AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Bayesian Test-Time Adaptation via Dirichlet feature projection and GMM-Driven Inference for Motor Imagery EEG Decoding

Huan LUO, Na Lu, Xiaopeng Wang, Xu Niu

ICLR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Realistic and stable test-time adaptation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Dirichlet feature projection (with EA) (ablation)SincAdaptNet · BNCI2014001 cross-sessionaccuracy (%): 82.33 -> 84.04Improved
add GMM inference (with EA) (ablation)SincAdaptNet · BNCI2014001 cross-sessionaccuracy (%): 82.33 -> 82.47No clear effect
add GMM inference on top of Dirichlet projection (ablation)SincAdaptNet + EA + Dirichlet · BNCI2014001 cross-sessionaccuracy (%): 84.04 -> 86.50Improved
add EA to BTTA-DG (ablation)BTTA-DG · BNCI2014001 cross-sessionaccuracy (%): 81.88 -> 86.50Improved
add full BTTA-DG adaptation to source model (ablation)SincAdaptNet · BNCI2014001 cross-subjectaccuracy (%): 75.30 -> 78.70Improved
add Tent entropy-minimization adaptationSincAdaptNet · BNCI2014001 cross-subjectaccuracy (%): 77.03 -> 70.23Hurt
add pseudo-label (PL) adaptationSincAdaptNet · BNCI2014001 cross-subjectaccuracy (%): 77.03 -> 71.83Hurt
use BTTA-DG instead of OTTASincAdaptNet · BNCI2014001 cross-subjectaccuracy (%): 77.58 -> 78.70Improved
increase online class imbalance from 1:1 to 1:0.25BTTA-DG · BNCI2014001 cross-subjectoverall accuracy (%): 78.70 -> 68.77Hurt
increase online class imbalance from 1:1 to 1:0.25BTTA-DG · BNCI2014001 cross-subjectminority class 1 accuracy (%): 80.40 -> 85.19Improved
vary number of GMM components K in [2, 12]BTTA-DG · BNCI2014001 cross-subjectaccuracy (%): 78.2%–78.7%No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy