Improved margin multi-class classification using dendritic neurons with morphological learning
We present an architecture of a spike based multiclass classifier using neurons with non-linear dendrites and sparse synaptic connectivity where each synapse takes a binary value. The learning in this model happens not through weight updates but through structural changes, i.e. a change of connectiv...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/100389 http://hdl.handle.net/10220/25710 |
| _version_ | 1824453541141413888 |
|---|---|
| author | Hussain, Shaista Liu, Shih-Chii Basu, Arindam |
| author2 | School of Electrical and Electronic Engineering |
| author_facet | School of Electrical and Electronic Engineering Hussain, Shaista Liu, Shih-Chii Basu, Arindam |
| author_sort | Hussain, Shaista |
| collection | NTU |
| description | We present an architecture of a spike based multiclass classifier using neurons with non-linear dendrites and sparse synaptic connectivity where each synapse takes a binary value. The learning in this model happens not through weight updates but through structural changes, i.e. a change of connectivity between inputs and dendrites. Hence, it is well suited for implementation in neuromorphic systems using address event representation (AER). We present a new learning rule that allows better generalization of the system to noisy testing data making it feasible to transfer learnt weights in software to a hardware device interfacing with noisy spiking sensors. The new rule improves testing accuracy by 7 - 10% compared to earlier versions. We also present preliminary results for multi-class classification on handwritten digits from the MNIST database and show that our system can attain comparable performance (≈ 3% more error) with other reported spike based classifiers while using at least 50% less synaptic resources. |
| first_indexed | 2025-02-19T03:08:03Z |
| format | Conference Paper |
| id | ntu-10356/100389 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:08:03Z |
| publishDate | 2015 |
| record_format | dspace |
| spelling | ntu-10356/1003892020-03-07T13:24:50Z Improved margin multi-class classification using dendritic neurons with morphological learning Hussain, Shaista Liu, Shih-Chii Basu, Arindam School of Electrical and Electronic Engineering IEEE International Symposium on Circuits and Systems (ISCAS)(2014:Melbourne) DRNTU::Engineering::Electrical and electronic engineering::Electronic circuits We present an architecture of a spike based multiclass classifier using neurons with non-linear dendrites and sparse synaptic connectivity where each synapse takes a binary value. The learning in this model happens not through weight updates but through structural changes, i.e. a change of connectivity between inputs and dendrites. Hence, it is well suited for implementation in neuromorphic systems using address event representation (AER). We present a new learning rule that allows better generalization of the system to noisy testing data making it feasible to transfer learnt weights in software to a hardware device interfacing with noisy spiking sensors. The new rule improves testing accuracy by 7 - 10% compared to earlier versions. We also present preliminary results for multi-class classification on handwritten digits from the MNIST database and show that our system can attain comparable performance (≈ 3% more error) with other reported spike based classifiers while using at least 50% less synaptic resources. MOE (Min. of Education, S’pore) Accepted version 2015-06-02T00:58:05Z 2019-12-06T20:21:36Z 2015-06-02T00:58:05Z 2019-12-06T20:21:36Z 2014 2014 Conference Paper Hussain, S., Liu, S. C., & Basu, A. (2014). Improved margin multi-class classification using dendritic neurons with morphological learning. 2014 IEEE International Symposium on Circuits and Systems (ISCAS), 2640-2643. https://hdl.handle.net/10356/100389 http://hdl.handle.net/10220/25710 10.1109/ISCAS.2014.6865715 en © 2014 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. The published version is available at: [http://dx.doi.org/10.1109/ISCAS.2014.6865715]. 4 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Electrical and electronic engineering::Electronic circuits Hussain, Shaista Liu, Shih-Chii Basu, Arindam Improved margin multi-class classification using dendritic neurons with morphological learning |
| title | Improved margin multi-class classification using dendritic neurons with morphological learning |
| title_full | Improved margin multi-class classification using dendritic neurons with morphological learning |
| title_fullStr | Improved margin multi-class classification using dendritic neurons with morphological learning |
| title_full_unstemmed | Improved margin multi-class classification using dendritic neurons with morphological learning |
| title_short | Improved margin multi-class classification using dendritic neurons with morphological learning |
| title_sort | improved margin multi class classification using dendritic neurons with morphological learning |
| topic | DRNTU::Engineering::Electrical and electronic engineering::Electronic circuits |
| url | https://hdl.handle.net/10356/100389 http://hdl.handle.net/10220/25710 |
| work_keys_str_mv | AT hussainshaista improvedmarginmulticlassclassificationusingdendriticneuronswithmorphologicallearning AT liushihchii improvedmarginmulticlassclassificationusingdendriticneuronswithmorphologicallearning AT basuarindam improvedmarginmulticlassclassificationusingdendriticneuronswithmorphologicallearning |