Interspeech
Collection
Accepted papers for Interspeech (Annual Conference of the International Speech Communication Association), one dataset per year. • 14 items • Updated
paper_id stringlengths 15 35 | title stringlengths 33 190 | authors listlengths 1 20 | isca_url stringlengths 66 86 | pdf_url stringlengths 65 85 | doi stringlengths 28 30 | pages stringlengths 3 9 | bibtex large_stringlengths 294 724 | abstract large_stringlengths 555 1.51k | arxiv_id stringlengths 10 10 ⌀ | arxiv_id_source stringclasses 2
values |
|---|---|---|---|---|---|---|---|---|---|---|
barreiros26_interspeech | Massive Open-Vocabulary Keyword Spotting | [
"Leonor Barreiros",
"Raul Monteiro",
"Afonso Mendes",
"Gonçalo M. Correia"
] | https://www.isca-archive.org/interspeech_2026/barreiros26_interspeech.html | https://www.isca-archive.org/interspeech_2026/barreiros26_interspeech.pdf | 10.21437/Interspeech.2026-1444 | 1-5 | @inproceedings{barreiros26_interspeech,
title = {{Massive Open-Vocabulary Keyword Spotting}},
author = {Leonor Barreiros and Raul Monteiro and Afonso Mendes and Gonçalo M. Correia},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {1--5},
doi = {10.21437/Interspeech.2026-1444},... | Automatic speech recognition systems have been shown to under-perform when it comes to transcribing words rarely seen in the training data, namely specialized terminology. Open-vocabulary keyword spotting, combined with contextual biasing, has been shown to mitigate this issue. However, existing systems can only handle... | 2606.11279 | title_snapshot |
baek26_interspeech | SPARK: Efficient Audio-Text Matching for User-Defined Keyword Spotting via Spiking Neural Networks | [
"Seung-Yeop Baek",
"Sangho Han",
"Joon-Hyuk Chang"
] | https://www.isca-archive.org/interspeech_2026/baek26_interspeech.html | https://www.isca-archive.org/interspeech_2026/baek26_interspeech.pdf | 10.21437/Interspeech.2026-3336 | 6-10 | @inproceedings{baek26_interspeech,
title = {{SPARK: Efficient Audio-Text Matching for User-Defined Keyword Spotting via Spiking Neural Networks}},
author = {Seung-Yeop Baek and Sangho Han and Joon-Hyuk Chang},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {6--10},
doi = {10.... | With the increasing prevalence of voice-driven interaction, keyword spotting (KWS) has become an essential component of hands-free control. This has led to a growing demand for user-defined KWS, allowing users to customize target keywords via text. While various models have emerged to support this, their high computati... | null | null |
khaymonenko26_interspeech | Scalable Keyword Spotting via Modular Network Expansion | [
"Viktor Khaymonenko",
"Dzmitry Saladukha",
"Aliaksei Rak",
"Alexander Rostov"
] | https://www.isca-archive.org/interspeech_2026/khaymonenko26_interspeech.html | https://www.isca-archive.org/interspeech_2026/khaymonenko26_interspeech.pdf | 10.21437/Interspeech.2026-987 | 11-15 | @inproceedings{khaymonenko26_interspeech,
title = {{Scalable Keyword Spotting via Modular Network Expansion}},
author = {Viktor Khaymonenko and Dzmitry Saladukha and Aliaksei Rak and Alexander Rostov},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {11--15},
doi = {10.21437/I... | Keyword spotting (KWS) models on embedded devices often need to add new keywords after deployment, but updates are difficult when original training data are unavailable and regressions on existing triggers are unacceptable. At a fixed operating point, our method reduces average new-keyword false reject rate (FRR) from ... | 2607.19918 | title_snapshot |
chen26q_interspeech | Streaming Open-Vocabulary Keyword Spotting via Role Swapping in Cross-Attention | [
"Xi Chen",
"Haichuan Bai",
"Liming Song"
] | https://www.isca-archive.org/interspeech_2026/chen26q_interspeech.html | https://www.isca-archive.org/interspeech_2026/chen26q_interspeech.pdf | 10.21437/Interspeech.2026-1676 | 16-20 | @inproceedings{chen26q_interspeech,
title = {{Streaming Open-Vocabulary Keyword Spotting via Role Swapping in Cross-Attention}},
author = {Xi Chen and Haichuan Bai and Liming Song},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {16--20},
doi = {10.21437/Interspeech.2026-1676... | In recent years, attention-based multi-modal open-vocabulary keyword spotting (KWS) has attracted attention, yet its streaming deployment on device with such a design remains underexplored. We identify a data flow mismatch in traditional frameworks: speech as Key/Value requires global context but, in streaming, provide... | null | null |
zhang26fa_interspeech | MPA-KWS: Multi-Modal Phoneme-Level Alignment for Streaming Open-Vocabulary Keyword Spotting | [
"Jue Zhang",
"Guibin Zheng",
"Jiarui Zhang",
"Jiqing Han",
"Chenhao Jing"
] | https://www.isca-archive.org/interspeech_2026/zhang26fa_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhang26fa_interspeech.pdf | 10.21437/Interspeech.2026-2485 | 21-25 | @inproceedings{zhang26fa_interspeech,
title = {{MPA-KWS: Multi-Modal Phoneme-Level Alignment for Streaming Open-Vocabulary Keyword Spotting}},
author = {Jue Zhang and Guibin Zheng and Jiarui Zhang and Jiqing Han and Chenhao Jing},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {21--2... | For open-vocabulary keyword spotting, phoneme-level alignment has been introduced to improve the performance on acoustically confusable words. However, most existing studies utilize non-streaming methods, which are unsuitable for streaming scenarios. Recent methods achieve streaming phoneme alignment via connectionist ... | null | null |
laquatra26b_interspeech | SSL-based Sequence Matching for Unsupervised Audio Retrieval | [
"Moreno La Quatra",
"Alkis Koudounas",
"Sabato Marco Siniscalchi"
] | https://www.isca-archive.org/interspeech_2026/laquatra26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/laquatra26b_interspeech.pdf | 10.21437/Interspeech.2026-2369 | 26-31 | @inproceedings{laquatra26b_interspeech,
title = {{SSL-based Sequence Matching for Unsupervised Audio Retrieval}},
author = {Moreno {La Quatra} and Alkis Koudounas and Sabato Marco Siniscalchi},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {26--31},
doi = {10.21437/Interspee... | This paper investigates audio-to-audio retrieval using self-supervised learning (SSL) models to generate audio representations without labeled data. To enhance retrieval accuracy, we explore the use of SSL embeddings with sequence matching techniques, including Dynamic Time Warping (DTW), and clustering methods, such a... | null | null |
