Efficient Solutions for Mitigating Initialization Bias in Unsupervised Self-Adaptive Auditory Attention Decoding

2025-09-18

AI Summary

In unsupervised adaptive auditory attention decoding (AAD), initialization bias degrades decoding performance, while existing bias-correcting methods suffer from computational complexity that scales superlinearly with data volume. To address this, we propose three low-overhead, constant-complexity adaptive algorithms grounded in an enhanced stimulus reconstruction framework. Integrating unsupervised learning with adaptive signal processing, our approach enables robust electroencephalographic (EEG) attention tracking without labeled data. Crucially, the algorithms eliminate initialization bias while maintaining fixed per-iteration computational cost—unlike state-of-the-art (SOTA) methods, whose complexity grows linearly or worse with time-series length. Evaluated in multi-speaker scenarios, our methods achieve decoding accuracy comparable to SOTA while improving inference efficiency by an order of magnitude. The implementation is fully open-sourced, ensuring reproducibility.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Unsupervised & Self-Supervised Learning
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Decoding the attended speaker in a multi-speaker environment from electroencephalography (EEG) has attracted growing interest in recent years, with neuro-steered hearing devices as a driver application. Current approaches typically rely on ground-truth labels of the attended speaker during training, necessitating calibration sessions for each user and each EEG set-up to achieve optimal performance. While unsupervised self-adaptive auditory attention decoding (AAD) for stimulus reconstruction has been developed to eliminate the need for labeled data, it suffers from an initialization bias that can compromise performance. Although an unbiased variant has been proposed to address this limitation, it introduces substantial computational complexity that scales with data size. This paper presents three computationally efficient alternatives that achieve comparable performance, but with a significantly lower and constant computational cost. The code for the proposed algorithms is available at https://github.com/YYao-42/Unsupervised_AAD.

Problem

• Mitigating initialization bias in unsupervised self-adaptive auditory attention decoding
• Reducing computational complexity in EEG-based speaker attention decoding
• Eliminating need for labeled calibration data in neuro-steered hearing devices

Innovation

• Unsupervised self-adaptive auditory attention decoding
• Computationally efficient bias mitigation alternatives
• Constant low computational cost performance