Skip to content
View phierhager's full-sized avatar

Block or report phierhager

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
phierhager/README.md

Philip Hierhager

Machine Learning Researcher | Geometric DL & RL

🔬 Research: Self-Supervised Learning, Graph & Geometric Neural Networks | ⚡ Skills: JAX, PyTorch, TensorFlow, Julia, Python

LinkedIn GitHub


🛠️ Selected Skills

python pytorch tensorflow JAX Julia r

docker kubernetes git linux

🔭 Currently working on: Home Energy Simulation (JAX, MPC, RL) | Action-adaptive Continual Learning for RL Agents

🚀 Recent research: 3D Object Detection (Self-Supervised) | Molecule Modeling (GNNs) | Non-Convex Consensus Optimization


📄 Selected Projects & Research

Project / Paper Description Link
Home Energy Simulation High-performance high-fidelity simulation in JAX with differentiable MPC, continuous-time deep RL, neural ODEs and system identification. Running 200 million steps per second across parallel environments on a laptop GPU. GitHub
Action-adaptive Continual Learning Continual learning framework for reinforcement learning agents, adaptive to changing tasks. Coming Soon
3D Object Detection (Self-Supervised) Efficient self-supervised methods for autonomous driving 3D object detection. PDF
Molecule Modeling with GNNs Graph neural networks for molecular property prediction. GitHub
Consensus-Based Optimization Experimental framework and methods for non-convex consensus optimization. GitHub
Robust Quantum Variational Algorithms Influence of Noise on Quantum Variational Algorithms. PDF
Quantum-Classical MILP Optimization Evaluation of quantum-classical hybrid solution methods for 3SAT problems. (German) PDF

@phierhager's activity is private