loweimi26_interspeech | To Be Multimodal or Not to Be: Query-Adaptive Audio-Visual Person Retrieval via Active Modality Detection | [
"Erfan Loweimi",
"Mengjie Qian",
"Kate Knill",
"Guanfeng Wu",
"Chi-Ho Chan",
"Abbas Haider",
"Muhammad Awan",
"Josef Kittler",
"Hui Wang",
"Mark Gales"
] | https://www.isca-archive.org/interspeech_2026/loweimi26_interspeech.html | https://www.isca-archive.org/interspeech_2026/loweimi26_interspeech.pdf | 10.21437/Interspeech.2026-790 | 32-36 | @inproceedings{loweimi26_interspeech,
title = {{To Be Multimodal or Not to Be: Query-Adaptive Audio-Visual Person Retrieval via Active Modality Detection}},
author = {Erfan Loweimi and Mengjie Qian and Kate Knill and Guanfeng Wu and Chi-Ho Chan and Abbas Haider and Muhammad Awan and Josef Kittler and Hui Wan... | When retrieving a person from a video archive by voice and face, should the system be multimodal or not? In real-world broadcast archives, unlike curated benchmarks, a target may be heard but unseen, seen but unheard, or both. Fusing scores from an absent modality injects noise, degrading precision below the best unimo... | 2606.05931 | title_snapshot |
jeon26_interspeech | Disentangling Depression from Cognitive Decline in Elderly Speech Using Concurrent Clinical Assessments | [
"Woori Jeon",
"Seunghee Ha",
"Sang-Kyu Lee",
"Ji Hye Yoon",
"Tae-Jin Yoon",
"Seung Jin Lee",
"Woojae Han",
"Jungmin So"
] | https://www.isca-archive.org/interspeech_2026/jeon26_interspeech.html | https://www.isca-archive.org/interspeech_2026/jeon26_interspeech.pdf | 10.21437/Interspeech.2026-1247 | 37-42 | @inproceedings{jeon26_interspeech,
title = {{Disentangling Depression from Cognitive Decline in Elderly Speech Using Concurrent Clinical Assessments}},
author = {Woori Jeon and Seunghee Ha and Sang-Kyu Lee and Ji Hye Yoon and Tae-Jin Yoon and Seung Jin Lee and Woojae Han and Jungmin So},
year = {2026}... | Detecting depression from speech in elderly patients with mild cognitive impairment (MCI) is complicated by overlapping acoustic effects of cognitive decline. Without disentangling these, classifiers risk learning cognitive rather than depression-specific patterns. We present a Korean elderly speech corpus of 89 MCI sp... | null | null |
haghbin26_interspeech | From Black-Box to Clinical Insight: A Multi-Stage Explainable Framework for Speech-Based Cognitive Impairment Detection | [
"Yasaman Haghbin",
"Sina Rashidi",
"Ali Zolnour",
"Fatemeh Taherinezhad",
"Ali Fartoot",
"Hossein Azadmaleki",
"James M. Noble",
"Maryam Dadkhah",
"Maryam Zolnoori"
] | https://www.isca-archive.org/interspeech_2026/haghbin26_interspeech.html | https://www.isca-archive.org/interspeech_2026/haghbin26_interspeech.pdf | 10.21437/Interspeech.2026-1252 | 43-47 | @inproceedings{haghbin26_interspeech,
title = {{From Black-Box to Clinical Insight: A Multi-Stage Explainable Framework for Speech-Based Cognitive Impairment Detection}},
author = {Yasaman Haghbin and Sina Rashidi and Ali Zolnour and Fatemeh Taherinezhad and Ali Fartoot and Hossein Azadmaleki and James M. No... | Speech-based cognitive impairment detection offers a noninvasive, accessible alternative to costly biomarker assays, yet transformer-based models remain clinically uninterpretable. We propose a multi-stage explainability framework that translates black-box transformer predictions into clinically grounded narratives by ... | 2606.27973 | title_snapshot |
gonzalezmachorro26_interspeech | Towards Speech Impairment Prediction in German-Speaking Individuals with Amyotrophic Lateral Sclerosis | [
"Monica Gonzalez-Machorro",
"Ricarda von Heynitz",
"Justin Hanslmeier",
"Finja Grimm",
"Alexandra-Iulia Deac",
"Anne Gründel",
"Isabell Cordts",
"Bjoern Schuller"
] | https://www.isca-archive.org/interspeech_2026/gonzalezmachorro26_interspeech.html | https://www.isca-archive.org/interspeech_2026/gonzalezmachorro26_interspeech.pdf | 10.21437/Interspeech.2026-1052 | 48-53 | @inproceedings{gonzalezmachorro26_interspeech,
title = {{Towards Speech Impairment Prediction in German-Speaking Individuals with Amyotrophic Lateral Sclerosis}},
author = {Monica Gonzalez-Machorro and Ricarda von Heynitz and Justin Hanslmeier and Finja Grimm and Alexandra-Iulia Deac and Anne Gründel and Isa... | Amyotrophic Lateral Sclerosis (ALS) is a neurodegenerative disease, often affecting speech due to bulbar dysfunction. In this study, we predict speech impairment in people with ALS (pwALS) using two clinical speech-related scores. We evaluate cross-sectional (across speakers) and personalised (within-speaker) modelling... | 2606.17616 | title_snapshot |
xu26n_interspeech | Automatic Graphical Representations of Language for Dementia Detection | [
"Lingfeng Xu",
"Si-Ioi Ng",
"Pranav S. Ambadi",
"Fan Lei",
"Kimberly D. Mueller",
"Julie Liss",
"Visar Berisha"
] | https://www.isca-archive.org/interspeech_2026/xu26n_interspeech.html | https://www.isca-archive.org/interspeech_2026/xu26n_interspeech.pdf | 10.21437/Interspeech.2026-1233 | 54-59 | @inproceedings{xu26n_interspeech,
title = {{Automatic Graphical Representations of Language for Dementia Detection}},
author = {Lingfeng Xu and Si-Ioi Ng and Pranav S. Ambadi and Fan Lei and Kimberly D. Mueller and Julie Liss and Visar Berisha},
year = {2026},
booktitle = {{Interspeech 2026}},
pag... | The Cookie Theft picture description task is widely used to assess cognitive–linguistic abilities. Analyzing how speakers progress through the 23 content information units (CIUs) in the picture provides insight into the informational relevance and efficiency of their descriptions. Although prior CIU-based studies have ... | null | null |
haghbin26b_interspeech | Natural Speech Encodes Early Markers of Cognitive Decline: Evidence from Clinical Conversations | [
"Yasaman Haghbin",
"Sina Rashidi",
"Ali Zolnour",
"Margaret McDonald",
"Maryam Zolnoori"
] | https://www.isca-archive.org/interspeech_2026/haghbin26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/haghbin26b_interspeech.pdf | 10.21437/Interspeech.2026-1860 | 60-65 | @inproceedings{haghbin26b_interspeech,
title = {{Natural Speech Encodes Early Markers of Cognitive Decline: Evidence from Clinical Conversations}},
