Education

Publications

  1. [J1] RePReL: A Unified Framework for Integrating Relational Planning and Reinforcement Learning for Effective Abstraction in Discrete and Continuous Domains,
    Harsha Kokel, Sriraam Natarajan, Balaraman Ravindran, Prasad Tadepalli, In Neural Computing and Applications 2022.
    paper | DOI | cite
    @article{KokelNRT22,
       author={Harsha Kokel and Sriraam Natarajan and Balaraman Ravindran and Prasad Tadepalli},
       journal={Neural Computing and Applications},
       title={RePReL: a unified framework for integrating relational planning and reinforcement learning for effective abstraction in discrete and continuous domains},
       year={2022},
       doi={https://doi.org/10.1007/s00521-022-08119-y},
       }

  2. [C8] LARA -- Human-guided collaborative problem solver: Effective integration of learning, reasoning and communication,
    Harsha Kokel, Mayukh Das, Rakibul Islam, Julia Bonn, Jon Cai, Soham Dan, Anjali Narayan-Chen, Prashant Jayannavar, Janardhan Rao Doppa, Julia Hockenmaier, Sriraam Natarajan, Martha Palmer, Dan Roth, In The Tenth Annual Conference on Advances in Cognitive Systems (ACS) 2022.
    paper | blog | supplemental | video | cite
    @article{KokeletalACS22,
       title={LARA -- Human-guided collaborative problem solver: Effective integration of learning, reasoning and communication},
       author={Harsha Kokel, Mayukh Das, Rakibul Islam, Julia Bonn, Jon Cai, Soham Dan, Anjali Narayan-Chen, Prashant Jayannavar, Janardhan Rao Doppa, Julia Hockenmaier, Sriraam Natarajan, Martha Palmer, Dan Roth},
       journal={The Tenth Annual Conference on Advances in Cognitive Systems (ACS)},
       year={2022},
       }

  3. [C7] ORIENT: Submodular Mutual Information Measures for Data Subset Selection under Distribution Shift,* equal contributions
    Athresh Karanam*, Krishnateja Killamsetty*, Harsha Kokel*, Rishabh K Iyer, In NeurIPS 2022.
    paper | code | supplemental | DOI | cite
    @inproceedings{KaranamKKI22,
       author = {Athresh Karanam and Krishnateja Killamsetty and Harsha Kokel and Rishabh K Iyer},
       title = {{ORIENT}: Submodular Mutual Information Measures for Data Subset Selection under Distribution Shift},
       year = {2022},
       booktitle = {NeurIPS},
       }

  4. [C6] Hybrid Deep RePReL: Integrating Relational Planning and Reinforcement Learning for Information Fusion,
    Harsha Kokel, Nikhilesh Prabhakar, Balaraman Ravindran, Eric Blasch, Prasad Tadepalli, Sriraam Natarajan, In IEEE 25th International Conference on Information Fusion (FUSION) 2022.
    paper | blog | slide | DOI| video | cite
    @inproceedings{KokelPRBTN22,
       title={Hybrid Deep RePReL: Integrating Relational Planning and Reinforcement Learning for Information Fusion},
       author={Harsha Kokel and Nikhilesh Prabhakar and Balaraman Ravindran and Eric Blasch and Prasad Tadepalli and Sriraam Natarajan},
       journal={IEEE 25th International Conference on Information Fusion (FUSION)},
       year={2022},
       doi={https://doi.org/10.23919/FUSION49751.2022.9841246},
       }

  5. [W5] Action Space Reduction for Planning Domains,
    Harsha Kokel, Junkyu Lee, Michael Katz, Shirin Sohrabi, Kavitha Srinivas, In Planning and RL (PRL) Workshop at ICAPS 2022.
    paper | slide | video
  6. [A2] Dynamic probabilistic logic models for effective task-specific abstractions in RL (Abstract),
    Harsha Kokel, Sriraam Natarajan, Balaraman Ravindran, Prasad Tadepalli, In 5th Multidisciplinary Conference on Reinforcement Learning and Decision Making (RLDM) 2022.
    paper | code | poster
  7. [A1] How to Reduce Action Space for Planning Domains? (Student Abstract),* oral presentation (20/111 accepted abstracts)
    Harsha Kokel, Junkyu Lee, Michael Katz, Shirin Sohrabi, Kavitha Srinivas, In AAAI 2022.
    paper | DOI| video | cite
    @article{KokelLKSS22,
       title={How to Reduce Action Space for Planning Domains? (Student Abstract)},
       author={Kokel, Harsha and Lee, Junkyu and Katz, Michael and Sohrabi, Shirin and Srinivas, Kavitha},
       journal={AAAI},
       pages={12989--12990},
       year={2022},
       url={https://ojs.aaai.org/index.php/AAAI/article/view/21631},
       }

