Completed
Why Traditional Self-Training Does Not Work?
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Towards Open World Event Knowledge Extraction with Weak Supervision
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 What is event extraction?
- 3 Application - Intelligence Analysis
- 4 Traditional Event Extraction
- 5 How to solve the limitations?
- 6 Our First Solution: Query-and-Extract
- 7 Query and Extract: Type-oriented Binary Decoding
- 8 Approach Details - Event Trigger Detection
- 9 Approach Details - Event Argument Extraction
- 10 Experiments - Supervised Event Extraction
- 11 Experiments-Zero-Shot Event Extraction
- 12 Event Detection based on Type Specific Prompts
- 13 Few-Shot Event Detection
- 14 Can the model be continuously updated?
- 15 Episodic Memory Prompting
- 16 How well the model retrains the capability on old types?
- 17 Pros and Cons of Query-and-Extraction Paradigm
- 18 Our Second Solution: Self-Training
- 19 Why Traditional Self-Training Does Not Work?
- 20 Self-Supervised Event Extraction with Gradient Guidance
- 21 Self-Training with Gradient Guidance
- 22 Evaluation of the Scoring Model
- 23 How Effective is the STGG to Event Extraction?
- 24 Future Directions: Open-Environment Event Extraction