author = {Yasaman Haghbin and Sina Rashidi and Ali Zolnour and Margaret McDonald and Maryam Zolnoori},
year = {2026},
booktitle = {{Interspeech 2026}},
... | Alzheimer's disease and related dementia (ADRD) remain largely undiagnosed, as early cognitive symptoms are rarely captured by structured clinical data. Natural speech offers a sensitive, non-invasive window into cognitive decline that structured clinical data cannot provide. This study validates conversational speech ... | null | null |
kothare26_interspeech | Speech-based Digital Biomarkers can Accelerate ALS Clinical Trials: Insights from Time-to-Event and Hazard Rate Analysis | [
"Hardik Kothare",
"Michael Neumann",
"Vikram Ramanarayanan"
] | https://www.isca-archive.org/interspeech_2026/kothare26_interspeech.html | https://www.isca-archive.org/interspeech_2026/kothare26_interspeech.pdf | 10.21437/Interspeech.2026-2847 | 66-71 | @inproceedings{kothare26_interspeech,
title = {{Speech-based Digital Biomarkers can Accelerate ALS Clinical Trials: Insights from Time-to-Event and Hazard Rate Analysis}},
author = {Hardik Kothare and Michael Neumann and Vikram Ramanarayanan},
year = {2026},
booktitle = {{Interspeech 2026}},
pages... | Early detection of functional decline in amyotrophic lateral sclerosis (ALS) is critical for timely intervention and efficient clinical trial design. We evaluated speech-derived biomarkers to detect ALS-related functional decline events earlier than traditional clinical measures. Using longitudinal patient data, we app... | null | null |
zafar26_interspeech | Rethinking Acoustic Variability Of ADReSS and ADReSSo Datasets For Dementia Detection | [
"Muhammad Abdullah Zafar",
"Mostafa Shahin",
"Beena Ahmed"
] | https://www.isca-archive.org/interspeech_2026/zafar26_interspeech.html | https://www.isca-archive.org/interspeech_2026/zafar26_interspeech.pdf | 10.21437/Interspeech.2026-2862 | 72-76 | @inproceedings{zafar26_interspeech,
title = {{Rethinking Acoustic Variability Of ADReSS and ADReSSo Datasets For Dementia Detection}},
author = {Muhammad Abdullah Zafar and Mostafa Shahin and Beena Ahmed},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {72--76},
doi = {10.214... | The ADReSS and ADReSSo challenge datasets have become the de-facto standards for research on dementia detection through speech, with over half of recent ICASSP and Interspeech studies on the topic relying on them. Despite their widespread adoption, both datasets exhibit properties that may undermine the validity of rep... | null | null |
kopar26_interspeech | Beyond Binary: Speech Representations Across the Cognitive Score Hierarchy | [
"Serli Kopar",
"Roshan P. Rane",
"Christian Mychajliw",
"Lydia Federmann",
"Gerhard Eschweiler",
"Daniela Berg",
"Sam Gijsen",
"Paula Andrea Pérez-Toro",
"Kerstin Ritter"
] | https://www.isca-archive.org/interspeech_2026/kopar26_interspeech.html | https://www.isca-archive.org/interspeech_2026/kopar26_interspeech.pdf | 10.21437/Interspeech.2026-2725 | 77-81 | @inproceedings{kopar26_interspeech,
title = {{Beyond Binary: Speech Representations Across the Cognitive Score Hierarchy}},
author = {Serli Kopar and Roshan P. Rane and Christian Mychajliw and Lydia Federmann and Gerhard Eschweiler and Daniela Berg and Sam Gijsen and Paula Andrea Pérez-Toro and Kerstin Ritte... | This study examines the relationship between speech representations and the hierarchical structure of cognitive assessment in mild cognitive impairment. Utilizing 5,754 German neuropsychological assessment recordings, we evaluate six cognitive tasks across three score levels: task, domain, and global levels. We compare... | 2605.27189 | title_snapshot |
liu26p_interspeech | audiobook-cc: Controllable Long-context Speech Generation for Multicast Audiobook | [
"Min Liu",
"JingJing Yin",
"Xiang Zhang",
"JianHao Ye",
"Siyu Hao",
"Siwei Xia",
"Hongbin Zhou"
] | https://www.isca-archive.org/interspeech_2026/liu26p_interspeech.html | https://www.isca-archive.org/interspeech_2026/liu26p_interspeech.pdf | 10.21437/Interspeech.2026-2125 | 82-86 | @inproceedings{liu26p_interspeech,
title = {{audiobook-cc: Controllable Long-context Speech Generation for Multicast Audiobook}},
author = {Min Liu and JingJing Yin and Xiang Zhang and JianHao Ye and Siyu Hao and Siwei Xia and Hongbin Zhou},
year = {2026},
booktitle = {{Interspeech 2026}},
pages ... | Existing text-to-speech systems predominantly focus on single-sentence synthesis and lack adequate contextual modeling and fine-grained control for coherent multicast audiobooks. To address this, we propose a context-aware, emotion-controllable speech synthesis framework with three innovations: a context mechanism for ... | 2509.17516 | title_snapshot |
ghosh26f_interspeech | MagpieTTS-LF: Inference-Time Long-Form Speech Generation Without Training on Long-Form data | [
"Subhankar Ghosh",
"Jason Li",
"Paarth Neekhara",
"Shehzeen Hussain",
"Ryan Langman",
"Xuesong Yang",
"Roy Fejgin"
] | https://www.isca-archive.org/interspeech_2026/ghosh26f_interspeech.html | https://www.isca-archive.org/interspeech_2026/ghosh26f_interspeech.pdf | 10.21437/Interspeech.2026-1461 | 87-91 | @inproceedings{ghosh26f_interspeech,
title = {{MagpieTTS-LF: Inference-Time Long-Form Speech Generation Without Training on Long-Form data}},
author = {Subhankar Ghosh and Jason Li and Paarth Neekhara and Shehzeen Hussain and Ryan Langman and Xuesong Yang and Roy Fejgin},
year = {2026},
booktitle = ... | Neural Text-to-Speech (TTS) systems achieve remarkable quality on short utterances but long-form speech generation shows prosodic drift, speaker inconsistencies and sentence boundary artifacts. Existing approaches either compress sequences, increase context length or naively concatenate independently synthesized chunks... | 2606.18485 | title_snapshot |
ren26d_interspeech | AuDirector: A Self-Reflective Closed-Loop Framework for Immersive Audio Storytelling | [
"Yiming Ren",
"Ziyang Zhang",
"Wen Wu",
"Baoxiang Li",
"Chao Zhang",
"Xuenan Xu"
] | https://www.isca-archive.org/interspeech_2026/ren26d_interspeech.html | https://www.isca-archive.org/interspeech_2026/ren26d_interspeech.pdf | 10.21437/Interspeech.2026-1180 | 92-96 | @inproceedings{ren26d_interspeech,
title = {{AuDirector: A Self-Reflective Closed-Loop Framework for Immersive Audio Storytelling}},
