○ CLICK FOR LIVE TRAINING
Classification · Binary Cross-Entropy Loss
Step 000   Loss 0.000
○ CLICK FOR LIVE TRAINING
Regression · Mean Squared Error Loss
Step 000   Loss 0.000
○ CLICK FOR LIVE TRAINING
Digit Recognizer · Cross-Entropy Loss
Step 000   Loss 0.000
○ CLICK FOR LIVE TRAINING
CartPole-v1 · ARS
Episode 000   Return 000   Gen 1
○ CLICK FOR LIVE TRAINING
Pendulum-v1 · ARS
Episode 000   Return 000   Gen 1
○ CLICK FOR LIVE TRAINING
BipedalWalker · ARS
Episode 000   Return 000   Gen 1

Lukas Müller

Machine Learning & Robot Learning
ContactDownload Resume
01 About
More →
Lukas Müller

I am studying computer science, currently writing my bachelor thesis on robot learning. I want to get machines that learn to act, right now through imitation learning.

Field
Machine Learning & Robot Learning
Focus
Imitation Learning
Location
Frankfurt, Germany
Email
contact@lukasmueller.dev
02 Resume
Education, Experience & Skills
View Resume →
03 Projects
See all →
Jan–Jun 2026Aarhus UniversityOnline vs. Optimal Electrical Vehicle ChargingHow close can a simple, computation-free charging rule get to the cost-optimal policy for an EV facing uncertain trips and fluctuating electricity prices?Markov Decision ProcessesDynamic ProgrammingEV ChargingCodeOct 2025Open SourceMuJoPyA small Python package that gives pythonic, graph-structured access to MuJoCo models — the preprocessing layer from the robot-generation project, factored out into a standalone, pip-installable tool.MuJoCoPythonRobotics ToolingCodePyPIApr–Sep 2025Technical University of DarmstadtLarge Scale Robot GenerationCan a variational autoencoder over robot graphs learn a continuous latent space of morphologies, so that thousands of diverse, physically plausible embodiments can be sampled directly from existing robot designs?Graph Neural NetworksVariational AutoencoderRobot MorphologyMuJoCoCodePaper
04 Blog
See all →

Writing Coming Soon