is the CTO for the Centera division at
He is responsible for product strategy including customer requirements
gathering and feature design. Michael's team regularly works on
everything from product architecture to projects like the creation of
cell phone that automatically downloads photos to a Centera.
Previous to his current role, Michael reported to EMC's Chief
Technology Officer as a Senior Technologist analyzing technology and
markets of the storage industry at large. Michael built the strategy
that led to the acquisition of FilePool, the company that has written
the software that forms the core of the Centera platform.
Michael joined EMC in the spring of 1997. Before EMC, Michael
developed object oriented programming tools and an object-oriented
database for Kendall Square Research's massively parallel computer.
Michael also worked at Digital Equipment Corporation's Eastern Research
Lab from 1985 to 1992.
Michael Kilian holds a Ph.D. from Harvard University in Computer
Science with a thesis on object-oriented programming methodologies for
massively parallel computers. Centera uniquely blends object-oriented
approaches to storage using a highly parallel platform.
Title: Deterministic Inference in Constraint
Based on the constraint programming paradigm, we illustrate
algorithmic challenges when designing filtering methods for two
constraints of high practical relevance.
The first is the context-free grammar constraint that requires that an
assignment of values to an ordered set of variables must form a word in
a given context-free language. For this constraint, we devise an
efficient, complete filtering algorithm that has the same asymptotic
complexity as the Cocke-Younger-Kasami algorithm for parsing
The second constraint that we present in this talk is the Knapsack
constraint, for which complete filtering is intractable. By reverting
to a relaxed notion of consistency, we develop an efficient incremental
filtering algorithm that runs in expected sub-linear time for several
interesting input distributions. After a theoretical study, we
introduce heuristic enhancements and demonstrate the new algorithm's
Joint work with Irit Katriel, Eli Upfal, and Pascal Van Hentenryck.
Meinolf Sellmann serves as Assistant Professor at Brown
University since 2004. He received his Diploma of Computer Science in
1997 from Paderborn University (Germany). After a year at Lucent
Technology Bell Labs he returned to his alma mater in Paderborn in 1998
from where he received his PhD in 2002. Before coming to Brown, he
spent sixteen months as Postdoctoral Associate at Cornell University.
While an important aim for Professor Sellmann is to provide actual
software systems that can tackle real-world applications efficiently,
the abstraction and generalization of originally problem-tailored
approaches to standard solution methods that facilitate algorithm
design and algorithm engineering for constraint satisfaction and
constrained optimization is a key part of his work. Main methodological
contributions of his research are the development of Symmetry Breaking
by Dominance Detection, Structural Symmetry Breaking, Streamlined
Constraint Reasoning, CP-based Column Generation, CP-based Lagrangian
Relaxation, and the introduction of associated theoretical notions like
Relaxed Consistency and Approximated Consistency.
Title: On Iterons of Automata
Abstract Iterons of automata are periodic coherent propagating
structures-patterns of symbols- that emerge in cellular nets of
automata. In a sense, they are like fractal objects; they owe their
existence to iterated automata maps (IAMs) performed over a given
string of symbols.
The iterons comprise of particles and filtrons. The
particles, or signals, are well known in classical cellular automata
(CAs) where iterated parallel processing of strings occurs. They spread
and carry local results, synchronize various events, combine
information, transform data, and carry out many other actions necessary
to perform a computation or to complete a global pattern formation
process in extended dynamical systems.
The filtrons form another class of coherent objects
supported by IAMs. They emerge in iterated serial string processing
which is a sort of recursive filtering (IIR filtering). In many aspects
the filtrons are like solitons known from nonlinear physics; e.g. they
pass through one another, demonstrate elastic collisions, undergo
fusion, fission and annihilation, and form breathers as well as other
We start by showing some images of solitons which are
fundamental phenomena in nonlinear physics. However, the aim of the
talk is to present a unified automaton approach to discrete coherent
structures. We give basic definitions of these entities and present the
automata which are equivalent to some known processing mechanisms (eg.
box-ball systems) capable of supporting coherent soliton-like objects.
