# Breaking Shortcut Learning for Cross-Trial EEG-Guided Target Speech Extraction via Two-Stage Training

arXiv: https://arxiv.org/abs/2606.24164 · HTML: https://mlsanity.com/paper/2606.24164/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 34 references that are in the map.
- First posted on arXiv 23 Jun 2026.

## Builds on (most cited first, 11 of 11)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [SDR - half-baked or well done?](https://mlsanity.com/paper/1811.02508/) (2018)
- [Dual-path RNN: efficient long sequence modeling for time-domain single-channel speech separation](https://mlsanity.com/paper/1910.06379/) (2019)
- [EEG-based Auditory Attention Decoding: Towards Neuro-Steered Hearing Devices](https://mlsanity.com/paper/2008.04569/) (2020)
- [NeuroHeed: Neuro-Steered Speaker Extraction using EEG Signals](https://mlsanity.com/paper/2307.14303/) (2023)
- [NeuroHeed+: Improving Neuro-steered Speaker Extraction with Joint Auditory Attention Detection](https://mlsanity.com/paper/2312.07513/) (2023)
- [Beware of Overestimated Decoding Performance Arising from Temporal Autocorrelations in Electroencephalogram Signals](https://mlsanity.com/paper/2405.17024/) (2024)
- [BASEN: Time-Domain Brain-Assisted Speech Enhancement Network with Convolutional Cross Attention in Multi-talker Conditions](https://mlsanity.com/paper/2305.09994/) (2023)
- [Improved Feature Extraction Network for Neuro-Oriented Target Speaker Extraction](https://mlsanity.com/paper/2501.01673/) (2025)
- [M3ANet: Multi-scale and Multi-Modal Alignment Network for Brain-Assisted Target Speaker Extraction](https://mlsanity.com/paper/2506.00466/) (2025)
- [Geometry-Constrained EEG Channel Selection for Brain-Assisted Speech Enhancement](https://mlsanity.com/paper/2409.12520/) (2024)

## Abstract (arXiv)

Recent end-to-end models for EEG-guided target speech extraction report impressive results, underscoring potential for neuro-steered hearing technologies. However, our analysis reveals that high within-trial performance can be driven by trial-specific EEG structure that acts as shortcuts for target selection, leading to poor generalization on unseen trials. To overcome this gap, we propose TRUST-TSE, a two-stage framework to mitigate shortcut learning. By introducing contrastive pretraining with attended-speaker negative sampling, we encourage the EEG encoder to capture fine-grained EEG--speech alignment while suppressing trial-identity cues. We also employ a confidence-weighted extraction objective based on EEG--source similarity to guide extraction using the learned representations. Experiments on KUL and DTU datasets show that TRUST-TSE outperforms end-to-end baselines under strict cross-trial protocols, addressing a key reliability bottleneck of existing approaches.