Keynote: Learning from observation: Small-data approach to human common sense
Speaker: Katsushi Ikeuchi, Sr. Principal Research Manager, Microsoft Research Redmond
Learning-from-Observation (LfO), a robot-teaching paradigm, aims to build a robot system that understands what humans do through a small number of human observations and map them to action. Unlike the more popular Learning-from-Demonstration paradigm, in which the robot itself is directly manipulated and the demonstration is repeated many times, Learning-from-Observation uses accumulated robotics knowledge from a small number of demonstrations through explicit steps and formulates it into task models to achieve a goal. In this session, we’ll briefly introduce the basic design concepts of specific task models, shared and formulated through earlier system designs. We’ll then discuss the formulation of common sense, required to achieve household tasks such as wiping a tabletop or bringing a cup of tea without spilling, into explicit task models. We’ll also discuss task recognition through these task models from visual and verbal cues and the implementation of robust execution modules pre-trained by reinforcement learning.
Learn more about the 2021 Microsoft Research Summit: https://Aka.ms/researchsummit (opens in new tab)
- Track:
- Towards Human-Like Visual Learning & Reasoning
- Date:
- Speakers:
- Katsushi Ikeuchi
- Affiliation:
- Microsoft Research
-
-
Katsushi Ikeuchi
Sr. Principal Research Manager
-
-
Towards Human-Like Visual Learning & Reasoning
-
Opening remarks: Towards Human-Like Visual Learning and Reasoning
Speakers:- Wenjun Zeng,
- Wenjun Zeng
-
-
-
Research talks: Learning for interpretability
Speakers:- Yuwang Wang,
- Hanwang Zhang,
- Shujian Yu
-
Research talks: Few-shot and zero-shot visual learning and reasoning
Speakers:- Han Hu,
- Zhe Gan,
- Kyoung Mu Lee
-
-
Research talks: Generalization and adaptation
Speakers:- Suha Kwak,
- Chong Luo,
- Lu Yuan
-
