Papers with LibriSpeech
Augmenting Librispeech with French Translations: A Multimodal Corpus for Direct Speech Translation Evaluation (L18-1)
Copied to clipboard
| Challenge: | Recent work in spoken language translation (SLT) has attempted to build end-to-end speech-totext translation without using source language transcription during learning or decoding. |
| Approach: | They propose to augment an existing (monolingual) corpus: LibriSpeech. |
| Outcome: | The proposed corpus is derived from read audiobooks from the LibriVox project and has been carefully segmented and aligned. |
Privacy Preserving Data Selection for Bias Mitigation in Speech Models (2025.acl-industry)
Copied to clipboard
Alkis Koudounas, Eliana Pastor, Vittorio Mazzia, Manuel Giollo, Thomas Gueudre, Elisa Reale, Luca Cagliero, Sandro Cumani, Luca De Alfaro, Elena Baralis, Daniele Amberti
| Challenge: | Existing methods for identifying subgroups raise privacy concerns and gather sensitive information at runtime might be impractical. |
| Approach: | They propose a method to identify and train underperforming subgroups and train a model to predict if an utterance belongs to these subgroup. |
| Outcome: | The proposed method reduces biases and improves performance on intent classification and automatic speech recognition tasks. |
SEAM: Bridging the Temporal-Semantic Granularity Gap for LLM-based Speech Recognition (2026.findings-eacl)
Copied to clipboard
| Challenge: | Existing duration-based methods generate embeddings at fixed rates, creating distributional mismatch with LLM pre-training. |
| Approach: | They propose an encoder-decoder architecture that generates embeddings at variable rates through cross-attention between speech features and text embeddables. |
| Outcome: | The proposed architecture achieves competitive performance on LibriSpeech (2.6%/5.2% WER) and 4.7% WER on TED-LIUM-v2 with a multi-stage training strategy and First Token Guidance. |
Align-Refine: Non-Autoregressive Speech Recognition via Iterative Realignment (2021.naacl-main)
Copied to clipboard
| Challenge: | Non-autoregressive encoder-decoder models improve decoding speed, but generation quality suffers . editing at the level of output sequences limits model flexibility. |
| Approach: | They propose *iterative realignment* which iteratively realigns connectionist temporal alignments. |
| Outcome: | The proposed model matches an autoregressive baseline with a 14x speedup on the WSJ dataset; on LibriSpeech, it achieves an LM-free test-other WER of 9.0% (19% relative improvement on comparable work). |
RECAST: Retrieval-Augmented Contextual ASR via Decoder-State Keyword Spotting (2025.findings-emnlp)
Copied to clipboard
| Challenge: | RECAST is a lightweight retrieval-augmented approach for contextual ASR . it repurposes decoder states of a pretrained ASR model to retrieve relevant keywords . |
| Approach: | RECAST is a retrieval-augmented approach that repurposes decoder states of a pretrained ASR model to retrieve relevant keywords without requiring audio exemplars. |
| Outcome: | RECAST outperforms full-list prompt biasing and strong phonetic/text baselines on 4,000 keywords across diverse domains. |
VoiceStar: Robust Zero-Shot Autoregressive TTS with Duration Control and Extrapolation (2026.findings-acl)
Copied to clipboard
| Challenge: | Neural codec language models (NCLMs) lack fine-grained controllability and inability to extrapolate to sequence lengths much longer than those seen during training. |
| Approach: | They propose a novel autoregressive encoder-decoder neural codec language model that can be trained with a Continuation-Prompt Mixed training system. |
| Outcome: | The proposed model outperforms or is on par with current state-of-the-art models on short-form benchmarks such as LibriSpeech and Seed-TTS in terms of intelligibility and naturalness. |