DC Field | Value | Language |
---|---|---|
dc.contributor.author | Han, Donghyeon | ko |
dc.contributor.author | Im, DongSeok | ko |
dc.contributor.author | Park, Gwangtae | ko |
dc.contributor.author | Kim, Youngwoo | ko |
dc.contributor.author | Song, Seokchan | ko |
dc.contributor.author | Lee, Juhyoung | ko |
dc.contributor.author | Yoo, Hoi-Jun | ko |
dc.date.accessioned | 2023-01-31T13:00:39Z | - |
dc.date.available | 2023-01-31T13:00:39Z | - |
dc.date.created | 2023-01-09 | - |
dc.date.issued | 2022-06 | - |
dc.identifier.citation | 4th IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2022, pp.37 - 40 | - |
dc.identifier.uri | http://hdl.handle.net/10203/304905 | - |
dc.description.abstract | A DNN training processor with a maximum of 332 TOPS/W is proposed for efficient and robust object detection. The proposed processor is able to support both quantization and pruning-based personalization to make a user-optimized lightweight network. In addition to personalization, it supports real-time adaptation to compensate for accuracy degradation caused by environmental changes or unpredictable situations. It maintains conventional input slice skipping architecture and stochastic rounding-based computing for the efficient acceleration of the DNN training. It further improves efficiency by removing pseudo-RNGs during the stochastic rounding and adding blocks to pruning-aware training. Moreover, it suggests an LT-flag-based reconfigurable accumulation network and enables multi-learning-task-allocation for low-latency DNN training with the backward unlocking solution. Fabricated in 28-nm technology, the proposed processor demonstrates 46.6 FPS object detection with 0.95 mJ/frame energy consumption which is the state-of-the-art performance compared with the previous processors. | - |
dc.language | English | - |
dc.publisher | Institute of Electrical and Electronics Engineers Inc. | - |
dc.title | A 0.95 mJ/frame DNN Training Processor for Robust Object Detection with Real-World Environmental Adaptation | - |
dc.type | Conference | - |
dc.identifier.wosid | 000859273200011 | - |
dc.identifier.scopusid | 2-s2.0-85139063344 | - |
dc.type.rims | CONF | - |
dc.citation.beginningpage | 37 | - |
dc.citation.endingpage | 40 | - |
dc.citation.publicationname | 4th IEEE International Conference on Artificial Intelligence Circuits and Systems, AICAS 2022 | - |
dc.identifier.conferencecountry | KO | - |
dc.identifier.conferencelocation | Songdo Convensia, Incheon | - |
dc.identifier.doi | 10.1109/AICAS54282.2022.9869960 | - |
dc.contributor.localauthor | Yoo, Hoi-Jun | - |
dc.contributor.nonIdAuthor | Kim, Youngwoo | - |
dc.contributor.nonIdAuthor | Song, Seokchan | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.