Picture of Sourav Bhattacharya

Sourav Bhattacharya

Cambridge, UK
Research Scientist

Education

PhD, Computer Science (2014), University of Helsinki, Finland
M.Sc., Computer Science (2009), University of Helsinki, Finland
B.Tech., Computer Science and Engineering (2005), West Bengal University of Technology, India

Biography

Sourav joined Bell Labs as a Postdoctoral Researcher in February 2015, and since February 2016 Sourav has been a Member of the Technical Staff (MTS) at Nokia Bell Labs. He develops novel inference techniques using deep learning and focusses on model compression to bring new inference capabilities to mobile and wearable platforms. Prior to joining Bell Labs, Sourav was a postdoctoral researcher at Aalto University, Finland. Sourav received a PhD and M.Sc. in Computer Science from University of Helsinki, Finland. 

Research Interests

  • Deep Learning
  • IoT/M2M
  • Machine Learning

Honors and Awards

  • Nokia Scholarship 2012

Selected Articles and Publications

2017

Papers:

[1] "Scaling Health Analytics to Millions Without Compromising Privacy using Deep Distributed Behavior Models", Petar Veličković, Nicholas Lane, Sourav Bhattacharya, Angela Chieh, Otmane Bellahsen, and Matthieu Vegreville, PervasiveHealth, 2017.

[2] "Squeezing Deep Learning into Mobile and Embedded Devices", Nicholas Lane, Sourav Bhattacharya, Akhil Mathur, Petko Georgiev, Claudio Forlivesi and Fahim Kawsar, IEEE Pervasive. July-September 2017.

[3] "DeepEye: Resource Efficient Local Execution of Multiple Deep Vision Models using Wearable Commodity Hardware", Akhil Mathur, Nicholas D Lane, Sourav Bhattacharya, Aidan Boran, Claudio Forlivesi, Fahim Kawsar, MobiSys, 2017.

2016

Papers:

[1] "From Smart to Deep: Robust Activity Recognition on Smartwatches Using Deep Learning", Sourav Bhattacharya and Nicholas D. Lane, WristSense'16. (Best Paper Award)

[2] "DeepX: A Software Accelerator for Low-Power Deep Learning Inference on Mobile Devices", Nicholas D. Lane, Sourav Bhattacharya, Petko Georgiev, Claudio Forlivesi, Lei Jiao, Lorena Qendro, and Fahim Kawsar, International Conference on Information Processing in Sensor Networks (IPSN), 2016.

[3] "Understanding The Impact of Geographical Context on the Subjective Well-Being of Urban Citizens", Afra Mashhadi, Sourav Bhattacharya, and Fahim Kawsar, International Conference on IoT in Urban Space (Urb-IoT) 2016.

[4] "Activity Recognition on Smart Devices: Dealing with diversity in the wild", Henrik Blunck, Sourav Bhattacharya, Allan Stisen, Thor Siiger Prentow, Mikkel Baun Kjærgaard, Anind Dey, Mads Møller Jensen, and Tobias Sonne, GetMobile: Mobile Computing and Communications 20(1), 34--38, 2016.

[5] "Towards multimodal deep learning for activity recognition on mobile devices", Valentin Radu, Nicholas D Lane, Sourav Bhattacharya, Cecilia Mascolo, Mahesh K Marina, and Fahim Kawsar, Adjunct Proceedings of ACM International Joint Conference on Pervasive and Ubiquitous Computing, 2016.

[6] "Sparsification and Separation of Deep Learning Layers for Constrained Resource Inference on Wearables", Sourav Bhattacharya and Nicholas D Lane, ACM SenSys, 2016.

[7] "DXTK : Enabling Resource-efficient Deep Learning on Mobile and Embedded Devices with the DeepX Toolkit", Nicholas D. Lane, Sourav Bhattacharya, Akhil Mathur, Claudio Forlivesi, and Fahim Kawsar, MobiCase 2016.

Demo:

[1] "DeepX: A Software Accelerator for Squeezing Deep Learning onto Wearables, Phones and Things", Nicholas D. Lane, Sourav Bhattacharya, Petko Georgiev, Claudio Forlivesi and Fahim Kawsar, Workshop on Mobile Computing Systems and Applications (ACM HotMobile) 2016.

[2] "DeepX: A Software Accelerator for Embedded Deep Learning", Nicholas D. Lane, Sourav Bhattacharya, Petko Georgiev, Claudio Forlivesi, and Fahim Kawsar, International Conference on Information Processing in Sensor Networks (IPSN) 2016.

[3] "Demo: Accelerated Deep Learning Inference for Embedded and Wearable Devices using DeepX", Nicholas D. Lane, Sourav Bhattacharya, Petko Georgiev, Claudio Forlivesi, and Fahim Kawsar, Proceedings of the 14th Annual International Conference on Mobile Systems, Applications, and Services (MobiSys) 2016.

 

2015

Papers:

[1] An Early Resource Characterization of Deep Learning on Wearables, Smartphones and Internet-of-Things Devices, Nicholas D. Lane, Sourav Bhattacharya, Petko Georgiev, Claudio Forlivesi, Fahim Kawsar, IoT-App’15.

[2] Smart Devices are Different: Assessing and Mitigating Mobile Sensing Heterogeneities for Activity Recognition, Allan Stisen, Henrik Blunck, Sourav Bhattacharya, Thor Siiger Prentow, Mikkel Baun Kjærgaard, Anind Dey, Tobias Sonne, and Mads Møller Jensen, The 13th ACM Conference on Embedded Networked Sensor Systems (SenSys'15).

[3] LookAhead: Augmenting Crowdsourced Website Reputation Systems With Predictive Modeling, Sourav Bhattacharya, Otto Huhta, N Asokan, Trust and Trustworthy Computing 2015, LNCS.

[4] Checksum gestures: continuous gestures as an out-of-band channel for secure pairing, Imtiaj Ahmed, Yina Ye, Sourav Bhattacharya, N Asokan, Giulio Jacucci, Petteri Nurmi, and Sasu Tarkoma, Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp'15), Pages: 391-401, ACM.

 

Patents

[4] "A Method To Run Multiple Deep Learning Models On Resource Constrained Hardware", Akhil Mathur, Nicholas D. Lane, Sourav Bhattacharya and Fahim Kawsar,

820576-EP-EPA, 2017.

[3] "Optimization of deep learning models", Claudio Forlivesi, Fahim Kawsar, Sourav Bhattacharya, and Nicholas D. Lane, 819490-EP-EPA, 2016.

[2] "Runtime optimization of convolutional neural networks", Sourav Bhattacharya, Nicholas D. Lane, and Fahim Kawsar, 819710-EP-EPA, 2016.

[1] "Method and device for analyzing sensor data", Sourav Bhattacharya, Nicholas D. Lane, Claudio Forlivesi, and Fahim Kawsar, 819475-EP-EPA, 2016.