Supervised graduate and honors students:
- Redcenco Dmitri, "SPARQL Query Client: Semantic Web & Linked Data", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- Reuter Alex, "The Straight Road: Cloud Computing", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- DeCarlo Jonathan, "Search Heuristics with PSpace-Complete Decision Problems: Experiments with Computer Amazons", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- Kovach Robert Jr., "Comparison of Classification Methods on Website Leads", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- Cudjoe Robert, "Customer Management System Case: Budget One Hour Cleaners", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- Pawelczyk Greg, "Agile Software Development with Specialized Software", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- Braverman Jonathan, "Automatic Door Motion Analyzer", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2015.
- Micun Pawel, "Leveraging GPU Parallelism to Factor
Large Integers", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2014.
- Horner Joseph, "Search Based Software
Engineering Defect
Management Object Oriented Evolutionary Testing - Unit Test Automation
Using Evolutionary Computation and The $-Calculus Anytime Algorithm",
Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- Loomis Kyle, "Enabling Search on a Semantic Property Graph", Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- LaPointe David, "Performance Comparison of JDBC ResultSet Data
Getter Strategies", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2014.
- Wade David Joe, "Federating AMQP1.0
Message-Oriented Middleware (MOM) Brokers", Master Project, Rensselaer
Polytechnic Institute at Hartford, 2014.
- Zipkin Justin,
"Improving Resiliency and Scalability in the Real Traffic Grabber",
Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- Summers Matthew, "Relational Model to Entity Translator", Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- Geoghegan
John, "Popularity and Interlingual Occurrence of Wikipedia Articles",
Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- Greene
Daniel, "Design and Development of a Cross-Platform Game Engine
Utilizing HTML5 and JavaScript", Master Project, Rensselaer Polytechnic
Institute at Hartford, 2014.
- Uricchio Andrew, "An Overview
of Search Engines with a Focus on Google", Master Project, Rensselaer
Polytechnic Institute at Hartford, 2014.
- Resto Alexi, "Will NoSQL Speedup k-Means Clustering?", Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- English Alexander, "Recommender Systems", Master Project, Rensselaer Polytechnic Institute at Hartford, 2014.
- Booth
Patrick, "Particle Swarm Optimization Tuning an Inverted Pendulum
Mobile Robot", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2014.
- Donaldson Scott, "Upgrading a Production Microsoft Access
Database: Methodology and Proof-of-Concept", Master Project, Rensselaer
Polytechnic Institute at Hartford, 2013.
- Smith Jonathan, "$-CalcuLisp, an Implementation of $-Calculus
Process Algebra in Common Lisp", Master Project, Rensselaer Polytechnic
Institute at Hartford, 2013.
- Ansari Babar, "$-cala: An Embedded Programming Language Based
on $-Calculus and Scala", Master Thesis, Rensselaer Polytechnic
Institute at Hartford, 2013.
- Hird Ian, "A Dynamic Bootp Server Developed in LabVIEW", Master Project, Rensselaer Polytechnic Institute at Hartford, 2013.
- Chanko David, "Compound Support Vector Machines", Master Project, Rensselaer Polytechnic Institute at Hartford, 2012.
- Desai Shaunak, "Distributed Android^{TM} Applications Development with Ruby Frameworks", Master Project, Rensselaer Polytechnic Institute at Hartford, 2012.
- Goric Timothy, "On the Complexity of Computational Models for
Structured Finance Products", Master Project, Rensselaer Polytechnic
Institute at Hartford, 2012.
- Sanville William, "Cflat: A Turing Complete Language", Master Project, Rensselaer Polytechnic Institute at Hartford, 2011.
- Pham Hung Quoc, "Relational Semantic Integrity Constraints", Master
Project, Rensselaer Polytechnic Institute at
Hartford, 2010.
- Szczepanski Andy, "Neural Character Recognition", Master
Project, Rensselaer Polytechnic Institute at
Hartford, 2010.
- Falconer James, "A Comparison of Data Storage Techniques for
Linear Linked Lists and List Arrays and Their Performance Utilizing the
Google Android Operating System in a Handheld Configuration", Master
Project, Rensselaer Polytechnic Institute at
Hartford, 2009.
- Battad Joseph, "A Comparison of Performance Between Relational
Non-Dimensional and Relational Dimensional Model of Data Warehouse
Database Design: MySQL and ICE Case Study", Master Project, Rensselaer
Polytechnic Institute at
Hartford, 2009.
- Schleicher James, "A Comparison of Back Propagation Neural
Network and Support Vector Machine Based Licence Plate Character
Classification", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2009.
