Discussion: Curiosity Rewards
Maximilian Granz
Author: Tim Landgraf
Paper Discussion: Why Normalizing Flows Fail to Detect Out-of-Distribution Data.
Oana Juiliana Popescu – https://arxiv.org/pdf/2006.08545.pdf
Paper Discussion: World Models
Maximilian Granz – https://openreview.net/forum?id=S1lOTC4tDS
Paper Discussion: The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
Tarek Beutler – https://arxiv.org/abs/1803.03635
Paper Discussion: When Explanations Lie: Why Modified BP Attribution Fails
Leon Sixt – https://arxiv.org/abs/1912.09818
Paper Discussion: DeepProbLog: Neural Probabilistic Logic Programming
Chiara Lakomski – http://papers.nips.cc/paper/7632-deepproblog-neural-probabilistic-logic-programming.pdf
Paper Discoussion: Improving Generalization in Meta Reinforcement Learning Using Learned Objectives
Maximilian Granz – https://arxiv.org/abs/1910.04098
TV report: Tim gives talk on RoboFish and BeeCharge
The TV network “Bayerischer Rundfunk” records short science talks in Berlin’s famous “Roter Salon” and broadcasts them on ARD alpha. Tim told the audience about the RoboFish project (how we can make robot fish “learn” in real-time from interactions with… Continue Reading
Paper Discussion: Learning Robust Representations via Multi-View Information Bottleneck
Leon Sixt – https://openreview.net/forum?id=B1xwcyHFDr
Paper Discussion: Creative GANs
Chiara Lakomski – https://arxiv.org/abs/1706.07068