The paper investigates the role of intermediate neural networks in end‑to‑end speech parsing, a task that jointly predicts transcription and syntactic structure. By removing these intermediate units, the authors reduce model parameters by 12% while matching or surpassing prior performance on ASR and parsing across French, Slovenian, and Naija. They find that intermediate layers mainly help close the representational gap when the pretrained encoder is frozen and analyze how training data size and encoder depth affect parsing performance.
arXiv:2509.17930v3 Announce Type: replace-cross Abstract: Multilingual translation suffers from computational redundancy, especially when translating into multiple languages simultaneously. In additi...
arXiv:2610.19388v1 Announce Type: new Abstract: Massively multilingual speech modeling has made remarkable progress in recent years, yet existing training resources and benchmarks still underrepresen...
The paper introduces phoneme‑guided initialization for large language model‑based speech recognition. By pre‑training an audio encoder on speech‑to‑phoneme conversion and an LLM on phoneme‑to‑grapheme conversion, then fine‑tuning jointly on ASR, the method leverages phoneme‑mediated processing within an end‑to‑end framework. Experiments on Japanese, Chinese, Tatar, and Urdu demonstrate that this approach matches or surpasses both cascaded S2P‑P2G baselines and end‑to‑end models without phoneme initialization.
The paper introduces a training‑free speech‑and‑text‑to‑pronunciation (ST2P) pipeline that combines lexical candidates from G2P tools with acoustic rescoring using frozen pretrained S2P models. By performing a left‑to‑right greedy search over whole‑sequence negative log‑likelihoods, the method achieves a dramatic reduction in character error rate on Japanese corpora, outperforming both baseline G2P/S2P approaches and commercial multimodal LLMs. The approach is also significantly faster—3–3.5× faster than beam search and twice as fast as direct decoding—while maintaining high accuracy across multiple languages.
arXiv:2603. 05121v2 Announce Type: replace-cross Abstract: Speech Large Language Models route speech encoder representations into an LLM decoder that typically accounts for over 90% of total parameters.
arXiv:2610.19492v1 Announce Type: new Abstract: Streaming Inverse Text Normalization (ITN) is vital for converting spoken-form outputs from streaming Automatic Speech Recognition into formatted writt...
arXiv:2610.08085v1 Announce Type: new Abstract: Speech-LLMs often exhibit prompt overfitting, where models solely trained on automatic speech recognition (ASR) instruction fail to generalize to new i...
arXiv:2608. 01281v1 Announce Type: cross Abstract: Phoneme-based multilingual automatic speech recognition (ASR) can share acoustic evidence across languages more directly than language-specific subword modeling.
Connecting a pre-trained speech encoder to a Large Language Model (LLM) is the standard architecture for building Speech LLMs. However, a structural misalignment exists between the encoder and the LLM.
The paper introduces a unified phoneme‑based TTS‑to‑ASR augmentation pipeline that uses a multilingual TTS model with language‑ID conditioning and incorporates grapheme‑to‑phoneme conversion, reference‑speech filtering, and candidate‑text selection. It proposes phoneme‑frequency‑guided selection (PFGS) to rank sentences based on phoneme frequencies from real ASR labels, and demonstrates that random augmentation and PFGS both improve ASR performance across Arabic, French, Italian, and Portuguese test sets, with PFGS yielding up to a 19.3% relative WER reduction. The study also shows that filtering reference speech can further lower WER by up to 0.59 points on certain datasets.
The paper presents a method for improving simultaneous speech translation by adapting a full‑utterance speech language model with prefix supervision derived from its own complete and partial waveform translations, eliminating the need for transcripts or human translations. Experiments on FLEURS and CoVoST2 across three language directions show that prefix training enhances quality–latency trade‑offs, especially when combined with multi‑turn append‑only decoding, and that a confidence threshold effectively controls the inference‑time quality–latency balance. The study also explores the impact of synthesis margin on translation quality and calibration, finding a non‑monotonic relationship with latency.
arXiv:2503. 05500v3 Announce Type: replace-cross Abstract: General-purpose multilingual vector representations, used in retrieval, regression and classification, are traditionally obtained from bidirectional encoder models.