- Forst Bryan, "Addressing Contingency Problems Using kOmega
Continuous Planning: Protocol, Navigation and Planning for Multiple
Robot Agents", Master Project, Rensselaer Polytechnic Institute at
Hartford, 2008.
- DePratti Roland, "The Value of an RDBMS-Centric XML Referential
Integrity Subsystem", Master Thesis, Rensselaer Polytechnic Institute
at Hartford,
2008 (The 2008 Outstanding Student Award in Computer Science).
- Hasan Muhammad Emadul, "Development of the kOmega Optimization
for Multiple Autonomous Mobile Robots", Master Project, Rensselaer
Polytechnic Institute at Hartford, 2007.
- Gaurav Kunal, “Image Recognition with Mobile Agents Using a
Common Middleware Environment”, Master Thesis, Univ. of Massachusetts
Dartmouth, 2007.
- Ramanadham Sandeep, “Study of Polymorphic Computer Viruses”,
Master Project, Univ. of Massachusetts Dartmouth, 2006 (graduated
despite his terminal cancer illness).
- Hung Chang-Yang, "Testing Capabilities of Common Control Language
and ER1 Robot", Master Project, Univ. of Masachusetts Dartmouth, 2006.
- Buzzell Christine, “A Common Control Language for Multiple
Autonomous Undersea Vehicle Cooperation”, Master Thesis, Univ. of
Massachusetts Dartmouth, 2005 (first graduate of Intercampus Graduate
School of Marine Sciences and Technology).
- Atallah Lara, “The $-Calculus Process Algebra Applied to Selected
Algorithms in Bioinformatics”, Master Project, Univ. of Massachusetts
Dartmouth, 2005.
- Lu Chia-Hsin, “Sequence Alignment with Genetic Algorithm”, Master
Thesis, Univ. of Massachusetts Dartmouth, 2004.
- Chang Edmond Yuan-Ming, “Development of J2ME Application: A Stock
Quotas MIDP Application on Mobile Devices”, Master Project,
Univ. of Massachusetts Dartmouth, 2004.
- Gaurav Kunal “Implementation of kOmega Optimization Single Agent
Search”, Master Project, Univ. of Massachusetts Dartmouth, 2004.
- Tsai Jerry-Horng, Kuo Ying-Hsiang, “Application of PDA Program
for Restaurant System”, Master Project, Univ. of Massachusetts
Dartmouth, 2003.
- Wang Chiao “Java Beans Architecture for Ordering DVD Movies”,
Master Project, Univ. of Massachusetts Dartmouth, 2003.
- Peng Hui-Min, Lee Wen-Kuei, “Products Requirement Planning
System”, Master Project, Univ. of Massachusetts Dartmouth, 2003.
- Huang Wei-Ting “Online Transportation Reservation System”, Master
Project, Univ. of Massachusetts Dartmouth, 2003.
- Shah Nikunj, “Improving Performance of Website Using Database
Replication”, Master Project, Univ. of Massachusetts Dartmouth, 2003.
- Lu Chia-Hsin, “DNA Sequence Analysis”, Master Project, Univ. of
Massachusetts Dartmouth, 2002.
- Inn We-Long, “Database System”, Master Project, Univ. of
Massachsuetts Dartmouth, 2002.
- Hung Chao-Hsi, “E-Commerce of UMassDStore”, Master Project, Univ.
of Massachusetts Dartmouth, 2002.
- Hanashi Salah, "System Level Fault-Diagnosis in the Distributed
PVM Environment", Master Project, co-supervisor P.Bodorik, Technical
Univ.
of Nova Scotia/Dalhousie Univ., 1996.
- Tong Stanley Siu-Fung, "Concurrent Prograph: An Approach to
Macro-Dataflow, Object-Oriented Parallel Programming", Ph.D. Thesis,
co-supervisor T.Pietrzykowski, Technical Univ. of Nova Scotia/Dalhousie
Univ., 1995.
- Horree Swarajsingh, "Design and Implementation of a Version of
SEMAL Interpreter", Honors Thesis, Acadia Univ., 1994 (The
Governor-General of Canada Silver Medal and the Univ. Bronze Medal in
Comp. Science awards).
- Seunarine Davika, "The CSA-Based Evolutionary Computer Network
Routing Algorithms", Master Thesis, Acadia Univ., 1994.
- Lim Cheek Yong, "Three Approaches to Expert System Design: Exsys
Shell, Aedes Shell and CSA", Honors Thesis, Acadia Univ., 1994.
