Thesis Defense: Controlling Neural Language Generation

Speaker: Tianxiao Shen , MIT CSAIL

Date: Tuesday, May 03, 2022

Time: 1:30 PM to 3:00 PM Note: all times are in the Eastern Time Zone

Public: Yes

Location: 32-D463 (Star), https://mit.zoom.us/j/97034696195

Event Type: Thesis Defense

Room Description:

Host: Regina Barzilay, MIT CSAIL

Contact: Tianxiao Shen, tianxiao@csail.mit.edu

Relevant URL:

Speaker URL: http://people.csail.mit.edu/tianxiao/

Speaker Photo:
None

Reminders to: seminars@csail.mit.edu

Reminder Subject: TALK: Thesis Defense: Controlling Neural Language Generation

Thesis advisors: Tommi Jaakkola, Regina Barzilay
Thesis committee: Yoon Kim

Abstract:
Language models have achieved success in a wide range of applications such as machine translation, dialogue generation, and writing autocompletion. However, typical models operate in a left-to-right, unconstrained fashion with limited control over what is generated. In this talk, I’ll discuss three directions to control language generation. First, we’ll leverage distributional alignment to perform style transfer from non-parallel text. Next, we’ll study confounding factors and use invariance to isolate them to transfer text in the desired direction. Lastly, we’ll look at flexible generation location control, for which we developed sequence models that dynamically create and fill in blanks during the generation process.

Research Areas:
AI & Machine Learning

Impact Areas:

This event is not part of a series.

Created by Tianxiao Shen at Wednesday, April 27, 2022 at 2:21 PM.