Jan Peters

Proud to share our latest work, accepted at @iclr-conf.bsky.social 2026: APPLE! 🍎

TL;DR: APPLE is a novel reinforcement learning framework for solving active perception problems.

#ICLR2026 #Robotics #MachineLearning #ActivePerception #RL
@ias-tudarmstadt.bsky.social

🧵 Accepted at @iclr-conf.bsky.social!

Target networks stabilize bootstrapping in RL 🛡️
But induce slow-moving targets 🐢

Online networks adapt fast ⚡
But can diverge with function approximation 💥

𝗠𝗜𝗡𝗧𝗢 🌿 uses the online network 𝗼𝗻𝗹𝘆 𝗶𝗳 𝗶𝘁 𝗰𝗮𝗻 — yielding faster 𝘢𝘯𝘥 more stable RL.

Here’s how 👇

Should we use a target network in deep value-based RL?🤔

The answer has always been YES or NO, as there are pros and cons.

@iclr-conf.bsky.social, I will present iS-QN, a method that lies in between this binary view, collecting the pros while reducing the cons🚀

🥳Our paper "Floating-Base Deep Lagrangian Networks (FeLaN)" has been accepted to #ICRA2026.

FeLaN: a grey-box approach for physically consistent SysID of floating-base robots (humanoids, quadrupeds).

📄 arxiv.org/abs/2510.17270
💻 Soon!
🌐 schulze18.github.io/felan_website/

NVIDIA Awards up to $60,000 Research Fellowships to PhD Students The Graduate Fellowship Program announced the latest awards of up to $60,000 each to 10 Ph.D. students involved in research that spans all areas of computing innovation.

I'm super excited to have been named an #NVIDIA Graduate Fellowship Finalist! 💚

Huge thanks to my supervisor @jan-peters.bsky.social and all my collaborators.

Can't wait to join the NVIDIA Seattle Robotics Lab for my internship next summer! 🤖

blogs.nvidia.com/blog/graduat...

🎉 Really excited, our paper "XQC: Well-conditioned Optimization Accelerates Deep Reinforcement Learning" has been accepted at #ICLR2026.

If you are interested in reinforcement learning, sample-efficiency, compute-efficiency go check it out. See you in Rio!

Just graduated PhD student #45 ... Niklas Wilhelm Funk did an outstanding job defending a doctoral dissertation on Learning Robotic Manipulation through Vision, Touch, and Spatially Grounded Representations! Major insights on many different aspects of manipulation...

Just graduated PhD student #45 ... Niklas Wilhelm Funk did an outstanding job defending a doctoral dissertation on Learning Robotic Manipulation through Vision, Touch, and Spatially Grounded Representations! Major insights on many different aspects of manipulation...

@timschneider94.bsky.social will present "Analysing the Interplay of Vision and Touch for Dexterous Insertion Tasks" by Janis Lenz, Tim Schneider, Theo Gruner, @daniel-palenicek.bsky.social, and
@jan-peters.bsky.social

🗓️ 13.06, 16:30 - 19:30
📍 Poster 100
See bsky.app/profile/tims...

Our work introduces a geometrically-aware approach that brings motion planning into Bayesian goal inference—an early but promising direction.

With @anindex.bsky.social , Theo Gruner, @joemwatson.bsky.social , @georgiachal.bsky.social & @jan-peters.bsky.social

@kay-pompetzki.bsky.social will present "Geometrically-Aware Goal Inference: Leveraging Motion Planning as Inference" by Kay Pompetzki, @anindex.bsky.social, Theo Gruner, @georgiachal.bsky.social, and @jan-peters.bsky.social

🗓️ 13.06, 16:30 - 19:30
📍 Poster 86
See bsky.app/profile/kay-...

Massively Scaling Explicit Policy-conditioned Value Functions We introduce a scaling strategy for Explicit Policy-Conditioned Value Functions (EPVFs) that significantly improves performance on challenging continuous-control tasks. EPVFs learn a value function V(...

🦾 By combining EPVFs with massive parallelism and careful regularization, we close the gap with state-of-the-art DRL in complex environments.
🔗 Full paper: arxiv.org/abs/2502.11949
✨ Finally, many thanks to @jan-peters.bsky.social and
@ias-tudarmstadt.bsky.social for the support!

We will also present "Scaling Off-Policy Reinforcement Learning with Batch and Weight Normalization" by @daniel-palenicek.bsky.social, Florian Vogt, @joemwatson.bsky.social, and @jan-peters.bsky.social.

🗓️ 13.06, 16:30 - 19:30
📍 Poster 50
See bsky.app/profile/did:...

Or come to my talk @ International Symposium on Adaptive Motion of Animals and Machines and LokoAssist Symposium (AMAM) on Friday at TU Darmstadt

Thanks to @ias-tudarmstadt.bsky.social, @jan-peters.bsky.social

🎤 Very excited to give a talk @cohereforai.bsky.social next week Friday 🎤

I will be presenting the research I have been working on for the last 2 years with Carlo D'Eramo, @jan-peters.bsky.social, and many more collaborators!

