Hardware for machine learning: Challenges and opportunities
Machine learning plays a critical role in extracting meaningful information out of the zetabytes of sensor data collected every day. For some applications, the goal is to analyze and understand the data to identify trends (e.g., surveillance, portable/wearable electronics); in other applications, th...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2017
|
| Online Access: | http://hdl.handle.net/1721.1/112983 https://orcid.org/0000-0003-4841-3990 https://orcid.org/0000-0002-4403-956X https://orcid.org/0000-0002-0376-4220 https://orcid.org/0000-0002-0619-8199 |
| _version_ | 1811092195633004544 |
|---|---|
| author | Sze, Vivienne Chen, Yu-Hsin Einer, Joel Suleiman, Amr AbdulZahir Zhang, Zhengdong |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Sze, Vivienne Chen, Yu-Hsin Einer, Joel Suleiman, Amr AbdulZahir Zhang, Zhengdong |
| author_sort | Sze, Vivienne |
| collection | MIT |
| description | Machine learning plays a critical role in extracting meaningful information out of the zetabytes of sensor data collected every day. For some applications, the goal is to analyze and understand the data to identify trends (e.g., surveillance, portable/wearable electronics); in other applications, the goal is to take immediate action based the data (e.g., robotics/drones, self-driving cars, smart Internet of Things). For many of these applications, local embedded processing near the sensor is preferred over the cloud due to privacy or latency concerns, or limitations in the communication bandwidth. However, at the sensor there are often stringent constraints on energy consumption and cost in addition to throughput and accuracy requirements. Furthermore, flexibility is often required such that the processing can be adapted for different applications or environments (e.g., update the weights and model in the classifier). In many applications, machine learning often involves transforming the input data into a higher dimensional space, which, along with programmable weights, increases data movement and consequently energy consumption. In this paper, we will discuss how these challenges can be addressed at various levels of hardware design ranging from architecture, hardware-friendly algorithms, mixed-signal circuits, and advanced technologies (including memories and sensors). |
| first_indexed | 2024-09-23T15:14:29Z |
| format | Article |
| id | mit-1721.1/112983 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T15:14:29Z |
| publishDate | 2017 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1129832022-09-29T13:37:22Z Hardware for machine learning: Challenges and opportunities Sze, Vivienne Chen, Yu-Hsin Einer, Joel Suleiman, Amr AbdulZahir Zhang, Zhengdong Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Sze, Vivienne Chen, Yu-Hsin Einer, Joel Suleiman, Amr AbdulZahir Zhang, Zhengdong Machine learning plays a critical role in extracting meaningful information out of the zetabytes of sensor data collected every day. For some applications, the goal is to analyze and understand the data to identify trends (e.g., surveillance, portable/wearable electronics); in other applications, the goal is to take immediate action based the data (e.g., robotics/drones, self-driving cars, smart Internet of Things). For many of these applications, local embedded processing near the sensor is preferred over the cloud due to privacy or latency concerns, or limitations in the communication bandwidth. However, at the sensor there are often stringent constraints on energy consumption and cost in addition to throughput and accuracy requirements. Furthermore, flexibility is often required such that the processing can be adapted for different applications or environments (e.g., update the weights and model in the classifier). In many applications, machine learning often involves transforming the input data into a higher dimensional space, which, along with programmable weights, increases data movement and consequently energy consumption. In this paper, we will discuss how these challenges can be addressed at various levels of hardware design ranging from architecture, hardware-friendly algorithms, mixed-signal circuits, and advanced technologies (including memories and sensors). United States. Defense Advanced Research Projects Agency (Young Faculty Award) Massachusetts Institute of Technology. Center for Integrated Circuits and Systems TSMC University Shuttle Texas Instruments Intel 2017-12-29T18:59:23Z 2017-12-29T18:59:23Z 2017-07 2017-04 Article http://purl.org/eprint/type/ConferencePaper 978-1-5090-5191-5 http://hdl.handle.net/1721.1/112983 Sze, Vivienne, et al. "Hardware for Machine Learning: Challenges and Opportunities." Custom Integrated Circuits Conference (CICC), 30 April - 3 May, 2017, Austin, TX, IEEE, 2017, pp. 1–8. https://orcid.org/0000-0003-4841-3990 https://orcid.org/0000-0002-4403-956X https://orcid.org/0000-0002-0376-4220 https://orcid.org/0000-0002-0619-8199 en_US http://dx.doi.org/10.1109/CICC.2017.7993626 2017 IEEE Custom Integrated Circuits Conference (CICC) Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) arXiv |
| spellingShingle | Sze, Vivienne Chen, Yu-Hsin Einer, Joel Suleiman, Amr AbdulZahir Zhang, Zhengdong Hardware for machine learning: Challenges and opportunities |
| title | Hardware for machine learning: Challenges and opportunities |
| title_full | Hardware for machine learning: Challenges and opportunities |
| title_fullStr | Hardware for machine learning: Challenges and opportunities |
| title_full_unstemmed | Hardware for machine learning: Challenges and opportunities |
| title_short | Hardware for machine learning: Challenges and opportunities |
| title_sort | hardware for machine learning challenges and opportunities |
| url | http://hdl.handle.net/1721.1/112983 https://orcid.org/0000-0003-4841-3990 https://orcid.org/0000-0002-4403-956X https://orcid.org/0000-0002-0376-4220 https://orcid.org/0000-0002-0619-8199 |
| work_keys_str_mv | AT szevivienne hardwareformachinelearningchallengesandopportunities AT chenyuhsin hardwareformachinelearningchallengesandopportunities AT einerjoel hardwareformachinelearningchallengesandopportunities AT suleimanamrabdulzahir hardwareformachinelearningchallengesandopportunities AT zhangzhengdong hardwareformachinelearningchallengesandopportunities |