Tingxin Yan PhD Candidate Department of Computer Science University of Massachusetts Amherst Phone: (413)695-6816 Email: yan AT cs DOT umass DOT edu

Bio

I am a fifth-year PhD student in the Department of Computer Science, UMASS Amherst. My advisor is Professor Deepak Ganesan. I have also been working with Professor R. Manmatha, Mark Corner, and Don Towsley. I received my MS degree from Chinese Academy of Sciences, Beijing, China and BS degree from Nanjing University, Nanjing, China.

Research Interests

My research interests are primarily in the areas of mobile and ubiquitous computing, distributed systems, wireless sensor networks, and crowdsourcing. My current research include crowdsourcing-based mobile applications, context-aware mobile system, and distributed mobile sensing. I am also interested in data management for mobile phones, and data mining from large mobile application traces.

Publications

Fast App Launching for Mobile Devices using Predictive User Context. [PDF upon request]
Tingxin Yan, David Chu, Deepak Ganesan, Jie Liu, and Aman Kansal.
In submission to ACM MobiSys 2012

CrowdPark: A Crowdsourcing-based Parking Reservation System for Mobile Phones.[Technical Report]
Tingxin Yan, Baik Hoh, Deepak Ganesan, Ken Tracton, Toch Iwuchukwu, and Juong-Sik Lee.
UMASS Technical Report, UM-CS-2011-001.

CrowdSearch: Exploiting Crowds for Accurate Real-time Image Search on Mobile Phones. [Paper], [Slides], [Bibtex]
Tingxin Yan, Vikas Kumar, and Deepak Ganesan.
In Proceedings of MobiSys 2010. San Francisco, USA. June 2010. (Acceptance rate: 25/124=20.2%)

mCrowd: A Platform for Mobile Crowdsourcing. [Demo Abstract]
Tingxin Yan, Matt Marzilli, Ryan Holmes, Deepak Ganesan, and Mark Corner.
In Proceedings of SenSys 2009. Berkeley, USA. June 2009.(Demo)

Distributed Image Search in Camera Sensor Networks. [Paper], [Slides], [Bibtex]
Tingxin Yan, Deepak Ganesan and R. Manmatha.
In Proceedings of Sensys 2008. Raleigh, USA. Nov 2008. (Acceptance rate: 25/153=16.3%)

Design and Implementation of a Dual-Camera Wireless Sensor Network for Object Retrieval. [Paper], [Bibtex]
Dan Xie, Tingxin Yan, Deepak Ganesan and Allen Hanson.
In Proceedings of IPSN 2008. St. Louis, USA. April 2008. (Acceptance rate: 30/126=23.8%)

Multi-user Data Sharing in Radar Sensor Networks. [Paper], [slides], [Bibtex]
Ming Li, Tingxin Yan, Deepak Ganesan, Eric Lyons, Prashant Shenoy, Arun Venkataramani, and Michael Zink.
In Proceedings of Sensys 2007. Sydney, Australia. Nov 2007. (Acceptance rate: 25/149=16.8%)

Teaching

CS377 Operating Systems, Spring 2011

Projects

FALCON: A context-aware mobile app preloading component for mobile OS. Based on intensive data analysis of app usage across multiple mobile users, FALCON presents a decision engine which exploits temporal and spacial characters of user behaviour to pre-load apps ahead of time, thereby improves the responsiveness of smartphones.

CrowdPark: A crowdsourcing-based vacant parking information sharing system. CrowdPark incentivizes crowdsourcing users to contribute available parking information using mobile phones, and uses sensing-based approaches to imporve the accuracy of parking information reports, and avoid malicious users.

CrowdSearch: A realtime image search engine with real-time human validation to improve the accuracy of automated search results. The core part of this project is the analysis of the trade-offs between incentives, delay, and accuracy in crowdsourcing services, and the modeling of delay and accuracy behavior of crowdsourcing participants. A predictive algorithm based on the user behavior modeling is invented to provide highly accurate image search results within a bounded delay.

mCrowd: a platform that simplifies the process of publishing mobile sensing tasks and make them exposed to a large number of mobile users, as well as incentivizes mobile users to participate to mobile sensing related crowdsourcing tasks and simplifies the process of contributing sensing data to these tasks.

SenSearch: A distributed image search engine to efficiently search across images stored on distributed camera sensors, and/or cellphones.

DualCam: A sensor platform that uses heterogeneous cameras for efficient object detection and recognition.

Muds: Design of a progressive data transmission scheme to optimize utility across multiple diverse users (meteorologist, researcher, emergency manager) in a high data-rate radar sensor network.