Multimodal Interaction Systems

At Intel, I have been leading a group of Researchers to work on problems related to Multimodal Interactions for a couple of domain applications. We work on multimodal fusion of signals for Intent Understanding and Emotion recognition (both human to human and human to machine) and apply various learning techniques to develop robust and practical systems.

Semantic Search and Recommendation Systems

Cobot is a intelligent information agent for community based systems that analyzes conversations and social preferences to provide socially filtered, semantically analyzed conversational recommendations. These agents monitor user conversations with other users in the community and provide personalized as well as community based recommendations to users. The core focus of this research involves efficient integration of language understanding, user modeling and case based reasoning techniques for creating more usable end-user search, interaction and recommendation systems.
Keywords: User Modeling, Semantic Search, Recommendation Systems

Language and Context

With an explosion in proliferation of user-generated content, the productivity of time spent online is decreasing and quality of readily available online content is deteriorating. There is an increasing need for socially intelligent systems that can readily understand content, intent, context and interactions to leverage problem solving. We are modeling community interactions in conversations by studying both language and content to proactively target the community for information exchange. We envision a system that encourages user engagement and participation by prompting questions and asking to suggest answers based on factors like user's participation besides semantic understanding and social feedback. Sustenance of community participation is a major challenge; we strive to create different engagement models for different types of users to target them for recommendations for community sustenance.
Keywords:Syntax, Semantics, Sentiment Analysis, Question Answering

Knowledge Representation and Reasoning

The rapidly increasing volume of unstructured information poses the challenge of knowledge integration so as to build autonomic computing systems that can acquire, represent and learn such knowledge, and efficiently reason from it to aid in knowledge discovery and re-use. The construction of these automated systems to assist problem solving is impeded by difficulties in formalizing knowledge and in encoding that knowledge for use by computer systems. This research focuses on developing methods of knowledge representation to build a semantic intelligence infrastructure using techniques of language understanding, ontology based representation and semantic nets.
Keywords:Ontology Learning, Semantic Nets, Case based Reasoning