The paper compares encoder‑based and generative decoder‑based large language models for evaluating automatic speech recognition (ASR). It examines BERTScore and SemDist across various LLMs, layers, and pooling strategies, finding that both metrics can strongly correlate with human judgments when properly configured. For generative LLMs, the study explores pairwise hypothesis selection via prompting and direct error classification, showing that while encoder‑based metrics remain competitive, generative models excel in hypothesis comparison and enhance interpretability of ASR evaluation.
By Thibault Ba\~neras-Roux, Shashi Kumar, Driss Khalil, Sergio Burdisso, Petr Motlicek, Shiran Liu, Mickael Rouvier, Jane Wottawa, Richard Dufour
arXiv:2605. 20712v2 Announce Type: replace-cross Abstract: Automatic speech recognition replaces typing only when correction costs less than manual entry - a threshold determined by error types, not counts: fixing a misrecognized domain term costs far more than inserting a comma.
By Kavya Manohar, Arghya Bhattacharya, Kush Juvekar, Kumarmanas Nethil
EviSI is an evidence‑based evaluation agent for low‑latency simultaneous speech‑to‑speech translation. It combines Multidimensional Quality Metrics with interpreter‑developed criteria, using shared source evidence to assess four dimensions—Anchor, Event, Logic, and Fluency—while deduplicating verified errors before scoring. On English‑to‑Chinese and Chinese‑to‑English data, EviSI’s rankings correlate strongly with human judgments, outperforming BLEU and COMET, and its multilingual extension maintains these correlations across five language directions.
By Ben Yan, Zongyao Li, Xiaoyu Chen, Daimeng Wei, Weidong Liu, Huan Zhao, Chong Li, Yaode Wang, Yuzhe Shang
arXiv:2606. 17826v1 Announce Type: cross Abstract: Automatic speech recognition (ASR) in non-English clinical settings is challenged by multiscript variability, where the same term may appear in multiple valid orthographic forms.
By Jean Seo, Minkyu Kim, Jeonguk Lee, Jisoo Jung, Wooseok Han, Eunho Yang
EviSI is a large language model evaluation agent designed for simultaneous speech-to-speech translation. It adapts Multidimensional Quality Metrics to assess semantic fidelity and oral expression, using shared source evidence and deterministic scoring. In English‑to‑Chinese, EviSI achieves a mean Kendall agreement of 0.707 with human system rankings, outperforming baseline metrics, and shows positive concordance with COMET across five translation directions.
The paper introduces a three‑stage pipeline to improve accented conversational ASR for speakers from India, Indonesia, and Latin America. It uses heuristic SQL filters to curate entity‑rich training data, regional LoRA adapters fine‑tuned on Qwen2.5‑Omni‑3B to generate both verbatim and corrected transcripts, and a six‑category error taxonomy validated by an LLM judge. The approach raises entity recall to 80‑85% and filler recall to 76‑86%, while keeping WER low (6‑10%) and outperforming Whisper and a commercial ASR on entity recall.
By Fiza Husain, Ankit Pandey, Yash Singh
Automated detection of errors in clinical documentation is a promising application of large language models (LLMs), yet decisions to deploy such models rest on benchmarks that evaluate each clinical note in isolation. Error-detection benchmarks are typically constructed by injecting errors into notes, such that each erroneous note has a natural counterpart.
The paper investigates how to fairly compare language models across languages, noting that current evaluation methods vary widely and lack empirical validation. By training controlled monolingual models on parallel data and testing multilingual LLMs, the authors find that many normalized metrics suffer from biases due to tokenization, encoding, and orthographic differences. Instead, they recommend using sentence‑level negative log‑likelihood over semantically equivalent sequences for more reliable cross‑lingual comparisons.
By Xiulin Yang, Ethan Gotlieb Wilcox, Catherine Arnett
The paper introduces UPHELD, a large benchmark of human-to-human dialogues written by professional script writers, featuring realistic turn densities and over 36,000 per-turn human annotations. It evaluates existing automatic metrics and LLM-as-a-judge methods, finding them unreliable against expert human judgment. Using UPHELD, the authors develop a Mixture-of-Judges framework that improves correlation with human assessments by about 30%.
By Ilija Subasic, Andrew Rabinovich, Zhao Chen
arXiv:2608. 16643v1 Announce Type: cross Abstract: Automated detection of errors in clinical documentation is a promising application of large language models (LLMs), yet decisions to deploy such models rest on benchmarks that evaluate each clinical note in isolation.
By Yifan Zhang, Rahmatollah Beheshti
arXiv:2608. 09930v1 Announce Type: cross Abstract: Automated Text-to-Speech (TTS) evaluation methods (Mean Opinion Score (MOS) predictors and Audio Large Language Models (Audio-LLM) judges) are expected to reflect human perception, yet it is unclear how well they capture the distinct aspects of speech that listeners actually perceive.
By Oluwanifemi Bamgbose, Simon Rosen, Jash Shah, Lindsay Devon Brin, Hoang H Nguyen, Anke Koelzer, Rachel Hansen, Tara Bogavelli, Fanny Riols
The paper introduces a register-aware framework to evaluate how human-like large language models (LLMs) are, focusing on linguistic feature distributions rather than factual correctness. It uses Maximum Mean Discrepancy (MMD) and 67 Biber lexico‑grammatical features to compare LLM‑generated texts with human reference corpora across different registers. Experiments on seven instruction‑tuned, open‑source models across five English datasets show that all LLMs deviate from human baselines, with closeness to human language varying by register and not by model size.
By Bj\"orn Nieth, Marianna Gracheva, Michaela Mahlberg, Bjoern Eskofier, Emmanuelle Salin