Anastasiya Belyaeva

Ph.D. Candidate
Massachusetts Institute of Technology
Email: belyaeva [at] mit.edu
[Google scholar] [Github] [CV]

I am a Ph.D. student at MIT Statistics and Data Science Center and Laboratory for Information and Decision Systems , advised by Caroline Uhler. My research focuses on developing and applying machine learning methods such as deep learning, representation learning, causal inference, and network analysis to large-scale data sets, with a special interest in biology & health.

Work Experience

  • Google Research, Research Intern
    June 2020 - Aug 2020
    I developed personalized computer vision models for gaze estimation from mobile phones, improving error by up to 67%.
  • Google Brain, Software Engineering Intern
    June 2019 - Sep 2019
    I developed semi-supervised deep generative models for sequences to enable more accurate receptor binding classification and generation.
  • Microsoft Research, Research Extern
    Jan 2019 - Feb 2020
    I worked on building predictive and interpretable supervised models for patient survival from high-dimensional data.

Publications

  1. Multi-Domain Translation between Single-Cell Imaging and Sequencing using Autoencoders. [PDF]
    K. D. Yang*, A. Belyaeva*, S. Venkatachalapathy, K. Damodharan, A. Katcoff, A. Radhakrishnan, G.V. Shivashankar and C. Uhler
    To appear in Nature Communications. (2020).
  2. Causal Network Models of SARS-CoV-2 Expression and Aging to Identify Candidates for Drug Repurposing. [PDF]
    A. Belyaeva*, L. Cammarata*, A. Radhakrishnan*, C. Squires, K.D. Yang, G.V. Shivashankar and C. Uhler.
    To appear in Nature Communications. Oral presentation at Machine Learning in Computational Biology Workshop 2020.
  3. Deconvolution of Bulk Genomics Data using Single-cell Measurements via Neural Networks. 
    A. Belyaeva and C. Uhler.
    NeurIPS Learning Meaningful Representations of Life Workshop (LMRL 2020).
  4. DCI: Learning Causal Differences between Gene Regulatory Networks. [PDF]
    A. Belyaeva, C. Squires, and C. Uhler.
    Under review at Bioinformatics. (2020).
  5. Anchored Causal Inference in the Presence of Measurement Error. [PDF]
    B. Saeed, A. Belyaeva, Y. Wang, and C. Uhler.
    Uncertainty in Artificial Intelligence (UAI 2020).
  6. Direct Estimation of Differences in Causal Graphs. [PDF]
    Y. Wang, C. Squires, A. Belyaeva and C. Uhler.
    Neural Information Processing Systems (NeurIPS 2018).
  7. Network Analysis Identifies Chromosome Intermingling Regions as Regulatory Hotspots for Transcription. [PDF]
    A. Belyaeva, S. Venkatachalapathy, M. Nagarajan, G.V. Shivashankar and C. Uhler.
    Proceedings of the National Academy of Sciences (PNAS 2017).

Research Highlights

Multi-Domain Translation between Single-Cell Imaging and Sequencing using Autoencoders

We present an approach for integrating vastly different modalities by learning a probabilistic coupling between the different data modalities using autoencoders to map to a shared latent space. [PDF]

Causal Network Models of SARS-CoV-2 Expression and Aging to Identify Candidates for Drug Repurposing

Given the severity of the SARS-CoV-2 pandemic, a major challenge is to rapidly repurpose existing approved drugs for clinical interventions. We propose a computational framework that predicts effect of a drug in a cell type of interest by using an autoencoder and style transfer/synthetic interventions trained on large-scale transcriptional drug screens. Furthermore, we provide putative drug mechanism by performing causal inference on the disease network.
[PDF]

Anchored Causal Inference in the Presence of Measurement Error

We develop a provably consistent algorithm for learning a causal graph in the presence of measurement error. We apply this method to learn gene regulatory networks from zero-inflated single-cell RNA-seq data.
[PDF]

Direct Estimation of Differences in Causal Graphs

We design an algorithm for directly learning differences between causal graphs from two conditions/data sets without learning each potentially dense graph separately.
[PDF]