http:^^www.cs.jhu.edu^~kasif^

来自「This data set contains WWW-pages collect」· EDU^~KASIF^ 代码 · 共 339 行

Date: Wed, 20 Nov 1996 23:04:16 GMT
Server: NCSA/1.5.1
Last-modified: Sun, 31 Mar 1996 17:20:42 GMT
Content-type: text/html
Content-length: 12622

<title> Simon Kasif's home page </title>
<h1> Simon Kasif's home page </h1>

<!WA0><img align = top src = "http://www.cs.jhu.edu/~kasif/kasif.gif"> 
<p>
<hr>


<h1> Research and Teaching </h1>

<pre>
High-Performance Intelligent Systems:
     Computational modelling and control of complex systems
     Machine Learning
     Systems for data analysis and data mining 
     Parallel Intelligent Systems
     Automated Reasoning 
General Interests:
     Parallel Computation 
     Computational Neuroscience
     Computational Biology 

</pre>

<p>

<H2> 
<!WA1><a href = "http://www.cs.jhu.edu/~kasif/send-out"> CFP: AAAI Symposium on Learning Complex Behaviors </a> </H2>

<h2> Research Interests </h2>

Intelligent information processing systems are likely to 
play a most significant role in the 21st century, and  
the design and analysis of these systems is one
of the great scientific challenges we face today.
Consider the following realistic possibilities, which are
currently the main focus of ongoing work in intelligent 
systems research around the globe.

<p>
An intelligent vision system installed in a car tracks 
the road and wakes up the dozing driver when the car is beginning to 
to head towards the shoulder. Thousands of lives could be saved 
annually.
Another data-driven system could help to perform precise
radioactive therapy on a cancerous tumor
in a way that minimizes damage to adjacent cells.  
An interactive automated system could consider a set of symptoms 
and warn of a likely heart attack in an emergency room patient.
Another medical system can perform sophisticated analysis 
of a collection of tests, help assess the probability of cancer 
recurrence and suggest a schedule for repeated evaluations. 

<p>
The basic principle driving these systems is that they can
LEARN to function from very large corpuses of data.
What makes this technology possible is the remarkable 
match between the massive amount of information that can be stored on 
modern computers and the amazing computational capabilities (speed) 
of these machines. 
<p>
Our society is currently entering a new period 
where massive amounts of heterogeneous information are generated 
and stored on computers and are readily available for exploration
over academic networks, digital libraries, and 
commercial information services.
At the same time, modern networks of computers can perform 
billions of computation steps per second.
This rather unique match of raw computer power 
and sea of available information creates many previously 
unfeasible opportunities for building intelligent systems. 
Coping with information explosion, scientific data mining, 
interactive modelling of unstructured data and complex processes, 
and intelligent data-driven decision-making with uncertainty
are among the major current focus areas in computer science.  

<p>

One striking example of the use of this computing power 
was a recent application developed at Johns Hopkins University.
A chess end-game analysis program (designed by Lewis Stiller) 
running on a 65,536 processor machine and using gigabytes of storage found 
chess end-games that absolutely require more than 220 moves to win a 
simple chess game.  Chess masters for generations 
used a rule that declared a draw after only 50 moves.
The old rule was dictated by human experience that could not 
perform this type of analysis.

<p>

A different program, implemented by an interdisciplinary group of 
researchers from Astronomy and Computer Science,
automatically finds stars, galaxies or filters cosmic rays in 
images obtained by the Hubble Telescope.  
This program is using a decision tree induction system 
implemented by S. Murthy at the Johns Hopkins University.
This program is very fast and can easily process millions of 
astronomical objects in a short amount of time.
<p>

My current research program focuses on 
high-performance computer systems and efficient algorithms
that provide the computational capabilities to support this 
critical information technology.
Our main focus is the design and analysis of
algorithms for high-performance intelligent systems. Specifically, we
are interested in efficient parallel and sequential algorithms for 
modelling and controlling complex processes (e.g., computer systems or 
neural networks) --
intelligent information retrieval and modelling of large data 
depositories (that can be used in educational environments) --
constraint networks and probabilistic networks (that can be used for 
decision-making with uncertainty, e.g., medical diagnosis) -- 
and machine learning.  The algorithms designed by our group
have been applied to a variety of applications, such as
public health,  computer vision, computer game playing, 
astronomy and several biomedical applications.

<p>


<h2> Academic Experience </h2>

<pre>

Jun 1991 & present Associate Professor, Dept. of Computer Science
                   joint appointment in Cognitive Science since 1989.

