Jesse Cresswell

Jesse Cresswell

Sr. Machine Learning Scientist

Enterprise ML grounded in fundamental research

I am a Staff Research Machine Learning Scientist with Layer 6 AI at TD, and lead the development of Trustworthy AI.

Designing trustworthy machine learning models is crucial in highly regulated industries like banking and healthcare. My research group at Layer 6 AI is advancing TD’s capabilities in privacy, fairness, robustness, and explainability, both for predictive and generative AI.

I have deep expertise in generative modelling - especially through the lens of the manifold hypothesis - conformal prediction, self-supervised learning, differential privacy, and federated learning.

News:

  • Dec. 2024 - I’ll be serving as an Area Chair for ICML 2024
  • Oct. 2024 - With NeurIPS workshop decisions out, I had four papers accepted to the TRL, FITML, ATTRIB, and SSL workshops.
  • Sept. 2024 - We had our paper on estimating local intrinsic dimension selected for a Spotlight at NeurIPS 2024
  • Aug. 2024 - I’ll be an Area Chair for next year’s ICLR 2025 in Singapore
  • July 2024 - This month two papers were accepted to TMLR, first my collaborative work with Vector Institute researchers, and second a comprehensive survey of deep generative modelling through the lens of the manifold hypothesis.
  • June 2024 - I had eight workshop papers accepted at ICML with a couple of them as Oral (*) presentations. SPIGM*, DAE*, GRaM x2, HAS, NGAIS, AI4Science, FMWild.
  • May 2024 - Two papers accepted to ICML 2024, one on conformal prediction and another that solves the longstanding paradox around OOD detection with deep generative models.
  • Apr. 2024 - I’ll be serving as an Area Chair for NeurIPS 2024
  • Mar. 2024 - We had a paper selected for a Highlight at CVPR 2024, and two posters accepted to the ICLR workshops BGPT and GenAI4DM
  • Feb. 2024 - At TD Bank’s annual patent awards I was nominated for both Inventor of the Year, and Invention of the year, while crossing the half-century mark for patents filed
  • Jan. 2024 - Another Poster accepted to ICLR 2024, which also had an Oral at the NeurIPS Table Representation Workshop
  • Dec. 2023 - I’m happy to say that our paper on topology-aware learning is published in TMLR
  • Dec. 2023 - I was named a Top Reviewer at NeurIPS 2023. I’ll also be reviewing for ICML 2024
  • Sept. 2023 - One Poster accepted at NeurIPS, a massive project with 10 authors from Layer 6 AI
  • May 2023 - Nature Communications has published my paper on decentralized federated learning with differential privacy guarantees
  • Jan. 2023 - We had two papers accepted at ICLR 2023 including one Spotlight (notable top 25%) and a Poster
  • Oct. 2022 - Five papers submitted to various NeurIPS workshops were all accepted this year, one with an oral and another spotlight. AFCP, ICBINB, FL, ML4PS, NeurReps
  • Sept. 2022 - I’ll be serving as a Program Chair for the federated learning workshop at this year’s NeurIPS
  • June 2022 - Two papers accepted to the ICML workshops TAG-ML and TPDP, headed to an in-person conference for the first time in years
  • Feb. 2022 - Some joint work on privacy enhancing technologies with Graham Taylor’s lab was accepted in Nature Scientific Reports
  • Aug. 2021 - We had a paper accepted at this year’s NeurIPS conference
  • Jan. 2021 - Our paper on reinforcement learning was accepted to ICLR
  • Sept. 2019 - Jumping full-time into ML, I joined Layer 6 AI, part of TD Bank, as a Machine Learning Scientist
  • Aug. 2019 - I defended my thesis to complete my PhD in physics at UofT
  • Nov. 2017 - I was very pleased to win a Vanier Canada Graduate Scholarship, Canada’s most competitive PhD level scholarship!