Ari Kobren

I am a research scientist in the Machine Learning Research Group of Oracle Labs in Burlington, MA. I am interested in studying how computational tools--especially those that are powered by machine learning--can be used safely and to promote social good. I am broadly interested in machine learning, NLP, efficient algorithms, and data structures.

Before Oracle, I completed a Ph.D. at UMass Amherst working in the Information Extraction and Synthesis Laboratory with Professor Andrew McCallum. I received a B.S. in Computer Science from Tufts University and worked as a researcher at MIT Lincoln Laboratory building intelligent decision support systems. I spent the summers of 2014 and 2016 interning at Google.

Publications

• Upstream Mitigation is Not All You Need: Testing the Bias Transfer Hypothesis in Pre-Trained Language Models.
  Steed, R., Panda, S., Kobren, A., Wick, M.
  Annual Meeting of the Association for Computational Linguistics (ACL). Dublin, Ireland. May, 2022.
  [bibtex]
• An Evaluative Measure of Clustering Methods Incorporating Hyperparameter Sensitivity.
  Mishra, S., Monath, N., Boratko, M., Kobren, A., McCallum, A.
  Conference on Artificial Intelligence (AAAI). Vancouver, BC, Canada. February, 2022.
  [bibtex]
• Online Post-Processing in Rankings for Fair Utility Maximization.Oral Presentation
  Gupta, A., Johnson, E., Roy, A. K., Payan, J., Kobren, A., Panda, S., Tristan, J.-B., Wick, M.
  Web Search and Data Mining (WSDM). Virtual. March, 2021.
  [bibtex]
• Leveraging Extracted Model Adversaries for Improved Black Box Attacks.
  Nizar, N. J., Kobren, A.
  Analyzing and Interpreting Neural Networks for NLP Virtual. November, 2020.
  [arXiv][bibtex]
• Predicting Institution Hierarchies with Set-based Models.
  Tam, D., Monath, N., Kobren, A., McCallum, A.
  Automated Knowledge Base Construction. Virtual. May, 2020.
  [bibtex]
• Offline Contextual Bandits with High Probability Fairness Guarantees.
  Metevier, B., Giguere, S., Brockman, S., Kobren, A., Brun, Y., Brunskill, E., Thomas, P.
  Neural Information Processing Systems. Vancouver, Canada. December 2019.
  [bibtex][press]
• Optimal Transport-based Alignment of Learned Character Representations for String Similarity. (Oral Presentation)
  Tam, D., Monath, N., Kobren, A., Traylor, A., Das, R., McCallum, A.
  Association of Computational Linguistics. Florence, Italy. July 2019.
  [arXiv][bibtex]
• Scalable Hierarchical Clustering with Tree Grafting. (Oral Presentation)
  Monath, N.*, Kobren, A.*, Krishnamurthy, A., Glass, M., Mccallum, A.
  The International Conference on Knowledge Discovery and Data Mining. Anchorage, Alaska. August 2019.
  [arXiv][bibtex][code]
• Paper Matching with Local Fairness Constraints. (Oral Presentation)
  Kobren, A., Saha, B., McCallum, A.
  The International Conference on Knowledge Discovery and Data Mining. Anchorage, Alaska. August 2019.
  [arXiv][bibtex][code][promo video][Press]
• Constructing High Precision Knowledge Bases with Subjective and Factual Attributes.
  Kobren, A., Barrio, P., Yakhnenko, O., Hibschman, J., Langmore, I.
  The International Conference on Knowledge Discovery and Data Mining. Anchorage, Alaska. August 2019.
  [arXiv][bibtex]
• Supervised Hierarchical Clustering with Exponential Linkage.
  Yadav, N., Kobren, A., Monath, N., McCallum, A.
  International Conference on Machine Learning (ICML). Long Beach, CA. June 2019.
  [arXiv][bibtex][supplement][code]
• Integrating User Feedback under Identity Uncertainty in Knowledge Base Construction.
  Kobren, A., Monath, N., McCallum, A.
  Automated Knowledge Base Construction (AKBC). Amherst, MA. May 2019.
  [bibtex]
• Compact Representation of Uncertainty In Clustering.
  Greenberg, C., Monath, N., Kobren, A., Flaherty, P., McGregor, A., McCallum, A.
  Neural Information Processing Systems. Montreal, Canada. December 2018.
  [bibtex]
• Gradient-based Hierarchical Clustering. (Oral Presentation)
  Monath, N.*, Kobren, A*., Krishnamurthy, A., McCallum, A. (*Equal Contribution)
  NIPS 17' Workshop on Discrete Structures in Machine Learning. Long Beach, California. December 2017.
  [bibtex]
• Entity-centric Attribute Feedback for Interactive Knowledge Bases.
  Kobren, A., Monath, N., McCallum, A. NIPS 17' Workshop on Automated Knowledge Base Construction. Long Beach, California. December 2017.
  [bibtex]
• A Hierarchical Algorithm for Extreme Clustering. (Oral Presentation)
  Kobren, A*., Monath, N.*, Krishnamurthy, A., McCallum, A. (*Equal Contribution)
  The International Conference on Knowledge Discovery and Data Mining. Halifax, Nova Scotia. August 2017.
  [arXiv][bibtex][code][promo video][talk]
• Getting More for Less: Optimized Crowdsourcing with Dynamic Tasks and Goals.
  Kobren, A., Tan, C.H., Ipeirotis, P., Gabrilovich, E.
  The International Conference on the World Wide Web. Florence, Italy. May 2015.
  [bibtex]
• Domain Specific Knowledge Base Construction via Crowdsourcing. (Outstanding Paper Award)
  Kobren, A., Logan, T., Sampangi, S., McCallum, A.
  NIPS '14 Workshop on Automated Knowledge Base Construction. Montreal, Canada. December 2014.
  [bibtex]
• Universal Schema for Slot Filling and Cold Start: UMass IESL at TACKBP 2013.
  Singh, S., Yao, L., Belanger, D., Kobren, A., Anzaroot, S., Wick, M., Passos, A., Pandya, H., Choi, J., Martin, B., and McCallum, A.
  Text Analysis Conference Knowledge Base Population Track. Gaithersburg, Maryland, USA. November 2013.
  [bibtex]
• Assessing Confidence of Knowledge Base Content With an Experimental Study in Entity Resolution.
  Wick, M.L., Singh, S., Kobren, A. and McCallum, A.
  CIKM '13 Workshop on Automated Knowledge Base Construction. San Francisco, CA, USA. October 2013.
  [bibtex]
• Large-scale Author Coreference via Hierarchical Entity Representations.
  Wick, M.L., Kobren, A. and McCallum, A.
  ICML '13 Workshop on Peer Reviewing and Publishing Models. Atlanta, GA, USA. June 2013.
  [bibtex]
• Probabilistic Reasoning about Human Edits in Information Integration.
  Wick, M.L., Kobren, A. and McCallum, A.
  ICML '13 Workshop on Machine Learning Meets Crowdsourcing. Atlanta, GA, USA. June 2013.
  [bibtex]

