English

JEPA as a Neural Tokenizer: Learning Robust Speech Representations with Density Adaptive Attention

Sound 2025-12-09 v1 Artificial Intelligence Machine Learning Audio and Speech Processing

Abstract

We introduce a two-stage self-supervised framework that combines the Joint-Embedding Predictive Architecture (JEPA) with a Density Adaptive Attention Mechanism (DAAM) for learning robust speech representations. Stage~1 uses JEPA with DAAM to learn semantic audio features via masked prediction in latent space, fully decoupled from waveform reconstruction. Stage~2 leverages these representations for efficient tokenization using Finite Scalar Quantization (FSQ) and a mixed-radix packing scheme, followed by high-fidelity waveform reconstruction with a HiFi-GAN decoder. By integrating Gaussian mixture-based density-adaptive gating into the JEPA encoder, the model performs adaptive temporal feature selection and discovers hierarchical speech structure at a low frame rate of 2.5~Hz. The resulting tokens (47.5 tokens/sec) provide a reversible, highly compressed, and language-model-friendly representation that is competitive with, and often more efficient than, existing neural audio codecs.

Keywords

Cite

@article{arxiv.2512.07168,
  title  = {JEPA as a Neural Tokenizer: Learning Robust Speech Representations with Density Adaptive Attention},
  author = {Georgios Ioannides and Christos Constantinou and Aman Chadha and Aaron Elkins and Linsey Pang and Ravid Shwartz-Ziv and Yann LeCun},
  journal= {arXiv preprint arXiv:2512.07168},
  year   = {2025}
}

Comments

UniReps: Unifying Representations in Neural Models (NeurIPS 2025 Workshop)

R2 v1 2026-07-01T08:14:12.869Z