[
    {
        "id": "osp-20494",
        "type": "article-journal",
        "title": "GLaS-JEPA: Gaussian-Regularized Speech SSL without Engineered Prediction Targets",
        "author": [
            {
                "family": "Botté",
                "given": "Gaspard"
            },
            {
                "family": "Baroudi",
                "given": "Séverin"
            },
            {
                "family": "Sadok",
                "given": "Samir"
            },
            {
                "family": "Paissan",
                "given": "Francesco"
            },
            {
                "family": "Hueber",
                "given": "Thomas"
            },
            {
                "family": "Alameda-Pineda",
                "given": "Xavier"
            },
            {
                "family": "Marxer",
                "given": "Ricard"
            },
            {
                "family": "Ravanelli",
                "given": "Mirco"
            }
        ],
        "URL": "https://omanscience.com/en/articles/glas-jepa-gaussian-regularized-speech-ssl-without-engineered-prediction-targets",
        "language": "en",
        "issued": {
            "date-parts": [
                [
                    2026
                ]
            ]
        },
        "abstract": "Speech self-supervised learning aims to learn general-purpose representations for downstream speech tasks. However, current approaches rely on complex, carefully designed prediction targets. We challenge this necessity with GLaS-JEPA, a framework that directly predicts the current encoder's continuous representations at masked positions, without contrastive learning, discrete targets, or separate EMA target encoders. We prevent representation collapse using SIGReg representation-space regularization, eliminating the need for engineered target-generation mechanisms. Pretrained on 960 hours of LibriSpeech, our 57M-parameter model achieves a 6.89% WER on frozen-encoder SUPERB ASR and a 25.87% CER on slot filling, outperforming the best non-distilled sub-90M baselines by 43.1% and 22.0%, respectively. These results demonstrate that highly competitive speech representations can emerge from a radically simplified training recipe."
    }
]