Select a calendar:
Filter November Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for November 07, 2023
-
PhD Thesis Proposal - Mozhdeh Gheini
Tue, Nov 07, 2023 @ 09:00 AM - 11:00 AM
Thomas Lord Department of Computer Science
University Calendar
PhD Thesis Proposal - Mozhdeh Gheini
Committee Members: Jonathan May (Chair), Xiang Ren, Xuezhe Ma, Swabha Swayamdipta, Khalil Iskarous
Title: Inductive Biases for Data- and Parameter-Efficient Transfer Learning
Abstract: The widespread success of natural language processing (NLP) models, such as Large Language Models, and the subsequent attention from the public often conceal and distract from the sheer amount of data and computational resources they have relied on to reach this point. The very same models often fail to perform as well in the absence of sufficient data and computational resources. However, how to adjust methods under such constraints remains under-discussed. In this talk, I present work incorporating inductive biases during both pretraining and downstream transfer learning and showcase the boosted performance for machine translation and named entity recognition under resource limitations. Following that, I discuss our work on creating a pretrained model using MEGA, a novel architecture with extensions to Transformers, and our ongoing efforts to investigate MEGA's inductive biases that significantly set it apart from Transformer in low-resource scenariosLocation: Ronald Tutor Hall of Engineering (RTH) - 306
Audiences: Everyone Is Invited
Contact: Melissa Ochoa
Event Link: https://usc.zoom.us/j/6564802162
-
PhD Thesis Proposal - Paul Chiou
Tue, Nov 07, 2023 @ 01:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
University Calendar
PhD Thesis Proposal - Paul Chiou
Committee Members: William G.J. Halfond (chair), Nenad Medvidovic, Mukund Raghothaman, Gisele Ragusa, and Chao Wang
Title: Automated Detection of Keyboard Accessibility Issues in Web Applications
Abstract: The internet has become an important part of our daily lives, enabling us to complete everyday and essential tasks online. For the 15% of the global population with disabilities, accessing the internet is critical and can provide access to resources that would otherwise be unavailable. Many people with different disabilities rely on the keyboard interface to access the internet; however, studies found that web applications today largely remain inaccessible to keyboard users. Testing keyboard accessibility is a labor-intensive task currently done manually by skilled practitioners. In this thesis proposal, I propose to use program analysis techniques to automate the keyboard accessibility testing process to alleviate the manual effort involved. I developed a novel approach to automatically detect keyboard accessibility issues that negatively affect disabled users' ability to navigate web pages' user interface. The approach implements a dynamic crawler to build a model that captures a web page's interactivity from a keyboard person's perspective. The approach then analyzes the model to identify the inaccessible behaviors per accessibility guidelines. Finally, I propose to conduct an evaluation to show the approach’s ability to accurately detect these keyboard accessibility issues in real-world web applicationsLocation: Social Sciences Building (SOS) - B43
Audiences: Everyone Is Invited
Contact: Melissa Ochoa
