[Colloquium] Reminder - Chengcheng Wan Dissertation Defense/Apr 8, 2022

[Megan Woodward]

This is an announcement of Chengcheng Wan's Dissertation Defense.
===============================================
Candidate: Chengcheng Wan

Date: Friday, April 08, 2022

Time:  1:30 pm CST

Remote Location: https://zoom.us/j/93204773188?pwd=Qk54QUttYVZ0QnI4dHFGQnp3RzZoQT09<https://urldefense.com/v3/__https://zoom.us/j/93204773188?pwd=Qk54QUttYVZ0QnI4dHFGQnp3RzZoQT09__;!!BpyFHLRN4TMTrA!uaQlGUzXOvdN-8k2FEsmLW4dRpDd8iukHftTSJ9fquNWKTDzUGEfLqt5CFi4e-nKv5Vn3vF7$> Meeting ID: 932 0477 3188 Passcode: D70JDv


Title: Correctness, Performance, And Energy-Efficiency: Improving Software Systems That Use Machine Learning Components

Abstract: Machine learning (ML) provides efficient solutions for a number of problems that are difficult to solve with traditional computing techniques. Deep neural networks (DNNs) have become a key workload for many computing systems due to their high inference accuracy. This accuracy, however, comes at a cost of long latency, high energy usage, and engineering effort. It requires tremendous human effort and domain knowledge to implement a correct, efficient, and robust ML software.

To improve the correctness, performance, and energy-efficiency of machine learning software systems, this dissertation works on these three parts and makes the following contributions:
First, to improve the flexibility of neural networks, this dissertation proposes a novel neural network architecture and a customized optimizer that support anytime prediction. This design allows one neural network to generate a series of increasingly accurate outputs over time without sacrificing accuracy for flexibility.

Second, this dissertation designs a runtime scheduler ALERT, which holistically configures neural networks and system resources together to meet application-specific accuracy, performance, and energy-consumption constraints. It uses a probabilistic model to detect environmental volatility and makes use of the full potential of the DNN candidate set to optimize performance and satisfy constraints.

Third, in the scope of software applications, this dissertation conducts the first comprehensive study about how real-world applications are using machine learning cloud APIs. We generalize 8 anti-patterns that degrade functional, performance, or economical quality of the software. Guided by this study, we propose Keeper, a new testing framework for software systems that use machine learning APIs. Keeper automatically generates many test cases to thoroughly test every branch in the specified function and its callees. It analyzes the test runs and reports many failures, as well as potential patches, to developers.

Advisors: Shan Lu

Committee Members: Shan Lu, Hank Hoffmann, and Michael Maire


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20220407/d3c639c4/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: wan_thesis_draft.pdf
Type: application/pdf
Size: 4023163 bytes
Desc: wan_thesis_draft.pdf
URL: <http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20220407/d3c639c4/attachment-0001.pdf>