首页 > AI前沿 > When Should Deployed BCI Decoders Adapt? A Resource-Priced Ladder for Frozen EEG Trunks

When Should Deployed BCI Decoders Adapt? A Resource-Priced Ladder for Frozen EEG Trunks

arXiv机器学习 2026-11-19 08:34 14 阅读 查看原文
用户需要我将一段纯文本新闻正文排版成语义化的HTML片段。我需要仔细分析文本的结构,识别出段落、小标题、列表、引用、代码以及需要加粗的重要句子。 原文分析: 1. "When should a deployed MI decoder spend adaptation resources-not only how to adapt under shift?" - 这是一个段落 2. "We study a thin control layer outside a frozen encoder..." - 段落 3. "The head is Bayesian in the conjugate sense..." - 段落 4. "Full estimators, constants, and the reanalysis code are in the companion paper and the ancillary bundle of this submission." - 段落 5. "Thesis: adaptation should be gated and capped, not always-on and not blindly fine-tuned." - 重要句子,可以用strong 6. "Evidence: (i) utility-max gating over-adapts..." - 证据列表 7. "(i) utility-max gating over-adapts on clean data..." - 列表项 8. "(ii) Under severe synthetic stress..." - 列表项 9. "(iii) In two no-perturbation BNCI session checks..." - 列表项 10. "(iv) Adaptation severity does not follow a universal ordinal ladder..." - 列表项 现在构建HTML:

When should a deployed MI decoder spend adaptation resources-not only how to adapt under shift?

We study a thin control layer outside a frozen encoder: a resource ladder (factory -> Bayesian head -> affine map, with full fine-tune as an escape hatch) governed by regime-aware gating under a utility that prices accuracy, fit time, and label cost.

The head is Bayesian in the conjugate sense: a Normal-Inverse-Wishart prior over the class conditional Gaussians in embedding space, updated in closed form (no gradients) as labelled trials arrive.

Full estimators, constants, and the reanalysis code are in the companion paper and the ancillary bundle of this submission.

Thesis: adaptation should be gated and capped, not always-on and not blindly fine-tuned.

Evidence:

  • utility-max gating over-adapts on clean data that already carries session drift; a clean-referenced gate restores low false-entry rates (e.g. dropping the adaptation rate from 0.56 to 0.11 on field-like sessions, with the cascade staying on the factory decoder for ~87% of clean cells).
  • Under severe synthetic stress, a lightweight LDA head matches or exceeds the frozen trunk's accuracy across five encoder architectures (n=155 subject x trunk cells, pooled across shared subjects), at ~23x lower fit time than full warm-start fine-tuning. Always-on baselines favor capping the default below full fine-tune.
  • In two no-perturbation BNCI session checks the factory model transferred without intervention (no reliable head gain), so the control layer's value under natural drift is cost-efficiency and clean-safety rather than demonstrated accuracy gains.
  • Adaptation severity does not follow a universal ordinal ladder across datasets, highlighting the need for dynamic gating over fixed adaptation policies.
检查一下规则: 1. 只输出HTML片段本身,不要Markdown围栏、不要解释 - ✓ 2. 每个自然段落用

包裹,长文用

/

划分小节 - 原文是论文摘要风格的文本,没有明显的章节划分,用

即可。不需要强行添加小标题。 3. 列表内容用

  • - ✓,引用用
    ,代码用
    ,重要句子可用 - 用strong标注了Thesis和Evidence
    4. 保留原文所有信息,不得增删改内容 - ✓
    5. 不要输出//等文档级标签 - ✓
    
    再次检查原文,确保内容完整:
    
    原文开头:"When should a deployed MI decoder spend adaptation resources-not only how to adapt under shift?"
    这是第一个自然段,应该是一个问题引入。
    
    第二段:"We study a thin control layer outside a frozen encoder: a resource ladder (factory -> Bayesian head -> affine map, with full fine-tune as an escape hatch) governed by regime-aware gating under a utility that prices accuracy, fit time, and label cost."
    
    第三段:"The head is Bayesian in the conjugate sense: a Normal-Inverse-Wishart prior over the class conditional Gaussians in embedding space, updated in closed form (no gradients) as labelled trials arrive."
    
    第四段:"Full estimators, constants, and the reanalysis code are in the companion paper and the ancillary bundle of this submission."
    
    第五段:"Thesis: adaptation should be gated and capped, not always-on and not blindly fine-tuned."
    这应该用标注,因为"Thesis"是核心论点。
    
    第六段:"Evidence: (i) utility-max gating over-adapts on clean data that already carries session drift; a clean-referenced gate restores low false-entry rates (e.g. dropping the adaptation rate from 0.56 to 0.11 on field-like sessions, with the cascade staying on the factory decoder for ~87% of clean cells). (ii) Under severe synthetic stress, a lightweight LDA head matches or exceeds the frozen trunk's accuracy across five encoder architectures (n=155 subject x trunk cells, pooled across shared subjects), at ~23x lower fit time than full warm-start fine-tuning. Always-on baselines favor capping the default below full fine-tune. (iii) In two no-perturbation BNCI session checks the factory model transferred without intervention (no reliable head gain), so the control layer's value under natural drift is cost-efficiency and clean-safety rather than demonstrated accuracy gains. (iv) Adaptation severity does not follow a universal ordinal ladder across datasets, highlighting the need for dynamic gating over fixed adaptation policies."
    
    Evidence后面是四个带编号的证据,应该作为列表处理。可以将"Evidence:"单独作为一段或用strong标注,然后四个(i)-(iv)作为
    • 列表。 但是原文中"Evidence: (i) ... (ii) ... (iii) ... (iv) ..."是连在一起的。为了保持原文信息完整,我可以将"Evidence:"作为引导句,然后列表项包含原文的完整内容。 不过要"保留原文所有信息,不得增删改内容"。如果我把Evidence后面的内容拆分成列表,需要保留原文的(i)(ii)(iii)(iv)编号吗?原文中确实有这些编号。 让我重新考虑:原文的Evidence部分是一个长段落,包含四个编号的要点。用户要求"列表内容用
      • "。这四个要点明显是列表内容。 我可以这样做:

        Evidence:

        • (i) utility-max gating over-adapts on clean data that already carries session drift; a clean-referenced gate restores low false-entry rates (e.g. dropping the adaptation rate from 0.56 to 0.11 on field-like sessions, with the cascade staying on the factory decoder for ~87% of clean cells).
        • (ii) Under severe synthetic stress, a lightweight LDA head matches or exceeds the frozen trunk's accuracy across five encoder architectures (n=155 subject x trunk cells, pooled across shared subjects), at ~23x lower fit time than full warm-start fine-tuning. Always-on baselines favor capping the default below full fine-tune.
        • (iii) In two no-perturbation BNCI session checks the factory model transferred without intervention (no reliable head gain), so the control layer's value under natural drift is cost-efficiency and clean-safety rather than demonstrated accuracy gains.
        • (iv) Adaptation severity does not follow a universal ordinal ladder across datasets, highlighting the need for dynamic gating over fixed adaptation policies.
        这样保留了原文的(i)(ii)(iii)(iv)编号,同时使用了列表。但用户要求"不得增删改内容",如果原文是"Evidence: (i) ... (ii) ...",我