Research Interests
- Decision Theory, Preference Logic, CP-nets, TCP-nets, Preference-guided Search, Utility Theory, Multi-attribute Utility theory, Ordinal Value Theory
- Web Service Composition: Composition techniques, Formal methods and model checking approaches to generate functionally compliant compositions, Composition based on non-functional attributes, Generating "most preferred" compositions based on Qualitative Preferences
- Semantic Web Services: Semantic Description and Discovery of Web Services, Semantic annotations to Web Services, Automated Composition of Web Services using Functional and Non-functional requirements and constraints.
- Semantic Web: Knowledge Representation on the Web, Ontology, Reasoning, Description Logic
Publications
- On Utilizing Qualitative Preferences in Web Service Composition: A CP-net based approach -- Ganesh Ram Santhanam, Samik Basu, Vasant Honavar -- 2nd International Workshop on Web Service Composition and Adaptation (WSCA-2008)
- Towards Scalable Constant Time Reasoning Over Partial Order Ontology with Mappings -- Term Paper fulfilling the requirement of the course Com S 572 Principles of Artificial Intelligence. PDF
Scalable Constant Time Reasoner
In partial fulfillment for the course requirements of Com S 572 Principles of Artificial Intelligence.
Abstract One of the challenges of realizing the potential of the Semantic Web is to build efficient reasoners that can run in constant or near constant time and can scale easily. We present an approach that uses a relational database as a store to pre-compute the closure of the knowledge base in partial order ontology, consequently improving the reasoning efficiency.
Approach We propose to address the problem by storing the knowledge base in ontology, mappings in a relational database. The basic idea is to save all the knowledge present in the ontology in the form of relational tuples in the database. By all knowledge, we mean to store all the concepts, axioms, and the closure of the axioms in the database as tuples.
Publications
- "Knowledge Engineering of Creative Musical Expressions using Carnatic Music Ideology" at Florida Artificial Intelligence Research Conference (FLAIRS 2004) held in under the auspices of American Association of Artificial Intelligence (AAAI), May 2004, Florida, USA. Published in the Proceedings, AAAI press. PDF
- "A New Approach to Music Cognition using Carnatic Music Ideology" at International Conference on Advanced Computation and Communication (ADCOM 2003) Dec 17-20 2003 conducted by Advanced Computation Society, IISc Bangalore. PDF
Undergraduate Research Project
In partial fulfillment for the degree of Bachelor of Technology at University of Madras [2003 - 2004]
- Project Report: Pattern Classification of creative musical expressions using Carnatic Music Ideology
- Advisor: Prof. S.Rajendran, Head, Department of Information Technology at SRM Engineering College
- Summary: The project employed Statistical and Machine Learning methods to create a musical knowledge base. The project tracked two critical musical attributes that contribute to creative musical expression - patterns of notes and inflexions between notes. A case study of two musicians with their real time concert samples was taken and Statistical Pattern classification techniques were used to prove the research hypothesis. An application was developed in Java and two research papers were published in international conferences.
- Project Presentation: Download Powerpoint slides
- Project Report: Available on request.
Highlights
- Proposed, designed and implemented (in Java) a system that could classify creative music expressions using a tailor-made machine learning algorithm.
- Successfully classified sequential data (both standard, prescribed note patterns and extempore dynamic note patterns) from purely creative expositions of musicians by defining a generic grammar for musical knowledge representation.
- Using statistical methods, I was able to clearly discern the differences in the inherent musical style and creativity in musicians.
- Results of this work carry a direct consequence onto
- Sequential Supervised Learning
- Stylometry and Genre classification
- Recieved 100 % marks for the project color: #555A60;