  8. [W4] Deep RePReL-Combining Planning and Deep RL for acting in relational domains,
    Harsha Kokel, Arjun Manoharan, Sriraam Natarajan, Balaraman Ravindran, Prasad Tadepalli, In Deep RL Workshop at NeurIPS 2021.
    paper | code | supplemental | slide | video | cite
    @article{KokelMNBTDRL,
       title={Deep RePReL-Combining Planning and Deep RL for acting in relational domains},
       author={Kokel, Harsha and Manoharan, Arjun and Natarajan, Sriraam and Ravindran, Balaraman and Tadepalli, Prasad},
       journal={Deep {RL} Workshop at {NeurIPS}},
       year={2021},
       }

  9. [W3] Dynamic probabilistic logic models for effective abstractions in RL,
    Harsha Kokel, Arjun Manoharan, Sriraam Natarajan, Balaraman Ravindran, Prasad Tadepalli, In Statistical Relational AI (StarAI) Workshop at IJCLR 2021.
    paper | poster | cite
    @article{KokelMNBTSTARAI,
       title={Dynamic probabilistic logic models for effective abstractions in RL},
       author={Kokel, Harsha and Manoharan, Arjun and Natarajan, Sriraam and Ravindran, Balaraman and Tadepalli, Prasad},
       journal = {CoRR},
       volume = {abs/2110.08318},
       year={2021},
       }

  10. [C5] Human-guided Collaborative Problem Solving: A Natural Language based Framework,
    Harsha Kokel, Mayukh Das, Rakibul Islam, Julia Bonn, Jon Cai, Soham Dan, Anjali Narayan-Chen, Prashant Jayannavar, Janardhan Rao Doppa, Julia Hockenmaier, Sriraam Natarajan, Martha Palmer, Dan Roth, In ICAPS (demo track) 2021.
    paper | blog | video | cite
    @article{KokeletalDemo,
       title={Human-guided Collaborative Problem Solving: A Natural Language based Framework},
       author={Harsha Kokel, Mayukh Das, Rakibul Islam, Julia Bonn, Jon Cai, Soham Dan, Anjali Narayan-Chen, Prashant Jayannavar, Janardhan Rao Doppa, Julia Hockenmaier, Sriraam Natarajan, Martha Palmer, Dan Roth},
       journal={Thirty First International Conference on Automated Planning and Scheduling ({ICAPS}) Demo Track},
       year={2021},
       }

  11. [W2] RePReL: Integrating Relational Planning and Reinforcement Learning for Effective Abstraction (Extended Abstract),* contributed talk (11/25 accepted paper)
    Harsha Kokel, Arjun Manoharan, Sriraam Natarajan, Balaraman Ravindran, Prasad Tadepalli, In Planning and RL (PRL) Workshop at ICAPS 2021.
    paper | poster | video
  12. [C4] A Probabilistic Approach to Extract Qualitative Knowledge for Early Prediction of Gestational Diabetes,
    Athresh Karanam, Alexander L. Hayes, Harsha Kokel, David M. Haas, Predrag Radivojac, Sriraam Natarajan, In AIME 2021.
    paper | blog | DOI| video | cite
    @inproceedings{KaranamHKHRN21,
       author = {Athresh Karanam and Alexander L. Hayes and Harsha Kokel and David M. Haas and Predrag Radivojac and Sriraam Natarajan},
       title = {A Probabilistic Approach to Extract Qualitative Knowledge for Early Prediction of Gestational Diabetes},
       year = {2021},
       publisher = {Springer International Publishing},
       pages = {497--502},
       booktitle = {Artificial Intelligence in Medicine},
       isbn = {978-3-030-77211-6},
       DOI = {10.1007/978-3-030-77211-6_59},
       }

  13. [C3] RePReL: Integrating Relational Planning and Reinforcement Learning for Effective Abstraction,
    Harsha Kokel, Arjun Manoharan, Sriraam Natarajan, Balaraman Ravindran, Prasad Tadepalli, In ICAPS 2021.
    paper | blog | code | supplemental | DOI| video | cite
    @article{KokelMNRT21,
       title={RePReL: Integrating Relational Planning and Reinforcement Learning for Effective Abstraction},
       author={Kokel, Harsha and Manoharan, Arjun and Natarajan, Sriraam and Balaraman, Ravindran and Tadepalli, Prasad},
       journal={Proceedings of the International Conference on Automated Planning and Scheduling},
       number={1},
       volume={31},
       year={2021},
       month={May},
       pages={533--541},
       url={https://ojs.aaai.org/index.php/ICAPS/article/view/16001},
       }

