. FoggyCache: Cross-Device Approximate Computation Reuse. In MobiCom, 2018.

PDF Project Slides Video

. Tensor index for large scale image retrieval. In Multimedia Systems, 2015.


Research Projects

Approximate computation reuse

A novel caching/memoization paradigm for machine learning based applications. Approximate computation reuse aims to relax the computation reuse requirements from exact matching to approximate matching, so as to exploit the function approximation and input error tolerance properties of these emerging applications to improve system performance.

Dynamic task sizing for data analytic systems

Despite extensive investigation of job scheduling in data-intensive computation frameworks, less consideration has been given to optimizing job partitioning for resource utilization and efficient processing. In light of this, we design a module as a framework extension to automate real-time job partitioning on individual task granularity.


I was a teaching fellow for the following courses at Yale University:


  • Yale Institute for Network Science (YINS) Room 333, 17 Hillhouse Ave, New Haven, CT 06511, USA