Abstract
A novel hardware architecture of the competitive learning (CL) algorithm with k-winners-take-all activation is presented in this paper. It is used as a custom logic block in the arithmetic logic unit (ALU) of the softcore NIOS processor for CL training. Both the partial distance search (PDS) module and hardware divider adopt finite precision calculation for area cost reduction at the expense of slight degradation in training performance. The PDS module also employs subspace search and multiple-coefficient accumulation techniques for effective reduction of the computation latency for the PDS search. Experiment results show that the CPU time is lower than that of Pentium IV processors running the CL training program without the support of custom hardware.
| Original language | English |
|---|---|
| Title of host publication | Next-Generation Applied Intelligence - 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2009, Proceedings |
| Pages | 594-603 |
| Number of pages | 10 |
| DOIs | |
| Publication status | Published - 2009 Nov 16 |
| Event | 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2009 - Tainan, Taiwan Duration: 2009 Jun 24 → 2009 Jun 27 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 5579 LNAI |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Other
| Other | 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2009 |
|---|---|
| Country | Taiwan |
| City | Tainan |
| Period | 09/6/24 → 09/6/27 |
Fingerprint
ASJC Scopus subject areas
- Theoretical Computer Science
- Computer Science(all)
Cite this
High speed k-winner-take-all competitive learning in reconfigurable hardware. / Li, Hui Ya; Yeh, Yao Jung; Hwang, Wen-Jyi; Yang, Cheng Tsun.
Next-Generation Applied Intelligence - 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2009, Proceedings. 2009. p. 594-603 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5579 LNAI).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - High speed k-winner-take-all competitive learning in reconfigurable hardware
AU - Li, Hui Ya
AU - Yeh, Yao Jung
AU - Hwang, Wen-Jyi
AU - Yang, Cheng Tsun
PY - 2009/11/16
Y1 - 2009/11/16
N2 - A novel hardware architecture of the competitive learning (CL) algorithm with k-winners-take-all activation is presented in this paper. It is used as a custom logic block in the arithmetic logic unit (ALU) of the softcore NIOS processor for CL training. Both the partial distance search (PDS) module and hardware divider adopt finite precision calculation for area cost reduction at the expense of slight degradation in training performance. The PDS module also employs subspace search and multiple-coefficient accumulation techniques for effective reduction of the computation latency for the PDS search. Experiment results show that the CPU time is lower than that of Pentium IV processors running the CL training program without the support of custom hardware.
AB - A novel hardware architecture of the competitive learning (CL) algorithm with k-winners-take-all activation is presented in this paper. It is used as a custom logic block in the arithmetic logic unit (ALU) of the softcore NIOS processor for CL training. Both the partial distance search (PDS) module and hardware divider adopt finite precision calculation for area cost reduction at the expense of slight degradation in training performance. The PDS module also employs subspace search and multiple-coefficient accumulation techniques for effective reduction of the computation latency for the PDS search. Experiment results show that the CPU time is lower than that of Pentium IV processors running the CL training program without the support of custom hardware.
UR - http://www.scopus.com/inward/record.url?scp=71049188555&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=71049188555&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-02568-6_60
DO - 10.1007/978-3-642-02568-6_60
M3 - Conference contribution
SN - 3642025676
SN - 9783642025679
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 594
EP - 603
BT - Next-Generation Applied Intelligence - 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2009, Proceedings
ER -