  14. [W1] Graph Sparsification via Meta-Learning,* contributed talk (4/22 accepted paper)
    Guihong Wan, Harsha Kokel, In Deep Learning for Graphs (DLG) Workshop at AAAI 2021.
    paper | slide | video | cite
    @article{WanK21,
       title={Graph Sparsification via Meta-Learning},
       author={Wan, Guihong and Kokel, Harsha},
       journal={Deep Learning for Graphs (DLG) Workshop at AAAI},
       year={2021}
       }

  15. [C2] A Unified Framework for Knowledge Intensive Gradient Boosting: Leveraging Human Experts for Noisy Sparse Domain,
    Harsha Kokel, Phillip Odom, Shuo Yang, Sriraam Natarajan, In AAAI 2020.
    paper | blog | code | supplemental | poster | slide | DOI | cite
    @article{KokelOYN20,
       author = {Harsha Kokel and Phillip Odom and Shuo Yang and Sriraam Natarajan},
       title = {A Unified Framework for Knowledge Intensive Gradient Boosting: Leveraging Human Experts for Noisy Sparse Domains},
       journal = {Proceedings of the AAAI Conference on Artificial Intelligence}
       volume = {34},
       number = {04},
       pages = {4460--4468},
       year = {2020},
       month = {Apr.},
       DOI = {10.1609/aaai.v34i04.5873},
       url = {https://aaai.org/ojs/index.php/AAAI/article/view/5873},
       }

  16. [C1] Morpheme Extraction Task,
    Rashmi Sankepally, Komal Agarwal, Harsha Kokel, Prasenjit Majumder, In Forum for Information Retrieval Evaluation (FIRE) 2013.
    paper | DOI | cite
    @inproceedings{SankepallyKAM13,
       author = {Rashmi Sankepally and Harsha Kokel and Komal Agarwal and Prasenjit Majumder},
       title = {Morpheme Extraction Task at {FIRE} 2012-2013},
       booktitle = {Proceedings of the Forum on Information Retrieval Evaluation, {FIRE}},
       pages = {5:1--5:4},
       publisher = {{ACM}},
       year = {2013},
       url = {https://doi.org/10.1145/2701336.2701637},
       doi = {10.1145/2701336.2701637},
       }

Experience

Industry

  • Research Intern, IBM T. J. Watson Research Center, Yorktown Heights, NY, USA   [ Summer 2021, Summer 2022]
    Bridging the gap between Symbolic planning and Reinforcement Learning.
    Manager: Dr. Shirin Sohrabi
  • ML Intern, Turvo Inc., CA, USA   [ Summer 2018 ]
    Modeled a cost estimator that leverages the knowledge of the domain experts. Kokel et al. AAAI 2020 was motivated by the work at Turvo.
  • Senior Software Engineer, Amadeus Software Labs, Bangalore, India   [ 2016-17 ]
    Implemented efficient low fare search for Air Canada.
  • Associate Technology, Publicis Sapient, Bangalore, India   [ 2013-16 ]
    Provided content management solutions for enhanced digital presence.

Academia

  • Instructor, UT Dallas, TX, USA   [ Fall 2021 ]
    Designed and organized labs for an 8-Week course on Advanced Machine Learning for engineers from Vistra Corporation.
  • Research Assistant, Starling Lab, UT Dallas, TX, USA   [ Spring 2019 - present ]
  • Teaching Assistant, UT Dallas, TX, USA   [ Fall 2018 ]
    CS6343: Artificial Intelligence (graduate level)
    CS4365: Artificial Intelligence (undergrad level)
  • Research Assistant, IR Lab, DA-IICT, Gandhinagar, India   [ 2012-13 ]
    Worked on Sandhan, a multilingual search engine for 8 Indian languages. Including cross-lingual search.
    Advisor: Prasenjit Majumder
  • Research Intern, Mobile Research Lab, NID, Gandhinagar, India   [ Summmer 2011 ]
    Developed a Point of Interest (POI) collector with Google maps for Android Smart Phones.
    Advisor: Jignesh Khakhar

Technical skills

    Python, PyTorch, JAX, Java, C, Shell Scripting, MATLAB, R, Linux/Unix, Git, SQL, Prolog, PDDL, Jupyter.

Projects

Link prediction using NeSy models

Compared different Neurosymbolic and symbolic models for a simple task of link prediction. More details.

