Salvato in:
Dettagli Bibliografici
Autori principali: Kankanala, Sai Samrat, Chandra, Ram, Ganapathy, Sriram
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2509.17965
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908874720673792
author Kankanala, Sai Samrat
Chandra, Ram
Ganapathy, Sriram
author_facet Kankanala, Sai Samrat
Chandra, Ram
Ganapathy, Sriram
contents Auditory attention and selective phase-locking are central to human speech understanding in complex acoustic scenes and cocktail party settings, yet these capabilities in multilingual subjects remain poorly understood. While machine understanding of natural speech has advanced in recent years, questions persist about comprehension of overlapped and mixed-channel speech. We propose a systematic paradigm for studying humans and machines in speech question-answering tasks in multilingual settings with clean and mixed-channel speech. For human listeners, selective attention to a target speaker was significantly better in their native language (L1) than in their second language (L2). For machine listening, speech-based large language models (LLMs) match or exceed human performance in clean, single-speaker conditions but often struggle to selectively attend in two-speaker settings. These results reveal a key divergence: humans rely on attentional cues that are more streamlined in their native language, whereas LLMs default to parallel information extraction which exceed human skills.
format Preprint
id arxiv_https___arxiv_org_abs_2509_17965
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Benchmarking Humans and Machines on Complex Multilingual Speech Understanding Tasks
Kankanala, Sai Samrat
Chandra, Ram
Ganapathy, Sriram
Audio and Speech Processing
Auditory attention and selective phase-locking are central to human speech understanding in complex acoustic scenes and cocktail party settings, yet these capabilities in multilingual subjects remain poorly understood. While machine understanding of natural speech has advanced in recent years, questions persist about comprehension of overlapped and mixed-channel speech. We propose a systematic paradigm for studying humans and machines in speech question-answering tasks in multilingual settings with clean and mixed-channel speech. For human listeners, selective attention to a target speaker was significantly better in their native language (L1) than in their second language (L2). For machine listening, speech-based large language models (LLMs) match or exceed human performance in clean, single-speaker conditions but often struggle to selectively attend in two-speaker settings. These results reveal a key divergence: humans rely on attentional cues that are more streamlined in their native language, whereas LLMs default to parallel information extraction which exceed human skills.
title Benchmarking Humans and Machines on Complex Multilingual Speech Understanding Tasks
topic Audio and Speech Processing
url https://arxiv.org/abs/2509.17965