- Veerayah Kumaran, "A CSA Based Fault-Tolerant System", Honors
Thesis, Acadia Univ., 1994.
- Yeo Choon Lin, "Methods of Concurrency Control in Distributed
Databases and Knowledge Bases", Honors Thesis, Acadia Univ., 1993.
- Ramachandran Ganapathy, "Implementation of Concurrent Algorithms
in OCCAM on Transputers and Their Preliminary Specification in SEMAL",
Honors Thesis, Acadia Univ., 1993.
- Blondon Robin, "Towards an Implementation of SEMAL: Building the
Set of Equations", Honors Thesis, Acadia Univ., 1993.
- Ho Kwan Wing, "Resource Deadlocks and Their Detection in
Distributed Systems", Honors Thesis, Acadia Univ., 1993.
- Tan Van King, "Using a Calculus of Self-modifiable Algorithms to
Define a Plan Generator", Honors Thesis, Acadia Univ., 1992 (The
Governor -General of Canada Gold Medal award at TUNS).
- Rzasa Jacek, "Algorytm Minimalizacji Pasma dla Wielkich Układów
Liniowych Równań Algebraicznych w Metodzie Elementów Skończonych", ("An
Algorithm for Minimization of the Band for Large Sets of Linear
Algebraic Equations
Using Finite Element Method"), in Polish, Master Postgraduate Thesis,
Rzeszow Univ. of Technology, 1986.
- Celek Mariusz, "Analiza Rozproszonego Systemu Cyfrowego
Tolerującego Błędy", ("An Analysis of a Fault-Tolerant Distributed
Digital System"), in Polish, Master Thesis, Rzeszow Univ. of
Technology, 1984.
- Kapala Janusz, "Współbieżność w Sieciach Mikroprocesorowych",
("Concurrency in Microprocessor Nets"), in Polish, Master Thesis,
Rzeszow Univ. of Technology, 1984.
Master and Ph.D. Theses and Project proposals:
- $-calculus as a formal model of resource-bounded computation.
Resource
bounded computation attempts to find the best answer possible given
operational
constraints. The approach is known under a variety of names, including
flexible computation, anytime algorithms, imprecise-computation,
or
design-to-time scheduling (see, e.g. Dean, Horvitz, Zilberstein).
Anytime
algorithms together with decision-theoretic metareasoning (Russell) are
two new promising techniques and their application to general
decision-making
architectures has not yet been investigated in any depth.
- $-calculus as a universal model of computation and its
expressiveness.
It looks that interactive distributed systems (e.g., intelligent
agents)
require formalisms having richer behavior than Turing machines. It will
be investigated a formal mapping of interaction machines (Wegner),
cellular
automata, pi-calculus (Milner and Pierce), and random automata
networks
to $-calculus. It will be investigated a hierarchy of expressiveness
of
particular models. Problems of universal computation and universal
construction
will be studied.
- Formal semantics of $-calculus based on
non-well-founded
sets
and coinduction. Formal semantics of $-calculus is given in a
traditional form for process algebras, i.e. in the form of the Labeled
Transition System. Wegner has suggested that interactive systems
require models based on non-well-founded sets, coinduction and greatest
fixed points.
- Cost metrics performance measure library. A unique
aspect
of
$-calculus, compared to other models of computation, is its associated
cost mechanism. So far, three types of cost functions are predefined in
$-calculus: crisp, probabilistic and fuzzy-logic. The user is free to
define
own cost functions. It will be desirable to establish a library of most
common cost metrics spanning as many applications as possible. Such a
standardization
attempt should benefit also the field of evolutionary computation.
- Cost function profiling. This is a crucial problem for
applicability
of cost calculus, i.e. to establish the way to define costs of atomic
cost
expressions (an analogue of the elementary probability distribution).
Resource-bounded
optimization typically assumes that statistics from previous
experiments
are used to develop profiles of cost functions. Deduction and empirical
testing can be used to derive costs of atomic and composite cost
expressions.
- Meta-level control library. $-calculus uses a simple
modifying algorithm
for meta-level control, which is a generalization of control structures
used for adaptive expert systems, evolutionary computation and neural
nets.
However, in a similar was as an evolutionary algorithm may lead to a
more
complex form like the cultural, or genocop algorithm, $-calculus allows
for an arbitrary flexible meta-control. Thus a work on the library of
typical
meta-control is needed.
- Emerging global behavior, design and integration of reactive
and
deliberate
behaviors. There are three basic questions for autonomous
agents
(also valid for dynamical systems, Alife, robotics): how does an
ensemble
of agents produce intelligent or interesting output (an emerging global
behavior), how to best decompose a global job into the group of
interacting
agents, and how to integrate reactive and deliberative behaviors?
