Robust I/O-compute concurrency for machine learning pipelines in constrained cyber-physical devices
Pages 1 - 11
Abstract
Cyberphysical systems have numerous industrial and commercial applications. Such systems are often built using low-resource devices that gather and process data, using machine-learning (ML) models, to make intelligent decisions and provide value to users. Programming such low-resource devices with an impoverished system runtime is often challenging.
This paper presents a new domain-specific language called PiCon for programming ML pipelines in low-resource devices. PiCon allows safe I/O-compute concurrency, ruling out a large class of errors, while providing a simple and sequential coding abstraction to the programmer. PiCon compiles to C code and easily interfaces with existing C/C++ code. Furthermore, the generated code does not rely on multi-threading support or dynamic memory allocation, dramatically reducing its footprint on the device. We present experience porting two real-world ML applications that demonstrate simplification in programmability, in addition to several safe-by-construction guarantees.
References
[1]
Motor Industry Software Reliability Association. 2008. MISRA C++: 2008: guidelines for the use of the C++ language in critical systems. MIRA.
[2]
Charles Bell. 2017. MicroPython for the Internet of Things. Springer. isbn:978-1-4842-3123-4
[3]
Don Kurian Dennis, Shishir G Patil, Chirag Pabbaraju, Nadeem Shaheer, Harsha Simhadri, Vivek Seshadri, Manik Varma, and Prateek Jain. 2018. GesturePod: Gesture-based Interaction Cane for People with Visual Impairments. Technical Report MSR-TR-2018–14. Microsoft.
[4]
Dennis, Don Kurian and Gaurkar, Yash and Gopinath, Sridhar and Goyal, Sachin and Gupta, Chirag and Jain, Moksh and Jaiswal, Shikhar and Kumar, Ashish and Kusupati, Aditya and Lovett, Chris and Patil, Shishir G and Saha, Oindrila and Simhadri, Harsha Vardhan. [n.d.]. EdgeML: Machine Learning for resource-constrained edge devices. https://github.com/Microsoft/EdgeML
[5]
Adam Dunkels, Bjorn Gronvall, and Thiemo Voigt. 2004. Contiki-a lightweight and flexible operating system for tiny networked sensors. In 29th annual IEEE international conference on local computer networks. 455–462. https://doi.org/10.1109/LCN.2004.38
[6]
Adam Dunkels, Fredrik Osterlind, Nicolas Tsiftes, and Zhitao He. 2007. Software-based on-line energy estimation for sensor nodes. In Proceedings of the 4th workshop on Embedded networked sensors. 28–32. https://doi.org/10.1145/1278972.1278979
[7]
Adam Dunkels, Oliver Schmidt, Thiemo Voigt, and Muneeb Ali. 2006. Protothreads: Simplifying event-driven programming of memory-constrained embedded systems. In Proceedings of the 4th international conference on Embedded networked sensor systems. 29–42. https://doi.org/10.1145/1182807.1182811
[8]
Chirag Gupta, Arun Sai Suggala, Ankit Goyal, Harsha Vardhan Simhadri, Bhargavi Paranjape, Ashish Kumar, Saurabh Goyal, Raghavendra Udupa, Manik Varma, and Prateek Jain. 2017. ProtoNN: compressed and accurate kNN for resource-scarce devices. In International Conference on Machine Learning. 1331–1340.
[9]
John Hunt. 2019. Concurrency with AsyncIO. In Advanced Guide to Python 3 Programming. Springer, 407–417. https://doi.org/10.1007/978-3-030-25943-3_34
[10]
Texas Instruments. [n.d.]. MS430 Microcontroller. https://www.ti.com/lit/ug/slau395f/slau395f.pdf
[11]
Aditya Kusupati, Manish Singh, Kush Bhatia, Ashish Kumar, Prateek Jain, and Manik Varma. 2018. FastGRNN: A Fast, Accurate, Stable and Tiny Kilobyte Sized Gated Recurrent Neural Network. In Advances in Neural Information Processing Systems. 9031–9042.
