SPLASH 2020
Sun 15 - Sat 21 November 2020 Online Conference
Tue 17 Nov 2020 12:00 - 12:20 at SPLASH-I - T-3 Chair(s): Burcu Kulahcioglu Ozkan, Olivier Tardieu
Wed 18 Nov 2020 00:00 - 00:20 at SPLASH-I - T-3 Chair(s): Chengyu Zhang, Ting Cao

Traditionally, IoT devices send collected sensor data to an intelligent cloud where machine learning (ML) inference happens. However, this course is rapidly changing and there is a recent trend to run ML on the edge IoT devices themselves. An intelligent edge is attractive because it saves network round trip (efficiency) and keeps user data at the source (privacy). However, the IoT devices are much more resource constrained than the cloud, which makes running ML on them challenging. Specifically, consider Arduino Uno, a commonly used board, that has 2KB of RAM and 32KB of read-only Flash memory. Although recent breakthroughs in ML have created novel recurrent neural network (RNN) models that provide good accuracy with KB-sized models, deploying them on tiny devices with such hard memory requirements has remained elusive.

We provide, Shiftry, an automatic compiler from high-level floating-point ML models to fixed-point C-programs with 8-bit and 16-bit integers, which have significantly lower memory requirements. For this conversion, Shiftry uses a data-driven float-to-fixed procedure and a RAM management mechanism. These techniques enable us to provide first empirical evaluation of RNNs running on tiny edge devices. On simpler ML models that prior work could handle, Shiftry-generated code has lower latency and higher accuracy.

Tue 17 Nov
Times are displayed in time zone: Central Time (US & Canada) change

11:00 - 12:20: T-3OOPSLA at SPLASH-I +12h
Chair(s): Burcu Kulahcioglu OzkanMPI-SWS, Olivier TardieuIBM Research
11:00 - 11:20
Talk
OOPSLA
Ritwika GhoshUniversity of Illinois at Urbana-Champaign, Chiao HsiehUniversity of Illinois at Urbana-Champaign, Sasa MisailovicUniversity of Illinois at Urbana-Champaign, Sayan MitraUniversity of Illinois at Urbana-Champaign
Link to publication DOI Pre-print Media Attached
11:20 - 11:40
Talk
OOPSLA
Suvam MukherjeeMicrosoft Research, Pantazis DeligiannisMicrosoft Research, Arpita BiswasIISc Bangalore, Akash LalMicrosoft Research
Link to publication DOI Pre-print Media Attached
11:40 - 12:00
Talk
OOPSLA
Umar FarooqUniversity of California at Riverside, Zhijia ZhaoUniversity of California at Riverside, Manu SridharanUniversity of California at Riverside, Iulian NeamtiuNew Jersey Institute of Technology
Link to publication DOI Pre-print Media Attached
12:00 - 12:20
Talk
OOPSLA
Aayan KumarMicrosoft Research, Vivek SeshadriMicrosoft Research, Rahul SharmaMicrosoft Research
Link to publication DOI Media Attached
23:00 - 00:20: T-3OOPSLA at SPLASH-I
Chair(s): Chengyu ZhangEast China Normal University, Ting CaoMicrosoft Research
23:00 - 23:20
Talk
OOPSLA
Ritwika GhoshUniversity of Illinois at Urbana-Champaign, Chiao HsiehUniversity of Illinois at Urbana-Champaign, Sasa MisailovicUniversity of Illinois at Urbana-Champaign, Sayan MitraUniversity of Illinois at Urbana-Champaign
Link to publication DOI Pre-print Media Attached
23:20 - 23:40
Talk
OOPSLA
Suvam MukherjeeMicrosoft Research, Pantazis DeligiannisMicrosoft Research, Arpita BiswasIISc Bangalore, Akash LalMicrosoft Research
Link to publication DOI Pre-print Media Attached
23:40 - 00:00
Talk
OOPSLA
Umar FarooqUniversity of California at Riverside, Zhijia ZhaoUniversity of California at Riverside, Manu SridharanUniversity of California at Riverside, Iulian NeamtiuNew Jersey Institute of Technology
Link to publication DOI Pre-print Media Attached
00:00 - 00:20
Talk
OOPSLA
Aayan KumarMicrosoft Research, Vivek SeshadriMicrosoft Research, Rahul SharmaMicrosoft Research
Link to publication DOI Media Attached