Work

BiMamba2 Masked Discrete-Unit Prediction for Multilingual Speech Representation for Unsupervised Speech in the Wild Challenge

Year: 2026

Type: conference-paper

Abstract: We describe our submission to the Unsupervised Speech in the Wild (UPS) Challenge at Interspeech 2026, a bidirectional Mamba-2 (BiMamba2) encoder trained with masked discreteunit prediction following the HuBERT-style paradigm.The 47.88M-parameter model is trained on 250 hours of speech across 67 languages from the MLCommons Unsupervised People's Speech dataset, with no labeled data.The objective combines masked k-means pseudo-label prediction with language identification supervision and VICReg regularization.On official evaluation, the system achieves an Adjusted Rand Index of 0.735, exceeding four baselines on speaker clustering.Language identification macro-F1 (0.073) and character error rate (0.870) remain below supervised baselines.We analyze a local-official discrepancy in metric scale and checkpoint ranking, highlighting limitations of in-distribution diagnostics for predicting Dynabench probe outcomes.

Source: Interspeech (USB)

DOI: 10.21437/interspeech.2026-2966

Authors: Prakriti Subedi, Howard Prioleau, Saurav Aryal

Institution: Howard University

Language: English

Topic: Speech Recognition and Synthesis

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: bronze

Funders: U.S. Navy, Office of Naval Research, Common Fund

Award: G1729356799