Explore a robotics seminar on knowledge representation and reasoning for cognition-enabled robot manipulation. Delve into advanced topics including task planning, manipulation skills, cognitive architecture, and symbolic knowledge bases. Learn about failure taxonomy, inference, grounding, and scaling in robot control programs. Discover how robots navigate real-world environments, utilize mental simulation, and perform autonomous manipulation tasks. Examine the role of semantic representations, episodic memory, and search-based reasoning in robotic systems. Gain insights into the challenges and key problems facing cognition-enabled robot manipulation, as presented by Michael Beetz from the University of Bremen's Institute for Artificial Intelligence.
Overview
Syllabus
Welcome
Introduction
Manipulation
Robot Control Programs
Manipulation Skills
The Key Problem
The Challenge
Cognitive Architecture
Implicit
Low Level
Plan Language
Failure Taxonomy
Inference
Hypothesis
Demonstration
Grounding
Scaling
Supervised Learning
Knowledge Representation
Robots in the Supermarket
Building a Symbolic Knowledge Base
Autonomous Manipulation
RealWorld Environment
Theatre Twin
Simulation Based Reasoning
Symbolic Knowledgebase
Simulation
Mental Simulation
Mental Imagery
Manipulation Tasks
Semantic Representations
Manipulation Actions
Basic System Architecture
Hybrid Infrastructure
Episodic Memory
Data Structure
Searchbased Reasoning
Taught by
Paul G. Allen School
