Tiny machine learning (tinyML) is a fast-growing field of machine
learning technologies and applications including algorithms, hardware,
and software capable of performing on-device sensor (vision, audio, IMU,
biomedical, etc.) data analytics at extremely low power, typically in
the mW range and below, and hence enabling a variety of always-on
use-cases and targeting battery-operated devices. tinyML systems are
becoming “good enough” for (i) many commercial applications and new
systems on the horizon; (ii) significant progress is being made on
algorithms, networks, and models down to 100 kB and below; and (iii)
initial low power applications in vision and audio are becoming
mainstream and commercially available. There is growing momentum
demonstrated by technical progress and ecosystem development. The first
annual tinyML research symposium serves as a flagship venue for research
at the intersection of machine learning applications, algorithms,
software, and hardware in deeply embedded machine learning systems. We
solicit papers from academia and industry combining cross-layer
innovations across topics. Submissions must describe tinyML innovations
that intersect and leverage synergy between at least two of the
following subject areas:
tinyML Datasets
* Public release of new datasets to tinyML
* Frameworks that automate dataset development
* Survey and analysis of existing tiny datasets that can be used for research
tinyML Applications
* Novel applications across all fields and emerging use cases
* Discussions about real-world use cases
* User behavior and system-user interaction
* Survey on practical experiences
tinyML Algorithms
* Federated learning or stream-based active learning methods
* Deep learning and traditional machine learning algorithms
* Pruning, quantization, optimization methods
* Security and privacy implications
tinyML Systems
* Profiling tools for measuring and characterizing system performance and power
* Solutions that involve hardware and software co-design
* Characterization of tiny real-world embedded systems
* In-sensor processing, design, and implementation
tinyML Software
* Interpreters and code generator frameworks for tiny systems
* Optimizations for efficient execution
* Software memory optimizations
* Neural architecture search methods
tinyML Hardware
* Power management, reliability, security, performance
* Circuit and architecture design
* Ultra-low-power memory system design
* MCU and accelerator architecture design and evaluation
tinyML Evaluation
* Measurement tools and techniques
* Benchmark creation, assessment and validation
* Evaluation and measurement of real production systems
Accepted papers will be published in the form of peer-reviewed online
proceedings. An author of an accepted paper must attend the research
symposium to give a presentation.
== Program Chairs ==
Vijay Janapa-Reddi, Harvard Univ.
Boris Murmann, Stanford Univ.
== Program Committee ==
Edith Beigne, Facebook
Vikas Chandra, Facebook
Yiran Chen, Duke Univ.
Hiroshi Doyu, Ericsson
Adam Fuks, NXP
Wolfgang Furtner, Infineon
Song Han, MIT
Jeremy Holleman, Syntiant
Prateek Jain, Microsoft
Kurt Keutzer, Berkeley
H. T. Kung, Harvard
Matthew Mattina, ARM
Tinoosh Mohsenin, Univ. of Maryland
Edwin Park, Qualcomm
Priyanka Raina, Stanford Univ.
Jae-sun Seo, ASU
Mingoo Seok, Columbia Univ.
Dennis Sylvester, Univ. of Michigan
Jonathan Tapson, GrAI Matter Labs
Marian Verhelst, KU Leuven
Pete Warden, Google
Hoi-Jun Yoo, KAIST
== Publicity Chair ==
Theocharis Theocharides, Univ. of Cyprus
== Important Deadlines ==
Papers Due: Nov 23rd, 2020 (extended)
Author Notification: Jan 15th, 2021
Camera Ready: Feb 15th, 2021
== Submission Page Limit ==
6 – 8 pages
== Submission Website ==
https://openreview.net/group?id=tinyml.org/tinyML/2021/Research_Symposium
== Submission Format ==
https://www.acm.org/publications/proceedings-template