Select a calendar:
Filter September Events by Event Type:
SUNMONTUEWEDTHUFRISAT
University Calendar
Events for September
-
PhD Thesis Proposal - Tu Do
Thu, Sep 02, 2021 @ 08:30 AM - 10:00 AM
Thomas Lord Department of Computer Science
University Calendar
Date and Time: Thursday, September 2nd @ 8:30-10 AM PDT
Committee: Ewa Deelman, Rafael Ferreira Da Silva, Aiichiro Nakano, Ramesh Govindan, Viktor Prasanna, Michela Taufer
Title: Enabling Efficient Execution of In Situ Workflows
Abstract:
Advances in high-performance computing (HPC) allow scientific simulations to run at an ever-increasing scale, generating a large amount of data that needs to be analyzed over time. Conventionally, the simulation outputs the entire simulated data set to the file system for later post-processing.
Unfortunately, the slow growth of I/O technologies compared to the computing capability of present-day processors causes an I/O bottleneck of post-processing as saving data to storage is not as fast as data is generated. According to data-centric models, a new processing paradigm has recently emerged, called in situ, where simulation data is analyzed on-the-fly to reduce the expensive I/O cost of saving massive data for post-processing. Since an in situ workflow usually consists of
co-located tasks running concurrently on the same resources in an iterative manner, the execution yields complicated behaviors that create challenges in evaluating the efficiency of an in situ run. To enable efficient execution of in situ workflows, this proposal proposes a theoretical framework that models the efficiency of in situ execution for evaluating the performance of in situ workflows. By extending the proposed performance model to resource-aware performance indicators, we introduce a method to assess resource usage, resource allocation, and resource provisioning for in situ workflow ensembles. In support of the thesis, this work serves as a prerequisite of evaluating in situ scheduling decisions.
Zoon info:
https://usc.zoom.us/j/99685923807
WebCast Link: https://usc.zoom.us/j/99685923807
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
PhD Thesis Proposal --Akira Matsui
Fri, Sep 03, 2021 @ 01:00 PM - 02:30 PM
Thomas Lord Department of Computer Science
University Calendar
Date and Time: 1:00pm-2:30pm
Friday, Sep, 3rd
Committee: Emilio Ferrara, Nakano Aiichiro, Aram Galstyan, Xiang Ren, Marlon Twyman
Title: Mining and leveraging temporal structures of human behavior in Online Platforms
Abstract:
With the development of technology in recent years, detailed data on human temporal behavior has become available. Many methods have been proposed for mining these human dynamic behavioral data, which have provided valuable insights for research, especially in computational social science. However, most methods analyze only sequential behavior and do not comprehensively consider inter-temporal information such as time intervals between users' actions. In this proposal, we leverage the time intervals of actions to understand human behavior from a mid-term perspective. We first argue that the engagement analysis without considering time interval differences among individuals may return a biased result and propose a method to mitigate those biases. Then we also propose an embedding-based approach to incorporate inter-temporal information into human behavior analysis, which enables us to mine the temporal context of action in a unified manner. Our proposal will demonstrate the importance of inter-temporal information, especially in the application of computational social science.
Zoom: https://usc.zoom.us/j/98321628121
WebCast Link: https://usc.zoom.us/j/98321628121
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
PhD Thesis Proposal - Rujun Han
Fri, Sep 10, 2021 @ 01:00 PM - 02:30 PM
Thomas Lord Department of Computer Science
University Calendar
Name: Rujun Han
Time: 2-3:30pm, Sep 10th
Committee Members: Nanyun Peng (Chair), Aram Galstyan, Shrikanth Narayanan, Aiichiro Nakano, Xiang Ren
Title: Event-centric Reasoning in NLP with IE, QA, and NLG Applications
Abstract: Stories or narratives are composed based on events. Knowing how these events are connected logically and semantically is crucial for understanding human written texts. In the proposed thesis, we explore and invent algorithms, as well as construct novel datasets to facilitate future investigation and evaluation of automated event-centric reasoning. Specifically, we design structural deep neural networks, adapt continual pretraining methods, and leverage graph attention models to enhance machines' reasoning skills in text data. We then apply these techniques to solve event-centric tasks in information extraction, reading comprehension, question-answering, and natural language generation.
WebCast Link: https://usc.zoom.us/j/4738647635
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
-
PhD Thesis Proposal - James Preiss
Fri, Sep 24, 2021 @ 11:00 AM - 01:00 PM
Thomas Lord Department of Computer Science
University Calendar
Name:
James A. Preiss
Advisor:
Gaurav S. Sukhatme
Committee:
Nora Ayanian, Stefanos Nikolaidis, Heather Culbertson, Ashutosh Nayyar, Gaurav S. Sukhatme
Title:
Reasoning about sets of dynamical systems:
Identification, adaptive control, and suboptimal coverings
Abstract:
Robots often have physical dynamics that are unknown, but belong to a highly structured set. Knowledge of the set structure is useful for designing adaptive mechanisms, especially if synthesizing a good control policy for each system in the set is straightforward. However, when the systems in the set are complex enough to necessitate black-box methods like reinforcement learning, it is less clear how to use knowledge of the set structure. For example, neural network policies that take dynamics parameters as a secondary input fall short of the performance of single-system policies.
We propose both practical and theoretical inquiries into this difficulty. From the practical side, we propose an alternative to black-box learning methods that combines differentiable simulation, recurrent neural network models, and established methods from nonlinear control. From the theoretical side, we introduce the framework of suboptimal covering numbers to quantify "how much" a good policy must change with respect to dynamics parameters. We bound the suboptimal covering number for a simple class of linear-quadratic systems.
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