$-calculus
addresses all three problems in a unique way. The feasibility of
emerging
global behavior is investigated as emerging global optimization through
local optimizations. The design (synthesis) problem is done by
$-calculus
meta-control to partition job in the direction of minimal costs. The
integration
of reactive and deliberative behaviors is done by modifying the
parameter
k for the k Omega-optimization problems.
- Dealing with incomplete and uncertain knowledge. To
deal
with incomplete
and uncertain knowledge, $-calculus has built-in probabilistic or fuzzy
cost functions. Another innovative aspect of the approach is the k
W-optimization
using invisible cost expressions to mask unknown or intractable
portions
of the search space.
- Cost Language (CL) design and implementation. We can
distinguish
at least 7 main classes of programming languages: procedural,
object-oriented,
functional applicative and single assignment, logic, rule-based, and
semantic
network languages, Cost languages based on the $-calculus model of
computation
introduce a distinct new programming paradigm. The CL meta-language
will
be a new cost language (after the SEMAL research prototype) crucial for
all applications. CL should be inherently parallel, highly portable,
flexible,
dynamic, robust, modular, and have a 3D graphical interface. The
cost optimization mechanism differs it from other agent related
languages
like KIF, KQML, or KLAIM.
- The design and implementation of anytime evolvable robots.
It has
been suggested that traditional AI approach of sense-plan-act is too
slow
and brittle for producing effective robotic systems. A nouvelle
behavioral
approach to robotics of purely reactive systems (Brooks), which do not
build a symbolic world representation, are not scalable and not
sufficient
for a number of problems inherent in robotics. Instead of approaches
being
forcibly either reactive or deliberative, the use of $-calculus
process
algebra with meta-reasoning allows the systems to react flexibly using
either deliberation or reaction as appropriate. Combining the best of
both
worlds should be beneficial for robotics.
- The design and implementation of evolvable cost-driven
architectures.
Cost-driven computers are computer architecture representatives
of
$-calculus. Between the $-calculus, cost languages and cost-driven
computers
is the same relation as between Hoare's CSP, OCCAM and Transputer-based
architectures. The main principle of workof cost-driven computers is to
choose the cheapest instruction for execution from the set of
possible
ones. The cost-driven model of computation is different from
control-driven,
data-driven, demand-driven, or pattern-driven models. An evolvable
cost-driven
cellular computer CICADA should be re-configurable and each cell
should implement the $-calculus k Omega-optimization in FPGA.
- Electronic commerce market behavior analysis, prediction and
control.
$-calculus with its optimization cost mechanism could be useful for
distributed
market behavior analysis, prices and profits prediction and control. CL
(Cost Language) seems to be a good candidate for e-commerce
transactions.
- DNA-based agents and molecular biology. An agent
architecture provides
a more natural model for DNA-based computing. Algorithmic descriptions
of DNA computing are fundamentally flawed by failing to account for the
chemical complexity of collections of biomolecules. DNA computing
incorporates
self-modifiability through the action of enzymes. Through the
minimization
of free energy, a DNA-based computer is a cost-driven machine. The
$-calculus
can provide the theoretical framework to direct, control, and better
understanding
of DNA-based computing. In molecular biology, the $-calculus flexible
optimization
search mechanism could be investigated for solutions of hard problems
of
DNA sequence alignment, and the protein folding and structure
prediction
by energy minimization.
- Design and implementation of hybrid expert systems, neural
nets,
genetic
algorithms, and fuzzy systems. $-calculus can express an arbitrary
neural network, evolutionary computation, rule-based expert system, or
logic programming. Very often, it is unclear, which of these methods
will
work the best, or the system may consist of heterogeneous modules. Then
an integration under one theoretical framework subsymbolic and symbolic
systems should be beneficial, i.e. $-calculus will either allow to pick
up the best method, or to to glue heterogeneous components.
- Simulation and investigation of quantum computing and
algorithms.
It is expected that the progress in the miniaturization of computer
hardware
components will lead in 2020s, independently which computing paradigm
will
dominate, to the necessity to take into account quantum effects. Two
basic
features of quantum computing are quantum parallelism and quantum
superposition
of states in qubits. Quantum superposition could be interpreted as a
new
$-calculus cost metric in the domain of complex numbers, and
parallelism
is inherent in $-calculus. Whether a cost calculus could be useful for
quantum computing, it is an open research question and requires careful
studies.
Updated: 2015-05-13, 18:39