author = {Yiming Ren and Ziyang Zhang and Wen Wu and Baoxiang Li and Chao Zhang and Xuenan Xu},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {92--96}... | Despite progress in text and visual generation, coherent long-form audio storytelling remains challenging. Existing systems often suffer from mismatches between character settings and voice performance, weak self-correction, and limited user interaction. We propose AuDirector, a self-reflective closed-loop multi-agent ... | 2605.11866 | title_snapshot |
lee26i_interspeech | Designed Vocalizations Dataset: Sound-Designed Human and Animal Voices for Non-human Voice Conversion | [
"Seolhee Lee",
"Minsu Kang",
"Yangsun Lee",
"Woosun Min",
"Choonghyeon Lee",
"Namhyun Cho"
] | https://www.isca-archive.org/interspeech_2026/lee26i_interspeech.html | https://www.isca-archive.org/interspeech_2026/lee26i_interspeech.pdf | 10.21437/Interspeech.2026-932 | 97-101 | @inproceedings{lee26i_interspeech,
title = {{Designed Vocalizations Dataset: Sound-Designed Human and Animal Voices for Non-human Voice Conversion}},
author = {Seolhee Lee and Minsu Kang and Yangsun Lee and Woosun Min and Choonghyeon Lee and Namhyun Cho},
year = {2026},
booktitle = {{Interspeech 202... | Advances in AI-based voice conversion have enabled a wide range of media applications, including films, audiobooks, and games. However, most research and public benchmarks still focus on natural human speech, leaving designed vocalizations, such as monster growls and robotic voices, underexplored, partly due to the lac... | 2607.20951 | title_snapshot |
kim26_interspeech | ZipL-Dialog: Memory-Efficient Long-Form Spoken Dialog Synthesis via Latent Flow Matching | [
"Jihwan Kim",
"Nam Soo Kim"
] | https://www.isca-archive.org/interspeech_2026/kim26_interspeech.html | https://www.isca-archive.org/interspeech_2026/kim26_interspeech.pdf | 10.21437/Interspeech.2026-185 | 102-106 | @inproceedings{kim26_interspeech,
title = {{ZipL-Dialog: Memory-Efficient Long-Form Spoken Dialog Synthesis via Latent Flow Matching}},
author = {Jihwan Kim and Nam Soo Kim},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {102--106},
doi = {10.21437/Interspeech.2026-185},
i... | Zero-shot dialog TTS benefits from flow-matching, but minute-scale generation on dense mel-spectrograms causes severe memory bottlenecks, often forcing unnatural chunked synthesis. We propose ZipL-Dialog, which shifts conditional flow-matching into a 4x time-compressed (25 Hz) latent space. To preserve acoustic fidelit... | 2607.12496 | title_snapshot |
wang26ea_interspeech | Not Flat, But Dissociated: Prosodic and Segmental Divergence in Neural TTS | [
"Rong Wang",
"Kun Sun",
"Harald Baayen"
] | https://www.isca-archive.org/interspeech_2026/wang26ea_interspeech.html | https://www.isca-archive.org/interspeech_2026/wang26ea_interspeech.pdf | 10.21437/Interspeech.2026-2730 | 107-111 | @inproceedings{wang26ea_interspeech,
title = {{Not Flat, But Dissociated: Prosodic and Segmental Divergence in Neural TTS}},
author = {Rong Wang and Kun Sun and Harald Baayen},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {107--111},
doi = {10.21437/Interspeech.2026-2730},
... | Standard TTS metrics such as MOS and mel-cepstral distortion provide global scores but do not locate where synthesis diverges from natural speech. To examine this, four systems (Tacotron2-DDC, FastSpeech2, Glow-TTS, MixerTTS) are analysed across 13,100 matched LJ-TTS utterances. At the prosodic level, global F0 variabi... | null | null |
ye26b_interspeech | Refining Emphasis Control in Flow-Matching TTS via Preference Alignment and Reinforcement Learning | [
"Jiangnan Ye",
"Jiawei Jin",
"Pengfei Tan",
"Chao Yan",
"Xuerui Yang",
"Zhiyong Wu"
] | https://www.isca-archive.org/interspeech_2026/ye26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/ye26b_interspeech.pdf | 10.21437/Interspeech.2026-2284 | 112-116 | @inproceedings{ye26b_interspeech,
title = {{Refining Emphasis Control in Flow-Matching TTS via Preference Alignment and Reinforcement Learning}},
author = {Jiangnan Ye and Jiawei Jin and Pengfei Tan and Chao Yan and Xuerui Yang and Zhiyong Wu},
year = {2026},
booktitle = {{Interspeech 2026}},
page... | Achieving fine-grained emphasis control in speech synthesis remains challenging due to data scarcity and the inherent complexity of prosody. To address this issue, we extend F5-TTS with an additional Emphasis Encoder and propose a three-stage optimization framework that progressively enhances emphasis controllability. ... | null | null |
yang26l_interspeech | CraftTTS: Fine-Grained Prosody Control for Text-to-Speech | [
"Wenbing Yang",
"Qihang Lu",
"Bingsong Bai",
"Zihan Sun",
"Yueran Hou",
"Peilei Jia",
"Yingming Gao",
"Ya Li",
"Jun Gao"
] | https://www.isca-archive.org/interspeech_2026/yang26l_interspeech.html | https://www.isca-archive.org/interspeech_2026/yang26l_interspeech.pdf | 10.21437/Interspeech.2026-2018 | 117-121 | @inproceedings{yang26l_interspeech,
title = {{CraftTTS: Fine-Grained Prosody Control for Text-to-Speech}},
author = {Wenbing Yang and Qihang Lu and Bingsong Bai and Zihan Sun and Yueran Hou and Peilei Jia and Yingming Gao and Ya Li and Jun Gao},
year = {2026},
booktitle = {{Interspeech 2026}},
pag... | While zero-shot text-to-speech models perform well in global voice cloning, they struggle with fine-grained prosodic control, as strict word-level intensity and tempo manipulation can disrupt acoustic priors and introduce artifacts. We propose CraftTTS, a three-stage framework enabling stable word-level control without... | null | null |
mou26b_interspeech | Dynamic Prosody Prediction in LLM-based TTS for Improving Speaker Similarity | [
"Zhenwei Mou",
"Liping Chen",
"Yajun Hu",
"Zhen-Hua Ling",
"Xin Fang",
"Jian-Qing Gao"
] | https://www.isca-archive.org/interspeech_2026/mou26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/mou26b_interspeech.pdf | 10.21437/Interspeech.2026-2312 | 122-126 | @inproceedings{mou26b_interspeech,
title = {{Dynamic Prosody Prediction in LLM-based TTS for Improving Speaker Similarity}},
