About Me
I am Deyu Tian, a fifth-year Ph.D. student from the OS Lab of Software Institute of Peking University. I major in Software Engineering, advised by Prof. Yun Ma, Xuanzhe Liu, and Gang Huang. My research interest includes the Web ecosystem and deep learning infrastructures.Education
Ph.D. in Software Engineering
At present I am a fifth-year Ph.D. candidate at OS Lab, Peking University, and I major in Software Engineering. I research in the field of Web ecosystem and deep learning infrastructures, advised by Prof Yun Ma, Xuanzhe Liu, and Gang Huang.
I was a Teaching Assistant (TA) for Introduction to Computing (A) from 2018 to 2022. I was also a TA for Scientific Paper Writing at 2021.
I was granted Award for Scientific Research by Peking University at 2019, and Award for Academic Excellence by Peking University at 2021.
Bachelor in Computer Science and Technology
I enrolled in the EECS school of Peking University at 2014, and got my bachelor degree of computer science and technology at 2018.
I was the vice precident of Students' Computer Association at School of EECS at 2016, during which I helped undertake the 2016 ACM-ICPC Asia Beijing Regional Contest.
I was granted Merit Student Award and Founder Scholarship by Peking University at 2016.
Awards
2021
Award for Academic Excellence
Peking University2019
Award for Scientific Research
Peking University2016
Founder Scholarship
Peking UniversityMerit Student
Peking UniversityThird Prize of Students' Programming Contest
School of EECS, Peking University2015
Third Prize of Students' Programming Contest
School of EECS, Peking University2013
Bronze Medal of National Olympics in Informatics
China Computer Foundation2012
First Prize of National Olympics in Informatics in Provinces
China Computer FoundationProjects
Only some selected projects are introduced here.A heterogeneous DNN inference engine for Web apps
I designed a heterogeneous DNN engine for Web apps, named WEngine, which uses pipeline parallelism to optimize DNN inference throughput in Web browsers. I modified TensorFlow.js to implement WEngine, including programming about WebGL, WebAssembly and Web worker.
This work has been published in MobiSys 2022.
A measurement study of deep learning in Web browsers
We evaluated the performance of different deep learning inference frameworks for Web apps. I implemented a performance measurement tool, which can control the experiment procedure and collect the measurement results.
This work has been published in WWW 2019.
An empirical study of embedded Web pages in mobile apps
I conducted an empirical study about embedded Web pages in mobile applications. I implemented an automated testing tool targeting the embedded Web pages in mobile apps. Then I revealed the prevalence, characteristics, and performance of the embedded Web pages.
This work has been published in IEEE TMC 2021.