This advanced seminar introduces fundamental algorithms for Robotic Embodied-AI systems that can autonomously perceive, navigate and manipulate objects in unstructured environments like homes, restaurants, supermarkets, etc. It addresses the complexities of developing intelligent robotic agents that can interact and change their world. The seminar will discuss fundamental problems in embodied AI connecting Multimodal Perception to Action.
The seminar will combine an introductory lecture and a reading group to discuss and learn about advanced algorithmic approaches in robotics and embodied AI. After the introductory lecture, students will be assigned research papers (1-2 per person groups through the semester based on the class size). The students shall present the paper's basic concept in class and discuss the topics presented. Moreover, we will implement a structured reviewing process for the assigned papers. Each presenter must write a comprehensive review of the paper(s) they were assigned -- criticizing and questioning are key for developing critical thinking and a research-oriented mindset. Two other participants will act as meta-reviewers (once or twice during the semester) who will have to perform fact-checking of the review wrt the paper and evaluate the quality of the review. Each student's final grade will be calculated by their presentation, review, meta-review, and overall participation in the seminar.
List of topics (non-exhaustive):
- 2D and 3D vision
- Implicit Geometric Representations
- Multi-sensory robot perception
- Transformers and Language models
- Large Language & Vision Models for Robotics
- Imitation & Reinforcement Learning
- Pre-training for Robot Learning
Guest speakers:
The seminar will also include invited research talks by distinguished researchers in the field of Robot Learning. More details will be provided soon!
10/05: Katerina Fragkiadaki
14/6: Shuran Song
21/6: Danny Driess
Schedule:
Wednesdays from 13:30 to 15:10 in room: S202/C110
Start: Apr 12, 2023
End: Jul 12, 2023
- Lehrende: Georgia Chalvatzaki