author = {Zhenwei Mou and Liping Chen and Yajun Hu and Zhen-Hua Ling and Xin Fang and Jian-Qing Gao},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {122--126}... | Personalized text-to-speech (TTS) aims to clone the target speaker in the synthesized speech, imitating both the voice and speaking style. Current large language model (LLM)-based TTS methods ignore the style-specific prosodic patterns in generated speech, resulting in deficient style learning and thus limiting speaker... | 2606.15267 | title_snapshot |
dong26_interspeech | Membership Inference Attacks against Large Audio Language Models | [
"Jia-Kai Dong",
"Yu-Xiang Lin",
"Hung-yi Lee"
] | https://www.isca-archive.org/interspeech_2026/dong26_interspeech.html | https://www.isca-archive.org/interspeech_2026/dong26_interspeech.pdf | 10.21437/Interspeech.2026-514 | 127-132 | @inproceedings{dong26_interspeech,
title = {{Membership Inference Attacks against Large Audio Language Models}},
author = {Jia-Kai Dong and Yu-Xiang Lin and Hung-yi Lee},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {127--132},
doi = {10.21437/Interspeech.2026-514},
issn ... | We present the first systematic membership inference attack (MIA) evaluation of LALMs. Using Multi-modal Blind Baselines based on textual, spectral and prosodic features, we demonstrate that common audio datasets exhibit near-perfect train/test separability (AUC ≈ 1.0) even without model inference, thus MIA may primari... | 2603.28378 | title_snapshot |
saini26_interspeech | Listening Like a Judge: A Music-Aware Framework for Automatic Singing Performance Evaluation | [
"Neelam Saini",
"Sourav Ghosh"
] | https://www.isca-archive.org/interspeech_2026/saini26_interspeech.html | https://www.isca-archive.org/interspeech_2026/saini26_interspeech.pdf | 10.21437/Interspeech.2026-912 | 133-137 | @inproceedings{saini26_interspeech,
title = {{Listening Like a Judge: A Music-Aware Framework for Automatic Singing Performance Evaluation}},
author = {Neelam Saini and Sourav Ghosh},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {133--137},
doi = {10.21437/Interspeech.2026-... | Automatic singing quality assessment (SQA) requires evaluating lyrical correctness and musical fidelity while handling expressive variations. However, existing systems largely rely on either acoustic cues or lyric transcriptions exclusively, limiting holistic performance evaluation. Furthermore, their integration is no... | 2606.26451 | title_snapshot |
suzuki26_interspeech | ELSA: Acoustic Event-Level Semantic Alignment for Fine-Grained Reference-Free Text-to-Audio Evaluation | [
"Shuntaro Suzuki",
"Kento Tokura",
"Daichi Yashima",
"Kanon Amemiya",
"Komei Sugiura",
"Shinnosuke Takamichi"
] | https://www.isca-archive.org/interspeech_2026/suzuki26_interspeech.html | https://www.isca-archive.org/interspeech_2026/suzuki26_interspeech.pdf | 10.21437/Interspeech.2026-914 | 138-142 | @inproceedings{suzuki26_interspeech,
title = {{ELSA: Acoustic Event-Level Semantic Alignment for Fine-Grained Reference-Free Text-to-Audio Evaluation}},
author = {Shuntaro Suzuki and Kento Tokura and Daichi Yashima and Kanon Amemiya and Komei Sugiura and Shinnosuke Takamichi},
year = {2026},
booktit... | Text-to-audio (TTA) generation, synthesizing audio from natural language, has been widely studied for its ability to capture precise user intent. To effectively advance TTA models, it is essential to reliably evaluate generated audio without relying on costly human subjective ratings, motivating the development of auto... | 2606.17404 | title_snapshot |
fan26_interspeech | PrefSQA: Pairwise Preference Prediction for Speech Quality Assessment and the Critical Role of High Quality Datasets | [
"Junyi Fan",
"Donald S. Williamson"
] | https://www.isca-archive.org/interspeech_2026/fan26_interspeech.html | https://www.isca-archive.org/interspeech_2026/fan26_interspeech.pdf | 10.21437/Interspeech.2026-1512 | 143-147 | @inproceedings{fan26_interspeech,
title = {{PrefSQA: Pairwise Preference Prediction for Speech Quality Assessment and the Critical Role of High Quality Datasets}},
author = {Junyi Fan and Donald S. Williamson},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {143--147},
doi = ... | Mean opinion scores (MOS) are widely used for speech quality assessment, yet scalar labels are sensitive to rater variability and listening test differences. This introduces labeling noise, which limits the reliability of MOS prediction. Preference prediction reduces this variability as listeners compare signals direct... | 2606.19597 | title_snapshot |
zhou26c_interspeech | UG-Bench: A Comprehensive Benchmark for Evaluating Large Audio-Language Models | [
"Jiaming Zhou",
"Haoqin Sun",
"Hui Wang",
"Jinghua Zhao",
"Yuhang Jia",
"Shiyao Wang",
"Enzhi Wang",
"Shiwan Zhao",
"Yong Qin"
] | https://www.isca-archive.org/interspeech_2026/zhou26c_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhou26c_interspeech.pdf | 10.21437/Interspeech.2026-1517 | 148-153 | @inproceedings{zhou26c_interspeech,
title = {{UG-Bench: A Comprehensive Benchmark for Evaluating Large Audio-Language Models}},
author = {Jiaming Zhou and Haoqin Sun and Hui Wang and Jinghua Zhao and Yuhang Jia and Shiyao Wang and Enzhi Wang and Shiwan Zhao and Yong Qin},
year = {2026},
booktitle = ... | Evaluating Large Audio-Language Models (LALMs) is challenging due to the diverse speech and audio tasks. Existing benchmarks often lack a unified framework, focusing either on understanding or generation. We introduce UG-Bench, a comprehensive benchmark systematically assessing LALMs across four competencies: speech pe... | null | null |
sultana26_interspeech | A Fine-Grained Acoustically-Aware Pre-training Encoder for Speech Quality Assessment | [
"Subrina Sultana",
"Donald S. Williamson"
] | https://www.isca-archive.org/interspeech_2026/sultana26_interspeech.html | https://www.isca-archive.org/interspeech_2026/sultana26_interspeech.pdf | 10.21437/Interspeech.2026-1607 | 154-158 | @inproceedings{sultana26_interspeech,
title = {{A Fine-Grained Acoustically-Aware Pre-training Encoder for Speech Quality Assessment}},
author = {Subrina Sultana and Donald S. Williamson},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {154--158},
doi = {10.21437/Interspeech.... | Self-supervised learning (SSL) has become popular in speech processing because it generalizes well across downstream tasks. However, many SSL methods focus on capturing long-term contextual and speaker information, often making their representations invariant to background acoustics, which is an issue for speech qualit... | null | null |
