Methods to improve quality and diversity of language-vision models

Aneja, Jyoti

Permalink

https://hdl.handle.net/2142/114006 Copy

Description

Title

Methods to improve quality and diversity of language-vision models

Author(s)

Aneja, Jyoti

Issue Date

2021-12-03

Director of Research (if dissertation) or Advisor (if thesis)

Schwing, Alexander

Doctoral Committee Chair(s)

Hooberman, Benjamin

Committee Member(s)

• Lazebnik, Svetlana
• Cooper, Lance
• Forsyth, David

Department of Study

Physics

Discipline

Physics

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

Machine Learning, Computer Vision, Natural Language Processing

Abstract

Humans can describe images and, more generally, the world around them in an evocative manner using vivid language constructs. Designing neural network models that can attain results similar to those of humans on tasks like image-captioning and image-generation is a worthy goal in the overall pursuit of artificial general intelligence. Notwithstanding the tremendous recent progress in this area, current systems still cannot describe objects and scenes as creatively and accurately as humans. As a step in the direction of bridging this gap, this thesis proposes architectures and algorithms for generating high-quality, diverse outputs for the tasks of image-captioning and image-generation.

Graduation Semester

2021-12

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/114006

Copyright and License Information

Copyright 2021 Jyoti Aneja

Owning Collections