Vasanth Sarathy
Research Assistant Professor, Tufts University
vasanth [DOT] sarathy [AT] tufts.edu
About me
I'm a Research Assistant Professor of Computer Science at Tufts University studying Artificial Intelligence (AI).
My research combines neural and symbolic approaches to make generative AI systems more creative and inventive while staying within the bounds of human socio-cultural norms. The goal of my research is to build trustworthy and competent AI co-creators and teammates. A few topics of interest are:
- Making and Breaking Inductive Biases: How can AI systems identify their own biases and how can we supply useful biases to them?
- Reasoning with Implicit Knowledge: How can AI systems represent and reason about implicit knowledge (e.g, socio-cultural norms)?
- Pragmatics: How can context contribute to natural language understanding?
- Abstract Concepts: How can AI systems generate, use and compose novel abstract conceptual ideas?
- Accountability for Creativity: Who is accountable when AI systems create novel artifacts, unforseeable by the designer?
I have a Ph.D. in Computer Science and Cognitive Science from Tufts University, an MS in EECS from MIT, and a JD from Boston University School of Law. Prior to Tufts, I was a Senior Researcher at Smart Information Flow Technologies (SIFT) where I worked closely with the Department of Defense and the Intelligence Community to solve cutting-edge AI challenges. I was selected as the Grand Prize Winner of the NSF 2026 Idea Machine Competition and established Creative Problem Solving as an NSF research priority.
I am a licensed attorney in the State of Massachusetts and practiced Intellectual Property law for nearly a decade at the law firm of Ropes & Gray.
I also LOVE to draw cartoons and play chess.
Research
Current Research Projects
- LLM+Robots: Integrating large language models (LLMs) and vision-language models (VLMs) into a physical robot's natural language understanding and reasoning components, so these robots can better communicate with us and follow our instructions.
- Planning+RL: Developing hybrid Symbolic Planning and Reinforcement Learning approaches to enable embodied agents to learn conceptual knowledge of its environment through hypothesis generation and experimentation.
Recent Research Efforts
- Argumentation and Negotiation Support: extracting and evaluating argument structures from persuasive communications
- Automated Content Moderation: enhancing pro-sociality and mitigating anti-sociality in online social media forums
- Learning Socio-Cultural Norms: surfacing implicit knowledge about norms and using it for understanding language and selecting actions
- Affordance Perception: perceiving action possibilities and functions of objects
- Humor and Cartoon Generation: generating jokes, images, and captions for cartoons and monologues
Publications
Most recent publications on Google Scholar.
‡ indicates equal contribution.
BIPLEX -- Creative Problem-Solving by Planning for Experimentation
Sarathy, Vasanth, Scheutz, Matthias
International Conference on Computational Creativity
From Unstructured Text to Causal Knowledge Graphs -- A Transformer-Based Approach
Friedman, Scott, Magnusson, Ian, Sarathy, Vasanth, Schmer-Galunder, Sonja
arXiv preprint arXiv:2202.11768
Spotter -- Extending symbolic planning operators through targeted reinforcement learning
Sarathy, Vasanth, Kasenberg, Daniel, Goel, Shivam, Sinapov, Jivko, Scheutz, Matthias
arXiv preprint arXiv:2012.13037
Learning context-sensitive norms under uncertainty
Sarathy, Vasanth
Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society
When exceptions are the norm -- Exploring the role of consent in hri
Sarathy, Vasanth, Arnold, Thomas, Scheutz, Matthias
ACM Transactions on Human-Robot Interaction (THRI)
MacGyver Problems -- AI Challenges for Testing Resourcefulness and Creativity
Sarathy, Vasanth, EDU, TUFTS, Scheutz, Matthias
Advances in Cognitive Systems
Cultural Value Resonance in Folktales -- A Transformer-Based Analysis with the World Value Corpus
Sarathy, Vasanth, Kantharaju, Pavan, McLure, Matthew D, Goldman, Robert P
Social, Cultural, and Behavioral Modeling -- 15th International Conference, SBP-BRiMS 2022, Pittsburgh, PA, USA, September 20--23, 2022, Proceedings
BIPLEX -- Creative Problem-Solving by Planning for Experimentation
Sarathy, Vasanth, Scheutz, Matthias
International Conference on Computational Creativity
RAPID-Learn -- A Framework for Learning to Recover for Handling Novelties in Open-World Environments.