🎉Our paper "Context-Aware Deep Lagrangian Networks for MPC" was accepted at #IROS2025!

We present CaDeLaC: adaptive physics-consistent robot control via online SysID + MPC.

Paper: arxiv.org/abs/2506.15249

Big thanks to @ias-tudarmstadt.bsky.social @jan-peters.bsky.social @olegarenz.bsky.social

GitHub - TimSchneider42/franky: High-Level Control Library for Franka Robots with Python and C++ Support High-Level Control Library for Franka Robots with Python and C++ Support - GitHub - TimSchneider42/franky: High-Level Control Library for Franka Robots with Python and C++ Support

Pushing for #icra but still missing real robot experiments? 😰
Skip the ROS headaches — get your Franka robot running in minutes with franky! 🦾
Super beginner-friendly, Pythonic, and fast to set up.
🔗 github.com/TimSchneider...
@ias-tudarmstadt.bsky.social @jan-peters.bsky.social

🚀 New preprint! Introducing XQC— a simple, well-conditioned actor-critic that achieves SOTA sample efficiency in #RL
✅ ~4.5× fewer parameters than SimbaV2
✅ Scales to vision-based RL
👉 arxiv.org/pdf/2509.25174

Thanks to Florian Vogt @joemwatson.bsky.social @jan-peters.bsky.social

GitHub - Schulze18/cadelac: Context-Aware Deep Lagrangian Networks for Model Predictive Control (CaDeLaC) Context-Aware Deep Lagrangian Networks for Model Predictive Control (CaDeLaC) - Schulze18/cadelac

Warming up for #IROS2025 🔥

We’re releasing CaDeLaC as open source!
- Training pipeline
- Integration with acados for adaptive physics-consistent MPC
- Simulation + real Franka torque control

💻 github.com/Schulze18/ca...

@ias-tudarmstadt.bsky.social @jan-peters.bsky.social @olegarenz.bsky.social

🇰🇷 Conferences are about finally meeting your collaborators from all around the world!

Check out our work on Embodiment Scaling Laws @CoRL2025
We investigate cross-embodiment learning as the next axis of scaling for truly generalist policies 📈

🔗 All details: embodiment-scaling-laws.github.io

As usual, @ewrl18.bsky.social was a wonderful experience.

I had the pleasure of presenting my research as a Contributed Talk 🎉

Special thanks to the organizers for making it happen!

Today, @theovincent.bsky.social l will present "Eau De Q-Network: Adaptive Distillation of Neural Networks in Deep Reinforcement Learning" by Théo Vincent, @jan-peters.bsky.social, and Carlo D'Eramo.
🗓️ 12.06, 16:30 - 19:30
📍 Poster 28
See bsky.app/profile/theo...

@timschneider94.bsky.social will present "Active Perception for Tactile Sensing: A Task-Agnostic Attention-Based Approach" by Tim Schneider, Cristiana de Farias, Roberto Calandra, Liming Chen, and @jan-peters.bsky.social
🗓️ 12.06, 16:30 - 19:30
📍 Poster 105
See bsky.app/profile/tims...

🚀 Interested in Neuro-Symbolic Learning and attending #ICRA2025? 🧠🤖

Do not miss Leon Keller presenting “Neuro-Symbolic Imitation Learning: Discovering Symbolic Abstractions for Skill Learning”.

Joint work of Honda Research Institute EU and @jan-peters.bsky.social (@ias-tudarmstadt.bsky.social).

Thanks to my co-authors Florian Vogt, @joemwatson.bsky.social @jan-peters.bsky.social

@hessianai.bsky.social @ias-tudarmstadt.bsky.social @dfki.bsky.social @cs-tudarmstadt.bsky.social
#RL #ML #AI

Work conducted @ias-tudarmstadt.bsky.social headed by @jan-peters.bsky.social @cs-tudarmstadt.bsky.social @tuda.bsky.social

Many thanks to all #LocoMuJoCo Contributors!

First German Robotics Conference starting in Messezentrum Nuremberg! 400+ delegates plus RoboCup next door! @jan-peters.bsky.social @tuda.bsky.social

Many thanks to my colleagues and collaborators: Daniel Palenicek, Łukasz Antczak, @jan-peters.bsky.social and most importantly Jonathan Kinzel (@ibims1jfk.bsky.social), who interned at MAB Robotics and did the experiments.
Also thanks to MAB Robotics for providing the hardware and constant support!

We build on the efficient CrossQ DRL algorithm and combine it with two control architectures — Joint Target Prediction for agile maneuvers and Central Pattern Generators for stable, natural gaits — to train locomotion policies directly on the HoneyBadger quadruped robot from MAB Robotics.

Intrigued? Check out the paper and videos here: nico-bohlinger.github.io/gait_in_eigh...