Polina Kirichenko

I am a PhD student at New York University Center for Data Science working with professor Andrew Gordon Wilson. I am also a Visiting Researcher at Meta AI (FAIR Labs) mentored by Ramakrishna Vedantam, as a part of the FAIR-NYU AI Mentorship program. I am working on deep learning robustness to distribution shift and out-of-distribution generlization. I am also broadly interested in probabilistic deep learning learning, uncertainty estimation and generative models.

Before joining NYU, I was a PhD student in Operations Research and Information Engineering at Cornell University. I received a BS in Computer Science at Higher School of Economics. During my undergrad, I worked at Bayesian Methods research group with professor Dmitry Vetrov.

In summer 2022, I interned at Meta AI with Hamed Firooz and Randall Balestriero and worked on biases of data augmentations. In summer 2021, I interned at Cold Spring Harbor Laboratory with Tony Zador and worked on meta-learning and neuroscience applications. In summer 2020, I interned at DeepMind with Mehrdad Farajtabar, Razvan Pascanu and Balaji Lakshminarayanan, and worked on task-agnostic continual learning. In summer 2018, I interned at Machine Learning and Optimization Lab at EPFL and worked with professors Martin Jaggi and Dan Alistarh on low-precision training of neural networks. I also spent two summers (2016, 2017) at Google as a Software Engineering intern.

Publications

• On Feature Learning in the Presence of Spurious Correlations
  Pavel Izmailov*, Polina Kirichenko*, Nate, Gruver*, Andrew Gordon Wilson
  Neural Information Processing Systems (NeurIPS), 2022
  [arXiv]
• Chroma-VAE: Mitigating Shortcut Learning with Generative Classifiers
  Wanqian Yang, Polina Kirichenko, Micah Goldblum, Andrew Gordon Wilson
  Neural Information Processing Systems (NeurIPS), 2022
• Last Layer Re-Training is Sufficient for Robustness to Spurious Correlations
  Polina Kirichenko*, Pavel Izmailov*, Andrew Gordon Wilson
  2022
  [arXiv]
• Does Knowledge Distillation Really Work?
  Samuel Stanton, Pavel Izmailov, Polina Kirichenko, Alexander A. Alemi, Andrew Gordon Wilson
  Neural Information Processing Systems (NeurIPS), 2021
  [arXiv]
• Task-agnostic Continual Learning with Hybrid Probabilistic Models
  Polina Kirichenko, Mehrdad Farajtabar, Dushyant Rao, Balaji Lakshminarayanan, Nir Levine, Ang Li, Huiyi Hu, Andrew Gordon Wilson, Razvan Pascanu
  ICML Workshop on Invertible Neural Networks, Normalizing Flows, and Explicit Likelihood Models (spotlight talk), 2021
  [arXiv, poster]
• Why Normalizing Flows Fail to Detect Out-of-Distribution Data
  Polina Kirichenko*, Pavel Izmailov*, Andrew Gordon Wilson
  ICML workshop on Invertible Neural Networks and Normalizing Flows, 2020
  Neural Information Processing Systems (NeurIPS), 2020
  [arXiv, poster]
• Semi-Supervised Learning with Normalizing Flows
  Pavel Izmailov*, Polina Kirichenko*, Marc Finzi*, Andrew Gordon Wilson
  ICML workshop on Invertible Neural Nets and Normalizing Flows, 2019
  14th Women in Machine Learning workshop (co-located with NeurIPS), 2019
  International Conference on Machine Learning (ICML), 2020
  [arXiv, PMLR, code, poster]
• Subspace Inference for Bayesian Deep Learning
  Pavel Izmailov*, Wesley Maddox*, Polina Kirichenko*, Timur Garipov*, Dmitry Vetrov, Andrew Gordon Wilson
  ICML workshop on Uncertainty & Robustness in Deep Learning (contributed talk), 2019
  Uncertainty in Artificial Intelligence (UAI), 2019
  [arXiv, code, UDL poster, UAI poster]
• SWALP: Stochastic Weight Averaging in Low-Precision Training
  Guandao Yang, Tianyi Zhang, Polina Kirichenko, Junwen Bai, Andrew Gordon Wilson, Christopher De Sa
  International Conference on Machine Learning (ICML), 2019
  [arXiv, PMLR, code]

Workshop Papers

• Effective Surrogate Models for Protein Design with Bayesian Optimization
  Nate Gruver, Samuel Stanton, Polina Kirichenko, Marc Finzi, Phillip Maffettone, Vivek Myers, Emily Delaney, Peyton Greenside, Andrew Gordon Wilson
  ICML Workshop on Computational Biology, 2021
  [workshop PDF]
• Invertible Convolutional Networks
  Marc Finzi, Pavel Izmailov, Polina Kirichenko, Andrew Gordon Wilson
  ICML workshop on Invertible Neural Nets and Normalizing Flows (splotlight talk), 2019
  [workshop PDF, poster, slides]

Awards

• DeepMind Fellowship 2019-2020
• Google Anita Borg Memorial Scholarship 2015 (Women Techmakers Scholarship)
• Golden HSE Award (Alumni Success: Best Academic Achievement) 2019
• Ilya Segalovich Scholarship (Yandex) 2016, 2017

Talks

• Last Layer Re-Training is Sufficient for Robustness to Spurious Correlations
  Oral presentation at ICML 2022 Workshop on Spurious Correlations, Invariance, and Stability
  [video]
• Applications of normalizing flows: semi-supervised learning, anomaly detection, and continual learning
  Keynote talk at ICML 2021 Workshop on Representation Learning for Finance and E-Commerce Applications
  [video]
• Task-agnostic Continual Learning with Hybrid Probabilistic Models
  Spotlight talk at INNF+ workshop at ICML 2021
  [video]
• Continual Learning in Neural Networks: on Catastrophic Forgetting and Beyond
  Bayesian Methods Research Group seminar (in Russian)
  [video, slides]
• Understanding Semantic Anomaly Detection with Generative Networks
  ML Collective 2021, Deep Learning: Classics and Trends
  [slides]
• Normalizing flows for anomaly detection
  CogSys Talks, Technical University of Denmark
  [video]
• Anomaly detection via Generative Models
  Open Data Science DafaFest 2020, Uncertainty in ML Workshop
  [video]
• Why Normalizing Flows Fail to Detect Out-of-Distribution Data
  INNF+ workshop at ICML 2020; NeurIPS 2020
  [ICML video, NeurIPS video]
• Scalable Bayesian Inference in Low-Dimensional Subspaces
  Higher School of Economics
  [video]
• How do we build neural networks we can trust?
  Broad Institute of MIT and Harvard
  [video, slides]
• Subspace Inference for Bayesian Deep Learning
  ICML workshop on Uncertainty & Robustness in Deep Learning
  [video, slides]