July 1994 - Jun 95  Sabbatical Leave: 
                    NEC Research Laboratories and
                    Computer Science Department, Princeton University 

Sep 1987 - Jun 91   Assistant Professor, Dept. of Computer Science,
		    The Johns Hopkins University

Sep 1989 - Jun 91   Assistant Professor, JOINT appointment, Cognitive Science Center
                    The Johns Hopkins University

Sep 1985 - Sep 87   Assistant Professor, Dept. of Electrical Engineering and Computer Science
		    The Johns Hopkins University. 

Jun 1987 - Aug 87   Visiting Scientist, Weizmann Institute of Science

Jun 1985 - Aug 85   Research Scientist, University of Maryland. 

May 1985            Ph.D. Computer Science Department,  University of Maryland. 

</pre>

<h2> Recent Journal Publications (since 1990) </h2>


<ol>

<li>
Kasif, S., On the Parallel Complexity of Discrete Relaxation in Constraint Networks,
Artificial Intelligence Journal,  pp. 275-286, October 1990.
<li>
Delcher, A. and S. Kasif,
Efficient Parallel Term Matching and Anti-Unification,
Journal of Automated Reasoning,  pp. 391--406, 1992.

<li>
Kasif, S., S. Banerjee, A. Delcher and G. Sullivan,
Some Results on the Complexity of Symmetric Connectionist Networks,
Annals of Mathematics and Artificial Intelligence, Nov. 1993, 327-344.
<li>
Kasif, S. and A. Delcher,
Analysis of Local Consistency in Parallel Constraint Networks, 
Artificial Intelligence, 307--327, 1994.
<li>
Kasif, S., Optimal Parallel Algorithms for Quad-Tree Problems, 
Journal of Computer Vision and Image Processing, pp. 281-285, May 1994. 
<li>
Salzberg, S., D. Heath, A. Delcher and S. Kasif, 
Best Case Analysis of Nearest Neighbours Algorithms, 
IEEE Transaction on Pattern Analysis and Machine Intelligence,
June 1995, 17:6, 599-608.
<li>
Heath, D., S. Kasif, S. R. Kosaraju, S. Salzberg and G. Sullivan, 
Learning Nested Concept Classes with Limited Memory, to appear 
in the  Journal of Experimental and Theoretical AI, 1996.
<li>
Heath, D. and S. Kasif,
On Voronoi Covers with Applications to Machine Learning,
Computational Geometry: Theory and Applications, Nov. 1993, 289-305.
<li>
Delcher, A. and S. Kasif,
Term Matching on a Mesh-Connected Array of Processors,
Annals of Mathematics and Artificial Intelligence, Volume 14, 177--186, 1995.
<li>
Murthy, S., S. Kasif and S. Salzberg,
A System for Induction of Oblique Decision Trees,
Journal of Artificial Intelligence Research, 2:1 (1994),1--33.
<li>
Delcher, A, A. Grove, S. Kasif and J. Pearl, 
Logarithmic Time Queries and Updates in Probabilistic Networks, 
to appear, Journal of Artificial Intelligence Research, 1996.
<li>
D. Waltz, and S. Kasif, On Reasoning from Data, to appear, 
Computing Surveys}, 1996.
<li>
Delcher, A. and S. Kasif,
The Complexity of Incremental Parallel Computations'', in review.
<li>
Bright, J., Kasif, L. Stiller, Memory-Efficient Parallel Algorithms
for Bi-Directional Search'', in review. 
<li>
Kasif, S., Salzberg, S. and Waltz, D., Rachlin, J. and Aha, D.,
Towards of a Framework for Memory-Based Reasoning,  
in review. 
<li>
D. Dobkin, D. Gunopoulous, T. Fulton, S. Kasif, and S. Salzberg,
Induction of Shallow Decision Trees ,
IEEE Trans. on Pattern Analysis and Machine Intelligence, in review. 

