MABSplit: Faster Forest Training Using Multi-Armed Bandits Mo Tiwari, Ryan Kang*, Je-Yong Lee*, Chris Piech, Ilan Shomorony, Sebastian Thrun, Martin Jinye Zhang In submission.
NL-Augmenter: A Framework for Task-Sensitive Natural Language Augmentation Kausthubh D. Dhole, ..., Mo Tiwari, ..., Yue Zhang (122 authors) [arxiv] In submission.
Beyond the Imitation Game: Quantifying and Extrapolating the Capabilities of Language Models Aarohi Srivastava, ..., Mo Tiwari, ..., Ziyi Wu (444 authors, listed alphabetically) [preprint] In submission.
GFlowNet Foundations Yoshua Bengio, Tristan Deleu, Edward J. Hu, Salem Lahlou, Mo Tiwari, Emmanuel Bengio [arxiv] In submission.
Image Compression and Classification Using Qubits and Quantum Deep Learning Ali Mohsen, Mo Tiwari [arxiv] In submission.
BanditPAM: Almost Linear Time k-medoids Clustering via Multi-Armed Bandits Mo Tiwari, Martin Jinye Zhang, James Mayclin, Sebastian Thrun, Chris Piech, Ilan Shomorony [Conference paper] [Blog post] Neural Information Processing Systems (NeurIPS) 2020.
Differentiation of Active Corneal Infections from Healed Scars Using Deep Learning Mo Tiwari, Chris Piech, Namperumalsamy Prajna, Muthiah Srinivasan, Prajna Lalith, Natacha C. Villegas, Janice T. Chua, Travis K. Redd, Thomas M. Lietman, Sebastian Thrun, Charles C. Lin [Journal paper] Ophthalmology 2020. Best Poster Award at associated conference, American Academy of Ophthalmology Conference (AAO) 2020.
Additional technical reports for unpublished work are available upon request, especially for less recent work in quantum computation/physics and machine learning.
Highlights of Algorithms (HALG21) Conference "BanditPAM: Almost Linear Time k-Medoids Clustering via Multi-Armed Bandits." A conference presentation of our NeurIPS 2020 paper, BanditPAM. Virtual talk.
U.S. Food and Drug Administration (FDA) "BanditPAM: Almost Linear Time k-Medoids Clustering via Multi-Armed Bandits." An FDA-wide talk I gave about our NeurIPS 2020 paper, BanditPAM. Part 2 of 2. Virtual talk.
U.S. Food and Drug Administration (FDA) "An Introduction to Clustering and Multi-Armed Bandits" An FDA-wide introductory talk I gave covering common clustering techniques and multi-armed bandits. Part 1 of 2. Virtual talk.
Twitch "Novel Data Augmentation, Multi-Armed Bandits, and more: New Machine Learning Techniques for Twitch Safety" Presentation to the Safety organization at Twitch on how to use new data augmentation and multi-armed bandit methods for their uses cases. Virtual talk.
C3.ai "k-medoids Clustering and Multimodal Data Augmentation" Given to the ML and product teams at C3.ai to discuss our ongoing research efforts and ways in which they might be applicable to C3.ai's use cases.
Facebook "ThreatExchange v2.8 Webinar" A webinar for new features for the product I worked on at Facebook. Joint presentation with Julian Nagler and Amir Naor. Facebook Headquarters in Menlo Park, CA, USA in December 2016. https://www.youtube.com/watch?v=SVVC4ZLYHmk
Microsoft Security Research Alliance "Tracking Advanced Persistent Threats With ThreatExchange" Presented to 50+ security professionals; discussed and demonstrated tracking advanced persistent threats (APTs) using tools we built at Facebook. Joint presentation with April Eubank. Microsoft Headquarters in Seattle, WA, USA in July 2016.
BanditPAM A high-performance Python package, written in C+, that implements the algorithm from our NeurIPS 2020 paper and is pip-installable via "pip install banditpam". Primary author, 240+ stars.
BIG-Bench A set of benchmark tasks meant to probe the capabilities of large language models.
NL-Augmenter A set of data augmentations and filters for natural language data.
Course Assistant for Client-Side Technologies (CS193C) Graded assignments, provided feedback, and answered questions for over 100 students each quarter during the summers of 2020 and 2021.
EDGE Mentor Mentored three early Ph.D. students in Computer Science at Stanford University via a formal, funded position.
Ph.D. Student Mentor Managed over a dozen undergraduate and M.S. students at Stanford University. Upward performance reviews available upon request.