Bahareh Tolooshams
Bahareh Tolooshams

PhD Candidate, Electrical Engineering
CRISP Group, Harvard University


I received B.ASc. degree with distinction in Electrical Engineering from the University of Waterloo in 2017. I am currently a PhD Candidate, advised by Demba Ba, at Harvard University.

I spent summer 2019 at Amazon AI as an Applied Scientist Intern working on neural architectures for multichannel speech enhancement. I have joined the Applied Sciences Group at Microsoft as a Research Intern for summer 2021; I work on stereo speech enhancement.

Research

My research interests are at the intersection of deep learning, statistical signal processing, and computational neuroscience. My PhD research is divided into three tracks:

  • model-based deep learning: I develop networks using probabilistic generative models and dictionary learning, applied to inverse problems such as compressive sensing and denoising.
  • deep learning theory: I develop theory for gradient dynamics and implicit acceleration.
  • computational neuroscience: I develop algorithms for a variety of problems such as neural encoding in the olfactory system and spike sorting.

Publications
2021
  1. PUDLE: Implicit acceleration of dictionary learning by backpropagation
    B. Tolooshams, and D. Ba
    Submitted [paper]
  2. Gaussian process convolutional dictionary learning
    A. H. Song, B. Tolooshams, and D. Ba
    Submitted [paper]
  3. On the convergence of group-sparse autoencoders
    E. Theodosis, B. Tolooshams*, P. Tankala*, A. Tasissa, and D. Ba
    Submitted [paper]
  4. Towards improving discriminative reconstruction via simultaneous dense and sparse coding
    A. Tasissa, E. Theodosis*, B. Tolooshams*, and D. Ba
    Submitted [paper] [code]
  5. Unfolding neural networks for compressive multichannel blind deconvolution
    B. Tolooshams*, S. Mulleti*, D. Ba, and Y. C. Eldar
    IEEE ICASSP [paper] [slides] [poster]
  6. Unsupervised learning of a dictionary of neural impulse responses from spiking data
    B. Tolooshams, H. Wu, P. Masset, V. N. Murthy, and D. Ba
    COSYNE [paper] [poster]
2020
  1. Convolutional dictionary learning based auto-encoders for natural exponential-family distributions
    B. Tolooshams*, A. H. Song*, S. Temereanca, and D. Ba
    ICML [paper] [code] [slides]
  2. Convolutional dictionary learning of stimulus from spiking data
    A. H. Song*, B. Tolooshams*, S. Temereanca, and D. Ba
    COSYNE [paper] [poster]
  3. Deep residual autoencoders for expectation maximization-inspired dictionary learning
    B. Tolooshams, S. Dey, and D. Ba
    IEEE Transactions on Neural Networks and Learning Systems [paper] [code]
  4. Channel-attention dense u-net for multichannel speech enhancement
    B. Tolooshams, R. Giri, A. H. Song, U. Isik, and A. Krishnaswamy
    IEEE ICASSP [paper]
2019
  1. Convolutional dictionary learning in hierarchical networks
    J. Zazo, B. Tolooshams, and D. Ba
    IEEE CAMSAP [paper]
  2. RandNet: deep learning with compressed measurements of images
    T. Chang*, B. Tolooshams*, and D. Ba
    IEEE MLSP [paper] [code] [poster]
2018
  1. Scalable convolutional dictionary learning with constrained recurrent sparse auto-encoders
    B. Tolooshams, S. Dey, and D. Ba
    IEEE MLSP [paper] [code]
2017
  1. Robustness of frequency division technique for online myoelectric pattern recognition against contraction-level variation
    B. Tolooshams and N. Jiang
    Frontiers in Bioengineering and Biotechnology [paper]
Latest News
  • 05/2021: I joined the Applied Sciences Group at Microsoft as a Research Intern for summer 2021.
  • 05/2021: I prepared and presented a tutorial on “Model-based deep learning” at IEEE ICASSP Conference.
  • 04/2021: I gave a talk and designed workshop tutorials for Neurosur 2021.
  • 04/2021: I gave a presetation as guest lecturer on "Dictionary learning based autoencoders for inverse problems” at ES201 - Decision Theory course at Harvard University.
  • 11/2020: I gave a talk on “On the relationship between dictionary learning and sparse autoencoders” at Computational and Applied Math Seminar at Tufts University.
  • 09/2020: I am nominated for Google PhD Fellowship.
  • 03/2020: I will be giving a workshop at 2020 Harvard WECode Conference, the largest student-run undergraduate women in tech conference. (cancelled)
  • 03/2020: I presented a poster on “Image denoising and analysis of neural spiking data with recurrent autoencoders for natural exponential-family of distributions” at Women in Data Science Cambridge Conference (WiDS) 2020.
  • 01/2020: I judged at the 8th annual National Collegiate Research Conference organized by Harvard College Undergraduate Research Association.
  • 07/2019: I received AWS Machine Learning Research Awards (MLRA).
  • 04/2019: I joined Amazon AI as ML/DSP Research Intern for Summer 2019.
  • 04/2019: I received QBio Student Fellowship for 2019-2020 academic year.
  • 03/2019: I presented a poster on “Deep residual autoencoders for expectation maximization-based dictionary learning” at Women in Data Science Cambridge Conference (WiDS) 2019.
  • 10/2018: I am selected to attend the IBRO-Simons Computational Neuroscience Imbizo 2018.
  • 09/2018: I am nominated for Microsoft Ada Lovelace Fellowship.
  • 08/2018: I received QBio Student Award Competition Fellowship, offering $41,200 for 2018-19 academic year.
  • 07/2018: I received a travel grant of $500 for MLSP2018.
  • 09/2017: I started my graduate studies at CRISP Group at Harvard University.
  • 06/2017: I received my B.ASc. degree with distinction in Electrical Engineering from the University of Waterloo.