zhang26x_interspeech | VoxEffects: A Speech-Oriented Audio Effects Dataset and Benchmark | [
"Zhe Zhang",
"Yigitcan Özer",
"Junichi Yamagishi"
] | https://www.isca-archive.org/interspeech_2026/zhang26x_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhang26x_interspeech.pdf | 10.21437/Interspeech.2026-1621 | 159-163 | @inproceedings{zhang26x_interspeech,
title = {{VoxEffects: A Speech-Oriented Audio Effects Dataset and Benchmark}},
author = {Zhe Zhang and Yigitcan Özer and Junichi Yamagishi},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {159--163},
doi = {10.21437/Interspeech.2026-1621},... | Speech audio in the wild is often processed by post-production effects, but existing speech datasets rarely provide precise annotations of effects and parameters, limiting systematic study. We introduce VoxEffects, a speech audio effects dataset that pairs produced speech with exact effect-chain supervision at multiple... | 2604.12389 | title_snapshot |
lanzendoerfer26_interspeech | Evaluating Objective Speech Quality Metrics for Neural Audio Codecs | [
"Luca A. Lanzendöerfer",
"Florian Grötschla",
"Roger Wattenhofer"
] | https://www.isca-archive.org/interspeech_2026/lanzendoerfer26_interspeech.html | https://www.isca-archive.org/interspeech_2026/lanzendoerfer26_interspeech.pdf | 10.21437/Interspeech.2026-1809 | 164-168 | @inproceedings{lanzendoerfer26_interspeech,
title = {{Evaluating Objective Speech Quality Metrics for Neural Audio Codecs}},
author = {Luca A. Lanzendöerfer and Florian Grötschla and Roger Wattenhofer},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {164--168},
doi = {10.2143... | Neural audio codecs have gained recent popularity for their use in generative modeling as they offer high-fidelity audio reconstruction at low bitrates. While human listening studies remain the gold standard for assessing perceptual quality, they are time-consuming and impractical. In this work, we examine the reliabil... | 2511.19734 | title_snapshot |
kostenok26_interspeech | Calibration-Reasoning Framework for Descriptive Speech Quality Assessment | [
"Elizaveta Kostenok",
"Mathieu Salzmann",
"Milos Cernak"
] | https://www.isca-archive.org/interspeech_2026/kostenok26_interspeech.html | https://www.isca-archive.org/interspeech_2026/kostenok26_interspeech.pdf | 10.21437/Interspeech.2026-2362 | 169-173 | @inproceedings{kostenok26_interspeech,
title = {{Calibration-Reasoning Framework for Descriptive Speech Quality Assessment}},
author = {Elizaveta Kostenok and Mathieu Salzmann and Milos Cernak},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {169--173},
doi = {10.21437/Inters... | Explainable speech quality assessment requires moving beyond Mean Opinion Scores (MOS) to analyze underlying perceptual dimensions. To address this, we introduce a novel post-training method that tailors the foundational Audio Large Language Model for multidimensional reasoning, detection and classification of audio ar... | 2603.10175 | title_snapshot |
ferreira26_interspeech | CAL-MOS: Bridging Layers with Adapters for Robust MOS Prediction Across Speech Foundation Models | [
"Alef Iury Ferreira",
"Pedro Botelho",
"Fernanda Silva",
"Daniel Casanova",
"Rafael Faustino",
"Frederico Oliveira",
"Arlindo Galvão Filho",
"Anderson da Silva Soares"
] | https://www.isca-archive.org/interspeech_2026/ferreira26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ferreira26_interspeech.pdf | 10.21437/Interspeech.2026-2960 | 174-179 | @inproceedings{ferreira26_interspeech,
title = {{CAL-MOS: Bridging Layers with Adapters for Robust MOS Prediction Across Speech Foundation Models}},
author = {Alef Iury Ferreira and Pedro Botelho and Fernanda Silva and Daniel Casanova and Rafael Faustino and Frederico Oliveira and Arlindo Galvão Filho and An... | Speech Quality Assessment (SQA) is essential for modern speech technologies, and recent non-intrusive SQA predictors increasingly rely on Speech Foundation Models (SFMs). However, because SFMs expose representations from many layers, it remains unclear which depths are most informative for MOS prediction and how multi-... | 2609.14956 | title_snapshot |
park26h_interspeech | AnimeScore: A Preference-Based Dataset and Framework for Evaluating Anime-Like Speech Style | [
"Joonyong Park",
"Jerry Li"
] | https://www.isca-archive.org/interspeech_2026/park26h_interspeech.html | https://www.isca-archive.org/interspeech_2026/park26h_interspeech.pdf | 10.21437/Interspeech.2026-3025 | 180-184 | @inproceedings{park26h_interspeech,
title = {{AnimeScore: A Preference-Based Dataset and Framework for Evaluating Anime-Like Speech Style}},
author = {Joonyong Park and Jerry Li},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {180--184},
doi = {10.21437/Interspeech.2026-3025... | Evaluating 'anime-like' voices currently relies on costly subjective judgments, yet no standardized objective metric exists. A key challenge is that anime-likeness, unlike naturalness, lacks a shared absolute scale, making conventional Mean Opinion Score (MOS) protocols unreliable. To address this gap, we propose Anime... | 2603.11482 | title_snapshot |
bhattacharya26b_interspeech | Exploiting Neural Audio Codec Latents for Adversarial Audio Attacks | [
"Sameek Bhattacharya",
"Bharath Krishnamurthy",
"Ajita Rattani"
] | https://www.isca-archive.org/interspeech_2026/bhattacharya26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/bhattacharya26b_interspeech.pdf | 10.21437/Interspeech.2026-3055 | 185-189 | @inproceedings{bhattacharya26b_interspeech,
title = {{Exploiting Neural Audio Codec Latents for Adversarial Audio Attacks}},
author = {Sameek Bhattacharya and Bharath Krishnamurthy and Ajita Rattani},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {185--189},
doi = {10.21437/... | Deep learning–based audio classification systems, including automatic speaker verification, are vulnerable to adversarial attacks. Realistic real-time threat assessment remains difficult because optimization-based methods, such as projected gradient descent (PGD) and Carlini–Wagner, require costly iterative updates in ... | 2606.20893 | title_snapshot |
dixit26_interspeech | AURA Score: A Metric for Holistic Audio Question Answering Evaluation | [
