Technology Category
Machine Learning: Active LearningComputer Vision: Multi-modal VisionIntelligent Robots: Multimodal Perception & Sensor Fusion
Cross-Modal Consistency-Guided Active Learning for Affective BCI Systems
🤖 AI Summary
EEG-based emotion recognition suffers from severe label noise, pronounced inter-subject variability, and modality heterogeneity, leading to unstable neural decoding. Method: This paper proposes a cross-modal consistency-guided active learning framework. It introduces a deep representation alignment module to map EEG and facial expression signals into a shared latent space, coupled with a residual inconsistency detection mechanism that jointly models predictive uncertainty and cross-modal consistency—thereby distinguishing cognitive ambiguity from sensor-induced noise for more accurate sample querying. Results: Experiments on the ASCERTAIN dataset demonstrate that the method significantly reduces reliance on noisy labels: it achieves state-of-the-art performance even with 40% less annotation effort, confirming its data efficiency and robustness. Contribution: To our knowledge, this is the first work to integrate explicit cross-modal consistency modeling into active learning strategies, effectively mitigating emotion decoding bias under multi-source noise.
Technology Category
Application Category
📝 Abstract
Deep learning models perform best with abundant, high-quality labels, yet such conditions are rarely achievable in EEG-based emotion recognition. Electroencephalogram (EEG) signals are easily corrupted by artifacts and individual variability, while emotional labels often stem from subjective and inconsistent reports-making robust affective decoding particularly difficult. We propose an uncertainty-aware active learning framework that enhances robustness to label noise by jointly leveraging model uncertainty and cross-modal consistency. Instead of relying solely on EEG-based uncertainty estimates, the method evaluates cross-modal alignment to determine whether uncertainty originates from cognitive ambiguity or sensor noise. A representation alignment module embeds EEG and face features into a shared latent space, enforcing semantic coherence between modalities. Residual discrepancies are treated as noise-induced inconsistencies, and these samples are selectively queried for oracle feedback during active learning. This feedback-driven process guides the network toward reliable, informative samples and reduces the impact of noisy labels. Experiments on the ASCERTAIN dataset examine the efficiency and robustness of ours, highlighting its potential as a data-efficient and noise-tolerant approach for EEG-based affective decoding in brain-computer interface systems.
Problem
Research questions and friction points this paper is trying to address.
Addresses label noise and data scarcity in EEG-based emotion recognition systems
Leverages cross-modal consistency to distinguish cognitive ambiguity from sensor noise
Proposes active learning to improve robustness against subjective emotional labels
Innovation
Methods, ideas, or system contributions that make the work stand out.
Active learning framework leverages cross-modal consistency
Representation alignment embeds EEG and face features
Feedback-driven process reduces impact of noisy labels
🔎 Similar Papers
No similar papers found.