Goel, Shivam, Shukla, Yash, Sarathy, Vasanth, Scheutz, Matthias, Sinapov, Jivko
2022 IEEE International Conference on Development and Learning (ICDL)
From Unstructured Text to Causal Knowledge Graphs -- A Transformer-Based Approach
Friedman, Scott, Magnusson, Ian, Sarathy, Vasanth, Schmer-Galunder, Sonja
arXiv preprint arXiv:2202.11768
A novelty-centric agent architecture for changing worlds
Muhammad, Faizan, Sarathy, Vasanth, Tatiya, Gyan, Goel, Shivam, Gyawali, Saurav, Guaman, Mateo, Sinapov, Jivko, Scheutz, Matthias
Proceedings of the 20th International Conference on Autonomous Agents and MultiAgent Systems
Spotter -- Extending symbolic planning operators through targeted reinforcement learning
Sarathy, Vasanth, Kasenberg, Daniel, Goel, Shivam, Sinapov, Jivko, Scheutz, Matthias
arXiv preprint arXiv:2012.13037
Reasoning requirements for indirect speech act interpretation
Sarathy, Vasanth, Tsuetaki, Alexander, Roque, Antonio, Scheutz, Matthias
Proceedings of the 28th International Conference on Computational Linguistics
Sense-Making Machines
Sarathy, Vasanth
Tufts University
Developing a corpus of indirect speech act schemas
Roque, Antonio, Tsuetaki, Alexander, Sarathy, Vasanth, Scheutz, Matthias
Proceedings of the Twelfth Language Resources and Evaluation Conference
Toward the engineering of virtuous machines
Govindarajulu, Naveen Sundar, Bringsjord, Selmer, Ghosh, Rikhiya, Sarathy, Vasanth
Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society
Learning context-sensitive norms under uncertainty
Sarathy, Vasanth
Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society
On resolving ambiguous anaphoric expressions in imperative discourse
Sarathy, Vasanth, Scheutz, Matthias
Proceedings of the AAAI Conference on Artificial Intelligence
When exceptions are the norm -- Exploring the role of consent in hri
Sarathy, Vasanth, Arnold, Thomas, Scheutz, Matthias
ACM Transactions on Human-Robot Interaction (THRI)
Knowledge Acquisition in the Cockpit Using One-Shot Learning
Gizzi, Evana, Le Vie, Lisa, Scheutz, Matthias, Sarathy, Vasanth, Sinapov, Jivko
NAECON 2018-IEEE National Aerospace and Electronics Conference
MacGyver Problems -- AI Challenges for Testing Resourcefulness and Creativity
Sarathy, Vasanth, EDU, TUFTS, Scheutz, Matthias
Advances in Cognitive Systems
Learning Cognitive Affordances for Objects from Natural Language Instruction
Sarathy, Vasanth, EDU, TUFTS, Oosterveld, Bradley, Krause, Evan, Scheutz, Matthias
Quasi-Dilemmas for Artificial Moral Agents
Kasenberg, Daniel, Sarathy, Vasanth, Arnold, Thomas, Scheutz, Matthias, Williams, Tom
arXiv preprint arXiv:1807.02572
A Generalized Framework for Detecting Anomalies in Real-Time using Contextual Information
Gizzi, Evana, Le Vie, Lisa, Scheutz, Matthias, Sarathy, Vasanth, Sinapov, Jivko
Real world problem-solving
Sarathy, Vasanth
Frontiers in human neuroscience
An Overview of the Distributed Integrated Cognition Affect and Reflection DIARC Architecture
Scheutz, Matthias, Williams, Thomas, Krause, Evan, Oosterveld, Bradley, Sarathy, Vasanth, Frasca, Tyler
Learning behavioral norms in uncertain and changing contexts
Sarathy, Vasanth, Scheutz, Matthias, Malle, Bertram F
2017 8th IEEE International Conference on Cognitive Infocommunications (CogInfoCom)
Mental Representations and Computational Modeling of Context-Specific Human Norm Systems.
Sarathy, Vasanth, Scheutz, Matthias, Kenett, Yoed N, Allaham, Mowafak, Austerweil, Joseph L, Malle, Bertram F
CogSci
The macgyver test-a framework for evaluating machine resourcefulness and creative problem solving
Sarathy, Vasanth, Scheutz, Matthias
arXiv preprint arXiv:1704.08350
A logic-based computational framework for inferring cognitive affordances
Sarathy, Vasanth, Scheutz, Matthias
IEEE Transactions on Cognitive and Developmental Systems
Enabling basic normative HRI in a cognitive robotic architecture
Sarathy, Vasanth, Wilson, Jason R, Arnold, Thomas, Scheutz, Matthias
arXiv preprint arXiv:1602.03814
Inferring higher-order affordances for more natural human-robot collaboration
Sarathy, Vasanth
2016 11th ACM/IEEE International Conference on Human-Robot Interaction (HRI)
Beyond grasping-perceiving affordances across various stages of cognitive development
Sarathy, Vasanth, Scheutz, Matthias
2016 Joint IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)
Cognitive affordance representations in uncertain logic
Sarathy, Vasanth, Scheutz, Matthias
Fifteenth international conference on the principles of knowledge representation and reasoning
Semantic representation of objects and function
Sarathy, Vasanth, Scheutz, Matthias
Proceedings of the 2015 IROS Workshop on Learning Object Affordances
Vitæ
Full Resume in PDF.
-
Tufts 2023 - NowResearch Assistant Professor
Dept. of Comp. Sci. -
SIFT 2020 - 2023Senior Researcher
Artificial Intelligence -
Tufts 2015 - 2020Ph.D. Student
Computer Science and Cognitive Science
Ropes & Gray 2005-2015Lawyer (Associate)
Intellectual Property GroupBoston University School of Law 2007-2010Law Student
MIT 2003 – 2005Masters and Ph.D. Student in EECS (Course 6)
Research Laboratory for ElectronicsSchlumberger Product Center June 2003- Aug 2003Research Intern
Resistivity GroupSchlumberger-Doll Research Laboratory June 2001- Aug 2001Research Intern
Real-time Inversion GroupUniversity of Arkansas 1999-2003B.S. Student
Electrical Engineering
Summa Cum LaudeAcknowledgement
This website uses the website design and template by Martin Saveski