Hierarchical Semantic Correlation-Aware Masked Autoencoder for Unsupervised Audio-Visual Representation Learning

TL;DR

HSC-MAE is a novel unsupervised audio-visual learning framework that enforces semantic consistency across multiple levels of representation, improving alignment and discriminative features without labeled data.

Contribution

It introduces a hierarchical, multi-level correlation-aware autoencoder framework combining global, local, and sample-level semantic constraints for better multimodal embedding alignment.

Findings

Significant mAP improvements on AVE and VEGAS datasets.

Robust audio-visual representations validated by experiments.

Effective multi-level semantic correlation enforcement.

Abstract

Learning aligned multimodal embeddings from weakly paired, label-free corpora is challenging: pipelines often provide only pre-extracted features, clips contain multiple events, and spurious co-occurrences. We propose HSC-MAE (Hierarchical Semantic Correlation-Aware Masked Autoencoder), a dual-path teacher-student framework that enforces semantic consistency across three complementary levels of representation - from coarse to fine: (i) global-level canonical-geometry correlation via DCCA, which aligns audio and visual embeddings within a shared modality-invariant subspace; (ii) local-level neighborhood-semantics correlation via teacher-mined soft top-k affinities, which preserves multi-positive relational structure among semantically similar instances; and (iii) sample-level conditional-sufficiency correlation via masked autoencoding, which ensures individual embeddings retain discriminative semantic content under partial observation. Concretely, a student MAE path is trained with masked feature reconstruction and affinity-weighted soft top-k InfoNCE; an EMA teacher operating on unmasked inputs via the CCA path supplies stable canonical geometry and soft positives. Learnable multi-task weights reconcile competing objectives, and an optional distillation loss transfers teacher geometry into the student. Experiments on AVE and VEGAS demonstrate substantial mAP improvements over strong unsupervised baselines, validating that HSC-MAE yields robust and well-structured audio-visual representations.