File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1109/ICSICT49897.2020.9278257
- Scopus: eid_2-s2.0-85099173487
Supplementary
-
Citations:
- Scopus: 0
- Appears in Collections:
Conference Paper: HOTCAKE: Higher Order Tucker Articulated Kernels for Deeper CNN Compression
Title | HOTCAKE: Higher Order Tucker Articulated Kernels for Deeper CNN Compression |
---|---|
Authors | |
Keywords | Tensors Kernel Neural networks Matrix decomposition Faces |
Issue Date | 2020 |
Publisher | IEEE. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000707 |
Citation | 2020 IEEE 15th International Conference on Solid-State and Integrated Circuit Technology (ICSICT) Proceedings (ICSICT-2020), Kunming, China, 3-6 November 2020, p. 41-44 How to Cite? |
Abstract | The emerging edge computing has produced immense interests in compacting a neural network without sacrificing much accuracy. In this regard, low-rank tensor decomposition constitutes a powerful tool to compress convolutional neural networks (CNNs) by decomposing the 4-way kernel tensor into multi-stage smaller ones. Building on top of Tucker-2 decomposition, we propose a generalized Higher Order Tucker Articulated Kernels (HOT-CAKE) scheme comprising four steps: input channel decomposition, guided Tucker rank selection, higher order Tucker decomposition and fine-tuning. By subjecting each CONV layer to HOTCAKE, a highly compressed CNN model with graceful accuracy trade-off is obtained. Experiments show HOTCAKE can compress even pre-compressed models and produce state-of-the-art lightweight networks. |
Persistent Identifier | http://hdl.handle.net/10722/301890 |
ISBN |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Lin, R | - |
dc.contributor.author | Ko, C-Y | - |
dc.contributor.author | He, Z | - |
dc.contributor.author | Chen, C | - |
dc.contributor.author | Cheng, Y | - |
dc.contributor.author | Yu, H | - |
dc.contributor.author | Chesi, G | - |
dc.contributor.author | Wong, N | - |
dc.date.accessioned | 2021-08-21T03:28:29Z | - |
dc.date.available | 2021-08-21T03:28:29Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | 2020 IEEE 15th International Conference on Solid-State and Integrated Circuit Technology (ICSICT) Proceedings (ICSICT-2020), Kunming, China, 3-6 November 2020, p. 41-44 | - |
dc.identifier.isbn | 9781728162362 | - |
dc.identifier.uri | http://hdl.handle.net/10722/301890 | - |
dc.description.abstract | The emerging edge computing has produced immense interests in compacting a neural network without sacrificing much accuracy. In this regard, low-rank tensor decomposition constitutes a powerful tool to compress convolutional neural networks (CNNs) by decomposing the 4-way kernel tensor into multi-stage smaller ones. Building on top of Tucker-2 decomposition, we propose a generalized Higher Order Tucker Articulated Kernels (HOT-CAKE) scheme comprising four steps: input channel decomposition, guided Tucker rank selection, higher order Tucker decomposition and fine-tuning. By subjecting each CONV layer to HOTCAKE, a highly compressed CNN model with graceful accuracy trade-off is obtained. Experiments show HOTCAKE can compress even pre-compressed models and produce state-of-the-art lightweight networks. | - |
dc.language | eng | - |
dc.publisher | IEEE. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000707 | - |
dc.relation.ispartof | International Conference on Solid-State and Integrated Circuit Technology Proceedings | - |
dc.rights | International Conference on Solid-State and Integrated Circuit Technology Proceedings. Copyright © IEEE. | - |
dc.rights | ©2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. | - |
dc.subject | Tensors | - |
dc.subject | Kernel | - |
dc.subject | Neural networks | - |
dc.subject | Matrix decomposition | - |
dc.subject | Faces | - |
dc.title | HOTCAKE: Higher Order Tucker Articulated Kernels for Deeper CNN Compression | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Chesi, G: chesi@eee.hku.hk | - |
dc.identifier.email | Wong, N: nwong@eee.hku.hk | - |
dc.identifier.authority | Chesi, G=rp00100 | - |
dc.identifier.authority | Wong, N=rp00190 | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1109/ICSICT49897.2020.9278257 | - |
dc.identifier.scopus | eid_2-s2.0-85099173487 | - |
dc.identifier.hkuros | 324497 | - |
dc.identifier.spage | 41 | - |
dc.identifier.epage | 44 | - |
dc.publisher.place | United States | - |