</ol>

<h2> Selected Conference Publications (since 1990) </h2>
<ol>
<li>
Heath, D., S. Kasif, S. R. Kosaraju, S. Salzberg and G. Sullivan, 
``Learning Nested Concept Classes with Limited Storage'', 
Proceedings of the International Joint Conference on Artificial Intelligence
(IJCAI-91), pp. 777-782, 1991.
<li>
Salzberg, S., D. Heath, A. Delcher and  S. Kasif, 
``Learning with a Helpful Teacher'',
Proceedings of the International Joint Conference on Artificial Intelligence,
(IJCAI-91), pp. 705-711, 1991.
<li>
S. Kasif and A. Delcher, 
``Improved Decision Making in Game Trees: Recovering from Pathology'',	
Proceedings of the National Conference on Artificial Intelligence (AAAI-92), 
pp. 513-518, July 1992.
<li>
D. Heath, S. Kasif and S. Salzberg,
``Learning Oblique Decision Trees'', Computational Learning Theory and 
Natural Learning Systems, 1992.
<li>
Kasif, S. and A. Delcher,
Analysis of Local Consistency in Parallel Constraint Networks 
International Conference on Artificial Intelligence and Vision, 
pp. 217-231, 1992.
<li>
S. Kasif,
``Iterative Focusing and Hashing: An Alternative to Alpha-Beta'',
International Conference on Artificial Intelligence and Vision, 
pp. 59-72, 1992.
<li>
D. Heath, S. Kasif and S. Salzberg,
Learning Oblique Decision Trees,
Proceedings of the International Joint Conference on 
Artificial Intelligence, (IJCAI 93), pp. 1002--1007, 
August 1993. 
<li>
Murthy, S., S. Kasif, S. Salzberg and R. Beigel,
OC1: A Randomized Algorithm for Building Oblique Decision Trees,
Proceedings of the National Conference on Artificial Intelligence, (AAAI-93),
pp. 322--327, July 93. 
<li>
\item
Delcher, A., S. Kasif, H. Goldberg and W. Xsu,
Protein Secondary-Structure Modeling with Probabilistic Networks,
International Conference on Intelligent Systems and Molecular Biology,
pp. 109--117, 1993.    
<li>
Delcher, A., S. Kasif, H. Goldberg and W. Xsu,
Application of Probabilistic Causal Trees to Analysis of 
Protein Secondary Structure, 
Proceedings of the National Conference on Artificial Intelligence, 
(a short version of the above), pp. 316--321, July 1993. 
<li>
Bright, J., Kasif, L. Stiller, Exploiting Algebraic Structure 
in Parallel State-Space Search, 
Proc. of the 11-th National Conf. on Artificial 
Intelligence, July 1994, (AAAI-94), 
pp. 1341-1346.
<li>
Rachlin, J., S. Kasif, S. Salzberg and D. Aha,
Toward of a better understanding of Memory-Based Classifiers,  
Proceeding of the 11-th Intern. Conf. on Machine
Learning, pp. 242--250, July 1994.
<li>
Fulton, T., S. Kasif and S. Salzberg,  
Efficient Algorithms for Finding Multi-Way Splits for Decision Trees, 
JHU TR, December 1993,  Proceeding of the 12-th Intern. Conf. on Machine
Learning, July 1995.
<li>
Delcher, A, A. Grove, S. Kasif and J. Pearl, 
Logarithmic Time Queries and Updates in Probabilistic Networks, 
Proceedings of the 1995 Conference on Uncertainty in AI, August 1995.
<li>
D. Dobkin, D. Gunopoulous,  S. Kasif,
Induction of Low-Depth Decision Trees'' 
International Conference on Mathematics and Artificial Intelligence, January 1996.
<li>
S. Weiss, S. Kasif, and E. Brill, 
Towards a Framework for Adaptive Information Retrieval,
AAAI Spring Symposium on Information Retrieval (to appear 1996). 
</ol>

<h2> Ph.D Students</h2>

<ol>
<li> <!WA2><a href = "http://blaze.cs.jhu.edu/delcher/home.html"> Art Delcher, </a>
Ph.D. 1989, M.A. in Mathematics, Johns Hopkins University.  
Thesis area:  Parallel Algorithms for Artificial Intelligence.
Currently, Full Professor and 
Chairman CS Department Loyola College,
Adjunct Research Faculty Johns Hopkins Computer Science Department.
<li> <!WA3><a href = " http://mission.amil.jhu.edu/projects/liver"> David Heath </a>
Ph.D. 1992, BS in EE, Cal Tech, Wolman Fellow,
Thesis area: Algorithms for Machine Learning.
Currently, faculty in Johns Hopkins Medical School, Computer-Assisted Radiology.
<li>
Lewis Stiller, BS in Mathematics, Johns Hopkins University, 
Ph.D. May 1995, National Defense Fellowship,
Thesis area: Exploiting Symmetry in Parallel Computation, 
currently at Berkeley, CA.
<li> <!WA4><a href = "http://sugar.cs.jhu.edu/~trux/"> Truxton Fulton  </a>
BS in Computer Science, Cal. Tech, Wolman Fellow.
Research area: Adaptive Computer Systems, expected 1996. 
<li> <!WA5><a href = "http://www.cs.jhu.edu/~weiss/"> Scott Weiss </a>
BS in Computer Science, Carnegie-Mellon University, Wolman Fellow.
Research Area:  Adaptive Information Retrieval, expected 1996.

</ol>



<!WA6><A HREF="http://www.cs.jhu.edu/~kasif/hot.html">Information</A>