---

Invited Talks

• Online Algorithms for Extreme Clustering. Tufts CS Rising Star Colloquium Series. October, 2018. [Announcement][Slides]
• Getting More for Less: Optimized Crowdsourcing with Dynamic Tasks and Goals. UMass Machine Learning and Friends Lunch Seminar Series. February, 2015. [Announcement]

About Me

I received a B.S. in Computer Science from Tufts University and worked as a researcher at MIT Lincoln Laboratory building intelligent decision support systems for US intelligence analysts. I spent the summer of 2014 at Google MTV training models of worker-task interaction that I used to optimize a new crowdsourcing platform. I spent the summer of 2016 at Google NYC building deep networks to predict attributes of the locations in Google Maps.

Researcher.cc

I was the chief designer and maintainer of ReSearcher.cc, a database of scientific affiliations and collaborations and conference management. Our backend integration engine dynamically produced ReSearcher's content by synthesizing multiple sources of data including edits from users. ReSearcher was used by ICCV, ECCV, CVPR and MICCAI to help identify conflicts of interest in peer review and automatically produce matchings of papers to reviewers. The tool is no longer supported, but its functionality has been implemented in OpenReview.net, a fully-featured and customizable conference management tool, which I am also involved with.

Mentees, Students, and Interns

• Johnny Wei
• Siddarth Sampangi
• Aarsh Patel
• Rakshitha Bhat
• Nithya Govindarajan
• Aditya Srinivasan
• Chetan Manjesh
• Lynn Samson
• David Ter-Ovanesyan
• Saranya Krishnakumar
• Derek Tam
• Chen Pan
• Mohit Surana
• Jad Salem
• Ryan Steed

Service

• AAAI '22 Reviewer
• ICML '21 Reviewer
• KDD '21 Reviewer
• NeurIPS '21 Reviewer
• ICML '20 Reviewer (Top 33% of Reviewers)
• KDD '20 Reviewer
• NeurIPS '20 Reviewer
• NeurIPS '19 Conference, Reviewer
• EMNLP '19 Conference, Reviewer
• ICML '19 Conference, Reviewer
• AKBC '19 Conference, Local Co-Chair & Webmaster
• NAACL '19 Conference, Reviewer
• NIPS '18 Reviewer (Top 30% of Reviewers)
• Machine Learning and Friends Lunch, Organizer, 2017-2018
• TKDD '17, Reviewer
• WSDM '16, Reviewer
• EMNLP '15, Reviewer

Other Fun Stuff

In my spare time, I enjoy cooking (especially ethnic foods that involve uncommon spices), learning for free, climbing, playing soccer and other sports, composing/playing music, fermenting things, snowboarding, playing German-style board games, traveling the world, listening to NPR (e.g. The Moth Radio Hour and wNYC's RadioLab), eating gummy candy, and learning about philosophy, food systems, religion, and nutrition.

In a previous life, I was a varsity college soccer player, I co-found a start-up and was a junior olympic Tae Kwon Do champ.

My favorite book is Zen and the Art of Motorcycle Maintenance.

---