Robert Müller
After graduating in Mathematics at the Technical Univesity of Munich, Robert joined the ML department in Carnegie Mellon to work on generalization in reinforcement learning with information bottlenecks, then joined appliedAI to work on multimodal meta learning and batch reinforcement learning until January 2021. After some time at Sony AI, he is back with us working on self-play in RL and applications in NLP.

Robert's work

Pill

Hyperparameters in Reinforcement Learning and How To Tune Them

Finding good hyperparameters for reinforcement learning (RL) is a notoriously difficult task. Instead of luck or grid search, one can apply …

Robert Müller
Reinforcement Learning
Pill

Exploiting past success in Off-Policy Actor-Critic

Extracting knowledge from previously gathered data is the very core of exploitation in reinforcement learning (RL). During training, a …

Robert Müller
Reinforcement Learning
Pill

Implicit Q Learning

A core challenge when applying dynamic programming based approaches to offline reinforcement learning is the bootstrapping error. This pill …

Robert Müller
Reinforcement Learning
Pill

Reasoning Traces as Learning Signal

An important feature of large language models is their ability to provide detailed responses that resemble “thinking step by …

Robert Müller
Large Language Models
Seminar

Duality in RL and the [*]-DICE Family (part 2)

The goal of reinforcement learning is to find an optimal policy. Two commonly used classes of methods are i) policy gradient and ii) dynamic …

Robert Müller
Seminar

Duality in RL and the [*]-DICE Family (part 1)

The goal of reinforcement learning is to find an optimal policy. Two commonly used classes of methods are i) policy gradient and ii) dynamic …

Robert Müller
Seminar

Problems and solutions in offline learning

Current reinforcement learning algorithms require tons of data to solve new tasks. This “hunger for data” is especially onerous in high …

Robert Müller
Reinforcement Learning
Seminar

The Information Bottleneck in RL

State representations hold the promise of simplifying control, allowing a reinforcement learning (RL) agent to solve a task more quickly, …

Robert Müller