Relational Q Learning

Implemented Q-learning for relational domains like blocksworld, where the states are represented using first-order logic predicates. [ code ]

Effective Abstraction in Relational RL

Providing state abstraction for better task transfer in taskable reinforcement learning environments using D-FOCI statements. [ code ]

Communicating with Computers

This is a DARPA funded project to build intelligent minecraft agent that can communicate and collaborate with humans through chat to build structures. [ video ]

Knowledge Intensive Gradient Boosting

Leveraging qualitative domain knowledge while learning tree based gradient boosting models to improve predictions in regions where data is noisy or absent. Kokel et al. AAAI 2020.

Learning Sparse Graph for GNN

Used meta-learning techniques to optimize the graph structure for obtaining sparse graph. Wan and Kokel, DLG 2021.

Causal inference from Protein Expression Data

Discovering causal molecular relationships from the evaluation of observational data using do-calculus. More details.

Ja-Walk-ER

Developed an interface that allows users with basic understanding of ER Diagrams to provide search bias for Inductive Logic Programming based. As described in Hayes et al. 2017. [ code ]

Expression Detection

A small project to detect wink and shush expression using OpenCV. [ code ]

RL for Healthcare

Learning polices for management of children on ECMO using batch reinforcement learning techniques. More details.

SRL model for credit default

Learnt and evaluated a statistical relational model for Kaggle Home credit default risk dataset and compared it with propositional models.

Optimization

Implementation of various first-order and second-order gradient methods for optimization. Including Barzilai-Borwein Gradient Descent, Conjugate Gradient Descent, Limited Memory BFGS and Armijo Line-search. [ code ]

ML/AI basket

My basket of Machine Learning and AI algorithms implemented over time. [ code ]

Talks

  • Invited talk, Dynamic probabilistic logic models for effective task-specific abstractions in RL at Robotics Lab at the Brown University, Sep 16, 2022.
    slide | video
  • Presented paper, Hybrid Deep RePReL: Integrating Relational Planning and Reinforcement Learning for Information Fusion at IEEE 25th International Conference on Information Fusion (FUSION), Linköping, Sweden, Jul 5, 2022.
    slide | video
  • Invited talk, RePReL: Integrating Relational Planning and Reinforcement Learning for Effective Abstractions at Eleventh RBCDSAI Workshop on Recent progress in Data Science and AI, IIT Madras, India, Nov 17, 2021.
    slide | video
  • Ph.D. mixer talk, Human Allied Artificial Intelligence at UT Dallas, USA, Sep 10, 2021.
    slide
  • Invited talk, A Unified Framework for Knowledge Intensive Gradient Boosting: Leveraging Human Experts for Noisy Sparse Domain at AAIR Lab Spring Meeting, ASU, USA, Jan 22, 2021.

Academic Service

  • Assistant Electronic Publishing Editors for JAIR 2020 - present.
  • PC Member for Workshop on Graphs and more Complex structures for Learning and Reasoning (GCLR) at AAAI 2022.
  • Reviewed papers for I Can’t Believe It’s Not Better (ICBINB) Workshop at NeurIPS 2021.
  • Reviewer for Data Mining and Knowledge Discovery Journal.
  • Volunteered at ICAPS 2021.
  • Volunteered at AAAI 2021.
  • Reviewer for Big Data Journal.
  • Reviewed papers for AAAI 2021.
  • Student volunteer at ICDE 2020.
  • Reviewed papers for CODS-COMAD, 2020 and SDM, 2020.
  • Helped organize meeting of Forum for Information Retrieval Evaluation (FIRE), 2018, 2013.
  • Conducted a lab on Information Retrieval with Terrier, an open source search engine, in MSR & IRSI Pre-FIRE workshop 2013 at Microsoft Research India.

Awards and Recognitions

  • Our team won Hackathon at Amadeus in November 2016
  • Certified Lead Developer for Adobe Experience Manager (AEM) 6.0 and 5.6
  • Star Employee of Sapient Bangalore, awarded in 2015 for the work on Loblaws project in the United Content Workz vertical
  • Nominated and Recognised at Sapient for exhibiting the Core Value of Client-Focused Delivery
  • Ranked among top 15% performer of the 2013 batch at Sapient and rewarded accordingly
  • Student Representative in Disciplinary action Committee (2012-2013)
  • Deputy convener of Cultural Committee (2012-2013)
  • Elected as Deputy Convener of the Cultural Committee in 2012-13 at DAIICT
  • Elected as the first student representative in the Disciplinary Action Committee at DAIICT for academic year 2012-13

My Github Contribution

My Github chart