1 code implementation • 29 Apr 2024 • Xingyuan Zhang, Philip Becker-Ehmck, Patrick van der Smagt, Maximilian Karl
In this paper, we study Imitation Learning from Observation with pretrained models and find existing approaches such as BCO and AIME face knowledge barriers, specifically the Embodiment Knowledge Barrier (EKB) and the Demonstration Knowledge Barrier (DKB), greatly limiting their performance.
no code implementations • 6 Dec 2023 • Elie Aljalbout, Felix Frank, Maximilian Karl, Patrick van der Smagt
We study the choice of action space in robot manipulation learning and sim-to-real transfer.
1 code implementation • NeurIPS 2023 • Xingyuan Zhang, Philip Becker-Ehmck, Patrick van der Smagt, Maximilian Karl
Our method is "zero-shot" in the sense that it does not require further training for the world model or online interactions with the environment after given the demonstration.
no code implementations • 28 Nov 2022 • Elie Aljalbout, Maximilian Karl, Patrick van der Smagt
Multi-robot manipulation tasks involve various control entities that can be separated into dynamically independent parts.
no code implementations • ICML Workshop URL 2021 • Philip Becker-Ehmck, Maximilian Karl, Jan Peters, Patrick van der Smagt
We show that while such an agent is still novelty seeking, i. e. interested in exploring the whole state space, it focuses on exploration where its perceived influence is greater, avoiding areas of greater stochasticity or traps that limit its control.
1 code implementation • 19 Mar 2020 • Philip Becker-Ehmck, Maximilian Karl, Jan Peters, Patrick van der Smagt
Learning to control robots without requiring engineered models has been a long-term goal, promising diverse and novel applications.
Model-based Reinforcement Learning reinforcement-learning +1
no code implementations • 2 Nov 2019 • Neha Das, Maximilian Karl, Philip Becker-Ehmck, Patrick van der Smagt
Learning a model of dynamics from high-dimensional images can be a core ingredient for success in many applications across different domains, especially in sequential decision making.
no code implementations • 13 Oct 2017 • Maximilian Karl, Maximilian Soelch, Philip Becker-Ehmck, Djalel Benbouzid, Patrick van der Smagt, Justin Bayer
We introduce a methodology for efficiently computing a lower bound to empowerment, allowing it to be used as an unsupervised cost function for policy learning in real-time control.
no code implementations • 26 Sep 2016 • Christopher Wolf, Maximilian Karl, Patrick van der Smagt
Variational inference lies at the core of many state-of-the-art algorithms.
no code implementations • 23 Jun 2016 • Maximilian Karl, Justin Bayer, Patrick van der Smagt
Tactile information is important for gripping, stable grasp, and in-hand manipulation, yet the complexity of tactile data prevents widespread use of such sensors.
no code implementations • 21 Jun 2016 • Maximilian Karl, Artur Lohrer, Dhananjay Shah, Frederik Diehl, Max Fiedler, Saahil Ognawala, Justin Bayer, Patrick van der Smagt
We study the responses of two tactile sensors, the fingertip sensor from the iCub and the BioTac under different external stimuli.
4 code implementations • 20 May 2016 • Maximilian Karl, Maximilian Soelch, Justin Bayer, Patrick van der Smagt
We introduce Deep Variational Bayes Filters (DVBF), a new method for unsupervised learning and identification of latent Markovian state space models.
no code implementations • 28 Sep 2015 • Maximilian Karl, Justin Bayer, Patrick van der Smagt
This is a natural candidate for an intrinsic reward signal in the context of reinforcement learning: the agent will place itself in a situation where its action have maximum stability and maximum influence on the future.
no code implementations • 19 Jul 2015 • Justin Bayer, Maximilian Karl, Daniela Korhammer, Patrick van der Smagt
Marginalising out uncertain quantities within the internal representations or parameters of neural networks is of central importance for a wide range of learning techniques, such as empirical, variational or full Bayesian methods.
no code implementations • 29 Dec 2014 • Maximilian Karl, Christian Osendorfer
A process centric view of robust PCA (RPCA) allows its fast approximate implementation based on a special form o a deep neural network with weights shared across all layers.