Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE

Abstract: 

A promising way to deploy Artificial Intelligence (AI)-based services on mobile devices is to run a part of the AI model (a deep neural network) on the mobile itself, and the rest in the cloud. This is sometimes referred to as collaborative intelligence. In this framework, intermediate features from the deep network need to be transmitted to the cloud for further processing. We study the case where such features are used for multiple purposes in the cloud (multi-tasking) and where they need to be compressible in order to allow efficient transmission to the cloud. To this end, we introduce a new loss function that encourages feature compressibility while improving system performance on multiple tasks. Experimental results show that with the compression-friendly loss, one can achieve around 20% bitrate reduction without sacrificing the performance on several vision-related tasks.

up
0 users have voted:

Paper Details

Authors:
Submitted On:
20 September 2019 - 2:18pm
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Saeed Ranjbar Alvar
Paper Code:
3572
Document Year:
2019
Cite

Document Files

ICIP_2019.pptx

(9)

Subscribe

[1] , "MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4793. Accessed: Oct. 18, 2019.
@article{4793-19,
url = {http://sigport.org/4793},
author = { },
publisher = {IEEE SigPort},
title = {MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE},
year = {2019} }
TY - EJOUR
T1 - MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE
AU -
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4793
ER -
. (2019). MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE. IEEE SigPort. http://sigport.org/4793
, 2019. MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE. Available at: http://sigport.org/4793.
. (2019). "MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE." Web.
1. . MULTI-TASK LEARNING WITH COMPRESSIBLE FEATURES FOR COLLABORATIVE INTELLIGENCE [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4793