Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2601.17625 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915753054175232 |
|---|---|
| author | Xie, Yuhan Liu, Jinhan Ni, Xiaoyong Tan, Fei Sakr, Icare Collin, Thibault Sun, Shiqi Guajardo, Alejandro Rodriguez Fanny, Demon Latchoumane, Charles-francois Vincent Lorach, Henri Bloch, Jocelyne Courtine, Gregoire Shoaran, Mahsa |
| author_facet | Xie, Yuhan Liu, Jinhan Ni, Xiaoyong Tan, Fei Sakr, Icare Collin, Thibault Sun, Shiqi Guajardo, Alejandro Rodriguez Fanny, Demon Latchoumane, Charles-francois Vincent Lorach, Henri Bloch, Jocelyne Courtine, Gregoire Shoaran, Mahsa |
| contents | Transformer-based neural decoders with large parameter counts, pre-trained on large-scale datasets, have recently outperformed classical machine learning models and small neural networks on brain-computer interface (BCI) tasks. However, their large parameter counts and high computational demands hinder deployment in power-constrained implantable systems. To address this challenge, we introduce BrainDistill, a novel implantable motor decoding pipeline that integrates an implantable neural decoder (IND) with a task-specific knowledge distillation (TSKD) framework. Unlike standard feature distillation methods that attempt to preserve teacher representations in full, TSKD explicitly prioritizes features critical for decoding through supervised projection. Across multiple neural datasets, IND consistently outperforms prior neural decoders on motor decoding tasks, while its TSKD-distilled variant further surpasses alternative distillation methods in few-shot calibration settings. Finally, we present a quantization-aware training scheme that enables integer-only inference with activation clipping ranges learned during training. The quantized IND enables deployment under the strict power constraints of implantable BCIs with minimal performance loss. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_17625 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | BrainDistill: Implantable Motor Decoding with Task-Specific Knowledge Distillation Xie, Yuhan Liu, Jinhan Ni, Xiaoyong Tan, Fei Sakr, Icare Collin, Thibault Sun, Shiqi Guajardo, Alejandro Rodriguez Fanny, Demon Latchoumane, Charles-francois Vincent Lorach, Henri Bloch, Jocelyne Courtine, Gregoire Shoaran, Mahsa Machine Learning Artificial Intelligence 68T07 Transformer-based neural decoders with large parameter counts, pre-trained on large-scale datasets, have recently outperformed classical machine learning models and small neural networks on brain-computer interface (BCI) tasks. However, their large parameter counts and high computational demands hinder deployment in power-constrained implantable systems. To address this challenge, we introduce BrainDistill, a novel implantable motor decoding pipeline that integrates an implantable neural decoder (IND) with a task-specific knowledge distillation (TSKD) framework. Unlike standard feature distillation methods that attempt to preserve teacher representations in full, TSKD explicitly prioritizes features critical for decoding through supervised projection. Across multiple neural datasets, IND consistently outperforms prior neural decoders on motor decoding tasks, while its TSKD-distilled variant further surpasses alternative distillation methods in few-shot calibration settings. Finally, we present a quantization-aware training scheme that enables integer-only inference with activation clipping ranges learned during training. The quantized IND enables deployment under the strict power constraints of implantable BCIs with minimal performance loss. |
| title | BrainDistill: Implantable Motor Decoding with Task-Specific Knowledge Distillation |
| topic | Machine Learning Artificial Intelligence 68T07 |
| url | https://arxiv.org/abs/2601.17625 |