[12]
Edward A Lee. 2006. Cyber-physical systems-are computing foundations adequate. In Position Paper for NSF Workshop On Cyber-Physical Systems: Research Motivation, Techniques and Roadmap. 2, 1–9.
[13]
Edward A Lee. 2008. Cyber physical systems: Design challenges. In 2008 11th IEEE International Symposium on Object and Component-Oriented Real-Time Distributed Computing (ISORC). 363–369. https://doi.org/10.1109/ISORC.2008.25
[14]
Philip Levis, Sam Madden, Joseph Polastre, Robert Szewczyk, Kamin Whitehouse, Alec Woo, David Gay, Jason Hill, Matt Welsh, and Eric Brewer. 2005. TinyOS: An operating system for sensor networks. In Ambient intelligence. Springer, 115–148. https://doi.org/10.1007/3-540-27139-2_7
[15]
Nicholas D Matsakis and Felix S Klock. 2014. The rust language. ACM SIGAda Ada Letters, 34, 3 (2014), 103–104. https://doi.org/10.1145/2692956.2663188
[16]
William P McCartney and Nigamanth Sridhar. 2006. Abstractions for safe concurrent programming in networked embedded systems. In Proceedings of the 4th international conference on Embedded networked sensor systems. 167–180. https://doi.org/10.1145/1182807.1182825
[17]
Semih Okur, David L Hartveld, Danny Dig, and Arie van Deursen. 2014. A study and toolkit for asynchronous programming in C#. In Proceedings of the 36th International Conference on Software Engineering. 1117–1127. https://doi.org/10.1145/2568225.2568309
[18]
Benjamin C Pierce. 2005. Advanced topics in types and programming languages. MIT press.
[19]
J. A. Stankovic, I. Lee, A. Mok, and R. Rajkumar. 2005. Opportunities and obligations for physical computing systems. Computer, 38, 11 (2005), Nov, 23–31. issn:0018-9162 https://doi.org/10.1109/MC.2005.386
[20]
STMicroelectronics. [n.d.]. STM8L050J3 Microcontroller. https://www.st.com/en/microcontrollers-microprocessors/stm8l050j3.html
Index Terms
- Robust I/O-compute concurrency for machine learning pipelines in constrained cyber-physical devices
Recommendations
Compositional Falsification of Cyber-Physical Systems with Machine Learning Components
AbstractCyber-physical systems (CPS), such as automotive systems, are starting to include sophisticated machine learning (ML) components. Their correctness, therefore, depends on properties of the inner ML modules. While learning algorithms aim to ...
Machine learning for intrusion detection in industrial control systems: Applications, challenges, and recommendations
AbstractMethods from machine learning are used in the design of secure Industrial Control Systems. Such methods focus on two major areas: detection of intrusions at the network level using the information acquired through network packets, and detection ...
Enabling Machine Learning on Resource Constrained Devices by Source Code Generation of the Learned Models
Computational Science – ICCS 2018AbstractDue to the development of IoT solutions, we can observe the constantly growing number of these devices in almost every aspect of our lives. The machine learning may improve increase their intelligence and smartness. Unfortunately, the highly ...
Comments
Information & Contributors
Information
Published In
June 2021
162 pages
Copyright © 2021 ACM.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]
Sponsors
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 22 June 2021
Check for updates
Author Tags
Qualifiers
- Research-article
Conference
LCTES '21
Sponsor:
LCTES '21: 22nd ACM SIGPLAN/SIGBED International Conference on Languages, Compilers, and Tools for Embedded Systems
June 22, 2021
Virtual, Canada
Acceptance Rates
Overall Acceptance Rate 116 of 438 submissions, 26%
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 125Total Downloads
- Downloads (Last 12 months)8
- Downloads (Last 6 weeks)4
Reflects downloads up to 15 Oct 2024
Other Metrics
Citations
View Options
Get Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in