-
Apr 2026
MOMO: Seamless Physical, Verbal, and Graphical Robot Skill Adaptation
MOMO: A framework unifying physical, verbal, and graphical modalities for interactive robot skill learning and adaptation. Validated on a 7-DoF torque-controlled robot at Automatica 2025.
Machine-LearningRoboticsAI
-
Apr 2026
CLASP: Natural Language-Driven Robot Skill Selection and Composition
CLASP: A modular framework combining Vision-Language Models with Task-Parameterized KMPs for data-efficient, language-driven robot skill selection, composition, and active acquisition.
Machine-LearningRoboticsAI
-
Mar 2026
RA-L 2026: IROSA — Interactive Robot Skill Adaptation using Natural Language
IROSA: A tool-based framework enabling open-vocabulary robot skill adaptation through natural language using LLMs and Kernelized Movement Primitives. Published in IEEE RA-L 2026.
Machine-LearningRoboticsAI
-
Aug 2025
Summer Schools 2025: ELLIS Cambridge & OxML Oxford
My experience at two world-class summer schools: ELLIS Cambridge (Probabilistic ML) and OxML Oxford (Representation Learning & Generative AI). Key insights and program highlights.
Machine-LearningEducationConferences
-
Jun 2025
Automatica 2025: DLR Showcasing the Future of Robotics
DLR's robotics showcase at Automatica 2025 Munich: TORO humanoid, AIMM autonomous system, EDAN assistive robot, SARA with LLM+KMP integration, and HUG haptic interface.
RoboticsConferencesDLR
-
Apr 2025
ICRA 2025: RACCOON — Grounding Embodied Question-Answering in Robotics
Our RAG-based approach for truthful foundation model question-answering in robotics was accepted by IEEE ICRA.
AIRobotics
-
Feb 2025
RA-L 2025: LOCI — Interactive Incremental Learning of Generalizable Robot Skills
LOCI: A probabilistic machine learning method for interactive robot skill modulation using a task-parameterized, kernelized approach (TP-KMP). Published in IEEE RA-L 2025.
Machine-LearningRoboticsInteractive-LearningIncremental-LearningAI
-
Nov 2024
CoRL 2024: Spotlight Presentation at CoRoboLearn Workshop
Spotlight presentation and two papers at the CoRoboLearn Workshop (Advancing Learning for Human-Centered Collaborative Robots) at CoRL 2024 in Munich.
RoboticsConferences
-
Jul 2024
RSS 2024: Embodied Question-Answering at GenAI-HRI Workshop
Presenting our RAG-based approach for grounded robot question-answering at the GenAI-HRI Workshop at RSS 2024 in Delft, Netherlands.
AIRoboticsConferences
-
Apr 2024
Automatica 2023 & Hannover Fair 2024: Task-Parameterized Robot Demo
Live demonstration of task-parameterized robotic approach at Automatica 2023 and Hannover Fair 2024, showcasing interactive robot learning capabilities.
Robotics
-
Oct 2023
Project: BlenderProc — Photorealistic Rendering for Computer Vision (3.4k+ GitHub Stars)
Popular open source project for fast creation of training images for deep neural networks in computer vision.
Computer-VisionOpen SourceDeep LearningBlender
-
Oct 2022
Dataset: HOWS-CL-25 — A Continual Learning Dataset for Household Objects
HOWS-CL-25: A comprehensive synthetic dataset with 150,795 unique images across 25 household object categories for continual learning research, created using BlenderProc.
Computer-VisionOpen SourceDeep Learning
-
Oct 2022
IROS 2022: RECALL — Rehearsal-Free Continual Learning for Object Classification
RECALL: A novel rehearsal-free continual learning approach for object classification achieving SOTA results. Published at IEEE IROS 2022.
AIMachine-LearningComputer-Vision