"Satvik Dixit",
"Soham Deshmukh",
"Bhiksha Raj"
] | https://www.isca-archive.org/interspeech_2026/dixit26_interspeech.html | https://www.isca-archive.org/interspeech_2026/dixit26_interspeech.pdf | 10.21437/Interspeech.2026-3185 | 190-194 | @inproceedings{dixit26_interspeech,
title = {{AURA Score: A Metric for Holistic Audio Question Answering Evaluation}},
author = {Satvik Dixit and Soham Deshmukh and Bhiksha Raj},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {190--194},
doi = {10.21437/Interspeech.2026-3185}... | Audio Question Answering (AQA) is a key task for evaluating Audio-Language Models (ALMs), yet assessing open-ended responses remains challenging. Existing metrics used for AQA such as BLEU, METEOR and BERTScore, mostly adapted from NLP and audio captioning, rely on surface similarity and fail to account for question co... | 2510.04934 | title_snapshot |
hegde26_interspeech | Aligning Audio Captions with Human Preferences | [
"Kartik Hegde",
"Rehana Mahfuz",
"Yinyi Guo",
"Erik Visser"
] | https://www.isca-archive.org/interspeech_2026/hegde26_interspeech.html | https://www.isca-archive.org/interspeech_2026/hegde26_interspeech.pdf | 10.21437/Interspeech.2026-2052 | 195-199 | @inproceedings{hegde26_interspeech,
title = {{Aligning Audio Captions with Human Preferences}},
author = {Kartik Hegde and Rehana Mahfuz and Yinyi Guo and Erik Visser},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {195--199},
doi = {10.21437/Interspeech.2026-2052},
issn ... | Current audio captioning relies on supervised learning with paired audio-caption data, which is costly to curate and may not reflect human preferences in real-world scenarios. To address this, we propose a preference-aligned audio captioning framework based on Reinforcement Learning from Human Feedback (RLHF). To captu... | 2509.14659 | title_snapshot |
chang26_interspeech | TAD: Token-Adaptive Contrastive Decoding with Confidence-Guided Gating for Hallucination Mitigation in Large Audio-Language Models | [
"Heyu Chang",
"Nianwen Si",
"Hao Zhang",
"Wenlin Zhang",
"Dan Qu"
] | https://www.isca-archive.org/interspeech_2026/chang26_interspeech.html | https://www.isca-archive.org/interspeech_2026/chang26_interspeech.pdf | 10.21437/Interspeech.2026-637 | 200-205 | @inproceedings{chang26_interspeech,
title = {{TAD: Token-Adaptive Contrastive Decoding with Confidence-Guided Gating for Hallucination Mitigation in Large Audio-Language Models}},
author = {Heyu Chang and Nianwen Si and Hao Zhang and Wenlin Zhang and Dan Qu},
year = {2026},
booktitle = {{Interspeech... | Large audio-language models (LALMs) can hallucinate audio objects, answering "yes" to absent sound events, thus undermining reliability in audio question answering. We propose Token-Adaptive Decoding (TAD), a training-free strategy for hallucination mitigation that grounds the initial yes/no decision by contrasting log... | 2609.07286 | title_snapshot |
ly26_interspeech | TinyGiantALM: A Compact Audio-Language Model for Intent-Aware Reasoning under Resource Constraints | [
"Vinh-Thuan Ly"
] | https://www.isca-archive.org/interspeech_2026/ly26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ly26_interspeech.pdf | 10.21437/Interspeech.2026-491 | 206-210 | @inproceedings{ly26_interspeech,
title = {{TinyGiantALM: A Compact Audio-Language Model for Intent-Aware Reasoning under Resource Constraints}},
author = {Vinh-Thuan Ly},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {206--210},
doi = {10.21437/Interspeech.2026-491},
issn ... | Current advancements in Audio Reasoning rely on massive Large Audio-Language Models (LALMs), hindering deployment in resource-constrained environments. We introduce Tiny-GiantALM, a compact 1.5B efficiency-oriented alternative. Instead of brute-force scaling, we propose an Instruction-Aware Feature Refinement framework... | 2606.08425 | title_snapshot |
rong26b_interspeech | Beyond Symmetric Interaction: Capability-Aware Asymmetric Multi-Agent Collaboration for Audio Deep Reasoning | [
"Yan Rong",
"Jinting Wang",
"Tianxin Xie",
"Xiang He",
"Chenxing Li",
"Dong Yu",
"Li Liu"
] | https://www.isca-archive.org/interspeech_2026/rong26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/rong26b_interspeech.pdf | 10.21437/Interspeech.2026-2273 | 211-216 | @inproceedings{rong26b_interspeech,
title = {{Beyond Symmetric Interaction: Capability-Aware Asymmetric Multi-Agent Collaboration for Audio Deep Reasoning}},
author = {Yan Rong and Jinting Wang and Tianxin Xie and Xiang He and Chenxing Li and Dong Yu and Li Liu},
year = {2026},
booktitle = {{Intersp... | Audio deep reasoning demands expert-level perception and multi-step reasoning. Vanilla multi-agent paradigms struggle here due to three challenges: (1) underutilization of the complementary abilities of diverse Large Audio-Language Models (LALMs); (2) neglect of role differentiation and capability bias; and (3) samplin... | null | null |
tu26b_interspeech | VISA: A Visual Information Strengthened Audio-Reasoning System for the Interspeech 2026 ARC Agent Track | [
"Wenming Tu",
"Jian Gao",
"Yanru Huo",
"Yixuan Wang",
"Jing Peng",
"Bohan Li",
"Ziyang Ma",
"Tao Liu",
"Shuai Fan",
"Kai Yu",
"Xie Chen",
"Zilong Zheng"
] | https://www.isca-archive.org/interspeech_2026/tu26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/tu26b_interspeech.pdf | 10.21437/Interspeech.2026-2381 | 217-221 | @inproceedings{tu26b_interspeech,
title = {{VISA: A Visual Information Strengthened Audio-Reasoning System for the Interspeech 2026 ARC Agent Track}},
author = {Wenming Tu and Jian Gao and Yanru Huo and Yixuan Wang and Jing Peng and Bohan Li and Ziyang Ma and Tao Liu and Shuai Fan and Kai Yu and Xie Chen and... | Audio reasoning requires multi-step, evidence-grounded inference over temporally dynamic and acoustically mixed signals, exceeding conventional perception tasks such as ASR or captioning. We present VISA, our submission to the Interspeech 2026 Audio Reasoning Challenge (Agent Track), evaluated via the MMAR Rubrics for ... | 2606.07264 | title_snapshot |
ma26_interspeech | The Interspeech 2026 Audio Reasoning Challenge: Evaluating Reasoning Process Quality for Audio Reasoning Models and Agents | [
"Ziyang Ma",
"Ruiyang Xu",
"Yinghao Ma",
"Chao-Han Huck Yang",
"Bohan Li",
"Jaeyeon Kim",
"Jin Xu",
"Jinyu Li",
