Yaniv Yacoby

prof_pic.png

About

I am a PhD candidate in Machine Learning at Harvard University, working with Professor Finale Doshi-Velez at the Data to Actionable Knowledge Lab (DtAK). I work on uncertainty quantification and tractable approximate inference for deep Bayesian latent variable models with applications in health-care.

Before joining DtAK, I got a MM in Contemporary Improvisation from the New England Conservatory in 2016 and a AB in Computer Science from Harvard College in 2015. You can find my music page here.

Selected Publications

  1. Pre-PrintICML UDL
    Mitigating the Effects of Non-Identifiability on Inference for Bayesian Neural Networks with Latent Variables
    *Yacoby, Yaniv*Pan, Weiwei, and Doshi-Velez, Finale
    Pre-Print 2022. Previously accepted at ICML UDL 2019.
  2. Pre-PrintICML UDL
    Failure Modes of Variational Autoencoders and Their Effects on Downstream Tasks
    Yacoby, YanivPan, Weiwei, and Doshi-Velez, Finale
    Pre-Print 2022. Previously accepted at ICML UDL 2022.
  3. AABIPMLR
    Characterizing and Avoiding Problematic Global Optima of Variational Autoencoders
    Yacoby, YanivPan, Weiwei, and Doshi-Velez, Finale
    In Symposium on Advances in Approximate Bayesian Inference 2019.