My name is Martin Trapp, I’m a Ph.D. candidate in probabilistic machine learning. I have a particular interested in deep tractable probabilistic models, probabilistic programming and Bayesian nonparametrics.
My thesis is supervised by Prof. Franz Pernkopf from the Graz University of Technology and Asst. Prof. Robert Peharz from the Eindhoven University of Technology. Until 2019 I was also affiliated with the Austrian Research Institute for AI and worked under the supervision of Dr. Brigitte Krenn and Prof. em. Robert Trappl. Before I started my Ph.D. I worked as a researcher at the Research Institute for Virtual Reality and Visualisation under the supervision of Dr. Katja Bühler.
In my spare time, I’m helping in the development of the probabilistic programming language Turing.jl, which is an open-source project organised by Dr. Hong Ge.
To contact me, find me on Twitter or Github.
News & Selected Publications
You can find an extended list publications on Google scholar.
- [June 2020] Robert’s paper on Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuits has been accepted at ICML.
- [January 2020] My paper on Deep Structured Mixtures of Gaussian Processes has been accepted at AISTATS.
- [December 2019] Together with Antonio Vergari, I organised the first official NeurIPS social event on tractable probabilistic inference (T-PRIME).
- [September 2019] My paper on Bayesian Learning of Sum-Product Networks has been accepted at NeurIPS (acceptance rate 21.2%).
- [May 2019] Robert’s paper on Random Sum-Product Networks: A Simple and Effective Approach to Probabilistic Deep Learning has been accepted at UAI (acceptance rate 26.2%).
- [June 2017] My paper on Safe Semi-Supervised Learning of Sum-Product Networks has been accepted at UAI (acceptance rate 31.0%).
