Thesis Defense: Towards a Deeper Understanding of Neural Language Generation
, CSAIL MIT
Date: Wednesday, April 27, 2022
Time: 2:00 PM to 3:00 PM Note: all times are in the Eastern Time Zone
Location: 32-G882, also virtual at https://mit.zoom.us/j/98047395332
Event Type: Thesis Defense
Host: James Glass, CSAIL MIT
Contact: Tianxing He, firstname.lastname@example.org
Speaker URL: https://people.csail.mit.edu/cloudygoose/
TALK: Thesis Defense: Towards a Deeper Understanding of Neural Language Generation
In recent years, the field of language modelling has witnessed exciting developments. Researchers are able to train language models which can generate realistic text. However, our understanding of these powerful language models remains shallow. This talk highlights our efforts towards a deeper understanding of neural language generation. In the first part, we will discuss "how to do generation". We will review several popular sampling algorithms, and show their performance are on par with each other. Then, we will extract high-level shared properties from the definition of the algorithms, and design experiments to validate their importance. In the second part, we will discuss "what to generate". We will review some known problems arised from neural language generation, and propose the negative training algorithm to deal with those bad behaviors.
AI & Machine Learning
Created by Tianxing He at Friday, April 22, 2022 at 7:23 PM.