Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2504.11981 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910912647004160 |
|---|---|
| author | Ikeda, Sosei Awano, Hiromitsu Sato, Takashi |
| author_facet | Ikeda, Sosei Awano, Hiromitsu Sato, Takashi |
| contents | Reservoir computing (RC) is attracting attention as a machine-learning technique for edge computing. In time-series classification tasks, the number of features obtained using a reservoir depends on the length of the input series. Therefore, the features must be converted to a constant-length intermediate representation (IR), such that they can be processed by an output layer. Existing conversion methods involve computationally expensive matrix inversion that significantly increases the circuit size and requires processing power when implemented in hardware. In this article, we propose a simple but effective IR, namely, dot-product-based reservoir representation (DPRR), for RC based on the dot product of data features. Additionally, we propose a hardware-friendly delayed-feedback reservoir (DFR) consisting of a nonlinear element and delayed feedback loop with DPRR. The proposed DFR successfully classified multivariate time series data that has been considered particularly difficult to implement efficiently in hardware. In contrast to conventional DFR models that require analog circuits, the proposed model can be implemented in a fully digital manner suitable for high-level syntheses. A comparison with existing machine-learning methods via field-programmable gate array implementation using 12 multivariate time-series classification tasks confirmed the superior accuracy and small circuit size of the proposed method. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_11981 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Hardware-Friendly Delayed-Feedback Reservoir for Multivariate Time-Series Classification Ikeda, Sosei Awano, Hiromitsu Sato, Takashi Machine Learning Hardware Architecture Reservoir computing (RC) is attracting attention as a machine-learning technique for edge computing. In time-series classification tasks, the number of features obtained using a reservoir depends on the length of the input series. Therefore, the features must be converted to a constant-length intermediate representation (IR), such that they can be processed by an output layer. Existing conversion methods involve computationally expensive matrix inversion that significantly increases the circuit size and requires processing power when implemented in hardware. In this article, we propose a simple but effective IR, namely, dot-product-based reservoir representation (DPRR), for RC based on the dot product of data features. Additionally, we propose a hardware-friendly delayed-feedback reservoir (DFR) consisting of a nonlinear element and delayed feedback loop with DPRR. The proposed DFR successfully classified multivariate time series data that has been considered particularly difficult to implement efficiently in hardware. In contrast to conventional DFR models that require analog circuits, the proposed model can be implemented in a fully digital manner suitable for high-level syntheses. A comparison with existing machine-learning methods via field-programmable gate array implementation using 12 multivariate time-series classification tasks confirmed the superior accuracy and small circuit size of the proposed method. |
| title | Hardware-Friendly Delayed-Feedback Reservoir for Multivariate Time-Series Classification |
| topic | Machine Learning Hardware Architecture |
| url | https://arxiv.org/abs/2504.11981 |