Research

 

Overview:
 

My research revolves around using Artificial Intelligence for designing clever solutions to impactful, real-world problems. Specifically I'm interested in leveraging tools such as Machine Learning, Online Learning / Reinforcement Learning, Bandits/Restless Bandits, Meta-Learning, Probabilistic Modeling and Sequential Decision Making.

Currently, in my Ph.D., I've been focusing on using AI for public health. My recent work (NeurIPS-2020) involved designing planning algorithms for health workers, that would recommend to them which patients they must intervene on, so as to maximize the health outcomes of their patient cohort.

I have also worked on projects involving decision-focused learning (differentiable optimization), illegal smuggling prevention using Graph Convolution Networks (GCNs), COVID-19 modeling using Agent-Based Modeling. 

Collaboration: I am excited about high-impact AI research that is useful in the real-world. I am always enthusiastic to collaborate and design cool solutions to real-world problems where AI can help. If you have potential ideas, please feel free to reach out to me at aditya_mate@g.harvard.edu

 

Publications:
 

Conference Publications


[2021]
 

  • Aditya Mate,  Andrew Perrault and Milind Tambe.
    “Risk-Sensitive Interventions in Public Health: Planning with Restless Multi-Armed Bandits”
    [AAMAS-21] International Conference on Autonomous Agents and Multi-agent Systems, AAMAS 2021, London, UK.   [paper] [code]

[2020]
 

  • Aditya Mate*, Jackson Killian*, Haifeng Xu, Andrew Perrault and Milind Tambe.
    “Collapsing Bandits and Their Application to Public Health Intervention”,
    [NeurIPS-20] Advances in Neural and Information Processing Systems (NeurIPS) 2020, Vancouver, Canada (* = Equal contribution). [paper] [code] [talk]
     
  • Perrault Andrew, Bryan Wilder, Eric Ewing, Aditya Mate, Bistra Dilkina and Milind Tambe.
    “End-to-End Game-Focused Learning of Adversary Behavior in Security Games”, 
    [AAAI-20] AAAI Conference on Artificial Intelligence 2020, New York, USA[arxiv]
     
  • Wang Kai, Andrew Perrault, Aditya Mate and Milind Tambe.
    “Scalable Game-Focused Learning of Adversary Models: Data-to-Decisions in Network Security Games”,
    [AAMAS-20] International Conference on Autonomous Agents and Multi-agent Systems, AAMAS 2020, Auckland, New Zealand. [paper]
     
  • Palvi Aggarwal, Omkar Thakoor, Aditya Mate, Milind Tambe, Edward A. Cranford, Christian Lebiere, and Cleotilde Gonzalez.
    “An Exploratory Study of a Masking Strategy of Cyberdeception Using CyberVAN.”
    [HFES-20] In 64th Human Factors and Ergonomics Society (HFES) Annual Conference. [paper]
     
  • B. Sombabu, Aditya Mate, D. Manjunath, Sharayu Moharir.
    “Whittle Index for AoI-aware scheduling”,
    [COMSNETS-20] In 12th International Conference on Communication Systems and Networks (COMSNETS). IEEE, 2020 [paper]

 

Journal Publications


[2020]
 

  • Bryan Wilder, Marie Charpignon, Jackson A Killian, Han-Ching Ou, Aditya Mate, Shahin Jabbari, Andrew Perrault, Angel Desai, Milind Tambe, and Maimuna S. Majumder.
    “Modeling between-population variation in COVID-19 dynamics in Hubei, Lombardy, and NewYork City”,
    [PNAS] Proceedings of the National Academy of Sciences (PNAS), 2020. [paper]
     

Workshops


[2020]
 

  • Aviva Prins, Aditya Mate, Jackson Killian, Rediet Abebe and Milind Tambe
    "Incorporating Healthcare Motivated Constraints in Restless Bandit Based Resource Allocation",
    [NeurIPS-20]  Workshop on Machine Learning for Health (ML4H), NeurIPS 2020, Vancouver, Canada
    Best Thematic Submission 
     
  • Aviva Prins, Aditya Mate, Jackson Killian, Rediet Abebe and Milind Tambe.
    "Incorporating Healthcare Motivated Constraints in Restless Bandit Based Resource Allocation",
    [NeurIPS-20]  Workshop on Challenges of Real World Reinforcement Learning (RWRL), NeurIPS 2020, Vancouver, Canada
     
  • Aviva Prins, Aditya Mate, Jackson Killian, Rediet Abebe and Milind Tambe
    "Incorporating Healthcare Motivated Constraints in Restless Bandit Based Resource Allocation",
    [NeurIPS-20]  Workshop on Machine Learning in Public Health (MLPH), NeurIPS 2020, Vancouver, Canada
    Best Lightning Paper
     
  • Aditya Mate, Jackson A. Killian, Bryan Wilder, Marie Charpignon, Ananya Awasthi, Milind Tambe and Maimuna S. Majumder.
    "Evaluating COVID-19 Lockdown Policies For India: A Preliminary Modeling Assessment for Individual States",
    [KDD-20]  ACM SIGKDD 2020 Workshop on Humanitarian Mapping
     
  • Bryan Wilder, Marie Charpignon, Jackson Killian, Han-Ching Ou, Aditya Mate, Shahin Jabbari, Andrew Perrault, Angel Desai, Milind Tambe and Maimuna Majumder.
    “Integrating agent-based modeling and Bayesian inference to uncover between-population variation in COVID-19 dynamics”,
    [KDD-20]  In ACM SIGKDD 2020 Workshop on Humanitarian Mapping.
     
  • Bryan Wilder, Marie Charpignon, Jackson Killian, Han-Ching Ou, Aditya Mate, Shahin Jabbari, Andrew Perrault, Angel Desai, Milind Tambe and Maimuna Majumder. 
    “Bayesian inference of between-population variation in COVID-19 dynamics”,
    [ICML-20] Workshop on Machine Learning for Global Health, International Conference on Machine Learning. 2020.
     
  • Aditya Mate*, Jackson A. Killian*, Haifeng Xu, Andrew Perrault and Milind Tambe.
    "Building Decision Aids for Community Health Workers: Optimizing Interventions via RestlessBandits",
    [AAMAS-20]  OptLearnMAS, AAMAS 2020 Workshop, Auckland, New Zealand

[2019]
 

  • Kai Wang, Aditya Mate, Bryan Wilder, Andrew Perrault, and Milind Tambe. 2019.
    Using Graph Convolutional Networks to Learn Interdiction Games .
    [IJCAI-19] In AI for Social Good workshop (AI4SG) at International Joint Conference on Artificial Intelligence (IJCAI) 2019. [paper]
     
  • Andrew Perrault, Bryan Wilder, Eric Ewing, Aditya Mate, Bistra Dilkina, and Milind Tambe. 2019.
    “Decision-Focused Learning of Adversary Behavior in Security Games.”
    [AAMAS-19]  In GAIW: Games, Agents and Incentives Workshop at International Conference on Autonomous Agents and Multiagent Systems (AAMAS-19). [paper]

     

Pre-prints


[2020]
 

  • Aditya Mate, Jackson A. Killian, Bryan Wilder, Marie Charpignon, Ananya Awasthi, Milind Tambe, and Maimuna S. Majumder. “Evaluating COVID-19 Lockdown Policies For India: A Preliminary Modeling Assessment for Individual States.” SSRN. [paper] [code]

 

News Coverage:
 

  • Sakal Media House coverage, May 2020: Middle ground for India's lockdown situation [article]
     
  • Nature India coverage, April 2020: "Model finds 'middle ground' for India's lockdown exit" [article]