About

I am a PhD student 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.

Publications

You can find the most updated list on my google scholar page.

• Y Yacoby, W Pan, F Doshi-Velez. Failure Modes of Variational Autoencoders and Their Effects on Downstream Tasks. Pre-Print.

  • Previous Version: Y Yacoby, W Pan, F Doshi-Velez. Failures of Variational Autoencoders and Their Effects on Downstream Tasks. ICML Workshop on Uncertainty & Robustness in Deep Learning (UDL), 2020.

• S Thakur, C Lorsung, Y Yacoby, F Doshi-Velez, W Pan. Uncertainty-Aware (UNA) Bases for Bayesian Regression Using Multi-Headed Auxiliary Networks. Pre-Print.

  • Previous Version: S Thakur, C Lorsung, Y Yacoby, F Doshi-Velez, W Pan. Learned Uncertainty-Aware (LUNA) Bases for Bayesian Regression using Multi-Headed Auxiliary Networks. ICML Workshop on Uncertainty & Robustness in Deep Learning (UDL), 2020.

• T Guénais, D Vamvourellis, Y Yacoby, F Doshi-Velez, W Pan. BaCOUn: Bayesian Classifiers with Out-of-Distribution Uncertainty. ICML Workshop on Uncertainty & Robustness in Deep Learning (UDL), 2020.

• M Downs, J Chu, Y Yacoby, F Doshi-Velez, W Pan. CRUDS: Counterfactual Recourse Using Disentangled Subspaces. ICML Workshop on Human Interpretability in Machine Learning (WHI), 2020.

• Y Yacoby, W Pan, F Doshi-Velez. Characterizing and Avoiding Problematic Global Optima of Variational Autoencoders. Advances in Approximate Bayesian Inference (AABI), 2019, Proceedings of Machine Learning Research (PMLR) 118:1-17, 2020. Spotlight Talk.

• Y Yacoby, W Pan, F Doshi-Velez. Mitigating the Effects of Non-Identifiability on Inference for Bayesian Neural Networks with Latent Variables. Pre-Print.

  • Previous Version: Y Yacoby, W Pan, F Doshi-Velez. Mitigating Model Non-Identifiability in BNN with Latent Variables. ICML Workshop on Uncertainty & Robustness in Deep Learning (UDL), 2019. Spotlight Talk.

• D Vaughan, W Pan, Y Yacoby, EA Seidler, AQ Leung, F Doshi-Velez, D Sakkas. The application of machine learning methods to evaluate predictors of live birth in programmed thaw cycles, Clinical Abstract. American Society of Reproductive Medicine (ASRM), 2019.

Teaching & Mentoring

I created and organized a workshop for incoming PhD students at Harvard about managing the multi-faceted challenges of being a PhD student (Fall 2020).  Workshop content can be found here.

I was a final project mentor, as well as a research mentor for students continuing their research after completing “Stochastic Methods for Data Analysis – Inference and Optimization” (AM207) (Fall 2019 - present).

I am a mentor for the Women in Data Science (WiDS) Cambridge datathon workshop, 2019-2020.

I previously served as a teaching fellow for a number of courses at Harvard: Advanced Machine Learning (CS281) Fall 2018, Systems Programming and Machine Organization (CS61) Fall 2015, and Intro to Computer Science I (CS50) Fall 2012.

Contact

Email: yanivyacoby [AT] g [DOT] harvard [DOT] edu