Completed
Training Data
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Natural Language Processing - Yoav Artzi - Cornell University
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Mapping Instructions to Actions
- 3 Mapping Instructions to Control • Drone state is determined by a configuration of target velocities
- 4 Modular Approach
- 5 Single-model Approach
- 6 Visitation Distribution for Navigation
- 7 Simulation-Reality Joint Learning
- 8 Training Data
- 9 Learning Architecture Supervised Learning Learning
- 10 RL for Control
- 11 Reward Goals
- 12 SuReAL Supervised and Reinforcement Asynchronous Learning
- 13 Experimental Setup
- 14 Human Evaluation
- 15 Game Environment
- 16 Two Players
- 17 Game Objective
- 18 Action and Dynamic
- 19 The CerealBar Scenario
- 20 The Task Goal: learn to map leader instructions to follower actions
- 21 Instruction Following in Collaborations
- 22 Learning with Error Propagation
- 23 Generating Error Recovery Examples
- 24 Generated Recovery Example
- 25 Data Collection
- 26 Data Statistics
- 27 Data Under-utilization
- 28 Cascaded Evaluation Metrics
- 29 Results: Held-out Games