The examples of various phenomena of interacting particles
and filtrons, like multifiltron collisions, trapped bouncing filtrons,
cool filtrons, repelling filtrons, annihilation, fusion, fission, and
spontaneous decay or quasi-filtrons are presented also, using the
images of space-time diagrams of filtrons.
Potential applications of the theory of iterons are
mentioned also; they cover, among others, simulating nonlinear physics
phenomena, solitonic computations, complex systems behavior analysis
and synthesis, and solitary waves prediction.
The outline of the iterons of automata can be found in:
P. Siwak. Iterons of automata. In Collision-Based Computing. A.
Adamatzky (ed.), Springer-Verlag, London. 2002, pp. 299-354.
Title: iRobot Corporation R&D Overview
iRobot provides robots that enable people to complete complex tasks in
a better way. We have developed technology incorporating advanced
concepts in navigation, mobility, manipulation and artificial
intelligence to build industry-leading robots. The iRobot Roomba floor
vacuuming robot and Scooba floor washing robot perform time-consuming
domestic chores, and the iRobot PackBot tactical military robots
perform battlefield reconnaissance and bomb disposal. In addition, we
are developing the Small Unmanned Ground Vehicle reconnaissance robot
for the U.S. Army’s transformational Future Combat Systems program and,
in conjunction with Deere & Company, the R-Gator unmanned ground
vehicle. We also have an active Research Group whose mission is to
provide the innovation that drives the company’s products forward.
Robotic research has always been the foundation of iRobot, with each
advance driving the next one forward. From the insectoid Ghengis to the
crablike Ariel, from the battle-proven PackBot to the heavy-duty
Warrior, from the networked robots of the Swarm to the many generations
of Roomba, each iRobot offering is the result of vision, teamwork, and
world-class research. This talk will provide an overview of these and
other of iRobot’s past and present research and development programs as
well as the foundational skills and technologies that enabled them and
will drive the next-generation of robotic systems.
Speaker Bio Dr. Chris Jones is the Research Program Manager at iRobot
Corporation. iRobot delivers innovative robots that are making a
difference in people’s lives. From cleaning floors to disarming
explosives, iRobot constantly strives to find better ways to tackle
dull, dirty and dangerous missions -- with better results. The
multi-disciniplary Research Group Dr. Jones leads is engaged in
conceptual and developmental robotics research in the areas of mobile
robot autonomy, collaborative unmanned systems, human-robot interaction
and interfaces, and innovative unmanned systems design concepts. Prior
to joining iRobot in 2005, he was involved in robotics research and
development at the Center for Robotics and Embedded Systems at the
University of Southern California, the Artificial Intelligence Lab at
the University of Zurich, the Intelligent Systems and Robotics Center
at Sandia National Laboratories, and the Robotics Research Lab at Texas
A&M University. Dr. Jones received his Ph.D. and M.S. in Computer
Science from the University of Southern California and his B.S. in
Computer Engineering from Texas A&M University.
Title: Olafctory bulb neurotransmitter studies and
The sense of smell, called olfaction, involves the detection and
perception of different odors. For both humans and animals, it is one
of the important means by which they communicate with the environment.
This odor-detecting system is called olfactory bulb. Its functionality
is based on neurons, primarily mitral and granule cells, and
communication among them. This process is very complex and involves
different types of neurotransmitters. The basic function of
neurotransmitters is the realization of communication processes between
neurons. Additionally, they are responsible for the efficient and
accurate processing of the information, as well as for the generation
of cellular changes, which corresponds to the memory functionality. We
provide the description of different neurotransmitters and their
dynamics. The main focus of our work is to analyze the neurotransmitter
effects based on the computational simulations corresponding to the
biological environment in the olfactory bulb.