"Carlos Busso",
"Kai Yu",
"Eng Siong Chng",
"Xie Chen"
] | https://www.isca-archive.org/interspeech_2026/ma26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ma26_interspeech.pdf | 10.21437/Interspeech.2026-118 | 222-227 | @inproceedings{ma26_interspeech,
title = {{The Interspeech 2026 Audio Reasoning Challenge: Evaluating Reasoning Process Quality for Audio Reasoning Models and Agents}},
author = {Ziyang Ma and Ruiyang Xu and Yinghao Ma and Chao-Han Huck Yang and Bohan Li and Jaeyeon Kim and Jin Xu and Jinyu Li and Carlos Bus... | Recent Large Audio Language Models (LALMs) excel in understanding but often lack transparent reasoning. To address this "black-box" limitation, we organized the Audio Reasoning Challenge at Interspeech 2026, the first shared task dedicated to evaluating Chain-of-Thought (CoT) quality in the audio domain. The challenge ... | 2602.14224 | title_snapshot |
olev26_interspeech | Multi-Source Evidence Fusion for Audio Question Answering | [
"Aivo Olev",
"Tanel Alumäe"
] | https://www.isca-archive.org/interspeech_2026/olev26_interspeech.html | https://www.isca-archive.org/interspeech_2026/olev26_interspeech.pdf | 10.21437/Interspeech.2026-3297 | 228-233 | @inproceedings{olev26_interspeech,
title = {{Multi-Source Evidence Fusion for Audio Question Answering}},
author = {Aivo Olev and Tanel Alumäe},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {228--233},
doi = {10.21437/Interspeech.2026-3297},
issn = {2958-1796},
} | Large audio language models (LALMs) can answer questions about speech, music, and environmental sounds, yet their internal reasoning is largely opaque and difficult to validate. We describe TalTech's solution to the Agent Track of the Interspeech 2026 Audio Reasoning Challenge, in which systems are evaluated on reasoni... | 2603.17822 | title_snapshot |
noronha26_interspeech | Structured Prompting vs. Self-Training for Audio Reasoning Under Limited Data and Compute: Lessons from Interspeech Audio Reasoning Challenge 2026 | [
"Sujit Noronha",
"Steven Au",
"Kaushlendra Tripathi"
] | https://www.isca-archive.org/interspeech_2026/noronha26_interspeech.html | https://www.isca-archive.org/interspeech_2026/noronha26_interspeech.pdf | 10.21437/Interspeech.2026-2880 | 234-238 | @inproceedings{noronha26_interspeech,
title = {{Structured Prompting vs. Self-Training for Audio Reasoning Under Limited Data and Compute: Lessons from Interspeech Audio Reasoning Challenge 2026}},
author = {Sujit Noronha and Steven Au and Kaushlendra Tripathi},
year = {2026},
booktitle = {{Interspe... | This study consists of our approaches to the Interspeech 2026 Audio Reasoning challenge evaluating chain-of-thought-based reasoning traces on the Multi-Modal Audio Reasoning benchmark (MMAR). For audio reasoning with limited data and compute, practitioners and researchers must choose among various strategies ranging fr... | null | null |
li26o_interspeech | Audio-Cogito: Towards Deep Audio Reasoning in Large Audio Language Models | [
"Longhao Li",
"Hongjie Chen",
"Zehan Li",
"Qihan Hu",
"Jian Kang",
"Jie Li",
"Lei Xie",
"Yongxiang Li"
] | https://www.isca-archive.org/interspeech_2026/li26o_interspeech.html | https://www.isca-archive.org/interspeech_2026/li26o_interspeech.pdf | 10.21437/Interspeech.2026-988 | 239-244 | @inproceedings{li26o_interspeech,
title = {{Audio-Cogito: Towards Deep Audio Reasoning in Large Audio Language Models}},
author = {Longhao Li and Hongjie Chen and Zehan Li and Qihan Hu and Jian Kang and Jie Li and Lei Xie and Yongxiang Li},
year = {2026},
booktitle = {{Interspeech 2026}},
pages ... | Recent advances in reasoning models have driven significant progress in text and multimodal domains, yet audio reasoning remains relatively limited. Only a few Large Audio Language Models (LALMs) incorporate explicit Chain-of-Thought (CoT) reasoning, and their capabilities are often inconsistent and insufficient for co... | 2604.12527 | title_snapshot |
he26e_interspeech | Audio-DeepThinker: Progressive Reasoning-Aware Reinforcement Learning for High-Quality Chain-of-Thought Emergence in Audio Language Models | [
"Xiang He",
"Chenxing Li",
"Jinting Wang",
"Yan Rong",
"Tianxin Xie",
"Zeyu Xie",
"Wenfu Wang",
"Li Liu",
"Dong Yu"
] | https://www.isca-archive.org/interspeech_2026/he26e_interspeech.html | https://www.isca-archive.org/interspeech_2026/he26e_interspeech.pdf | 10.21437/Interspeech.2026-1720 | 245-250 | @inproceedings{he26e_interspeech,
title = {{Audio-DeepThinker: Progressive Reasoning-Aware Reinforcement Learning for High-Quality Chain-of-Thought Emergence in Audio Language Models}},
author = {Xiang He and Chenxing Li and Jinting Wang and Yan Rong and Tianxin Xie and Zeyu Xie and Wenfu Wang and Li Liu and... | Large Audio-Language Models (LALMs) excel at perception but lack grounded reasoning. Existing methods rely on supervised chain-of-thought (CoT) data or coarse Reinforcement Learning (RL) rewards that do not directly evaluate reasoning quality, yielding chains logically ungrounded in audio. To bridge this gap, we propos... | 2604.18187 | title_snapshot |
zhang26t_interspeech | EChO-Agent: Evidence Chain Orchestration Agent for Audio Reasoning | [
"Siyuan Zhang",
"Jian Zong",
"Junyu Wang",
"Peiyuan Jiang",
"Jiahao Yan",
"Jingyu Zhang",
"Tianrui Wang",
"Xiaobao Wang",
"Longbiao Wang",
"Jianwu Dang"
] | https://www.isca-archive.org/interspeech_2026/zhang26t_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhang26t_interspeech.pdf | 10.21437/Interspeech.2026-1313 | 251-255 | @inproceedings{zhang26t_interspeech,
title = {{EChO-Agent: Evidence Chain Orchestration Agent for Audio Reasoning}},
author = {Siyuan Zhang and Jian Zong and Junyu Wang and Peiyuan Jiang and Jiahao Yan and Jingyu Zhang and Tianrui Wang and Xiaobao Wang and Longbiao Wang and Jianwu Dang},
year = {2026}... | While LALMs show promise on audio question answering, they fail to focus on question-relevant segments of audio and provide a clear, checkable reasoning process when dealing with complex audio reasoning. Reinforcement learning and tool-augmented prompting can help models better relate questions to audio but lack a reli... | 2606.15141 | title_snapshot |
wang26t_interspeech | MATA: A Training-Free Approach to Mitigate Cross-Modal Attention Imbalance in Large Audio Language Models | [
"Junyu Wang",
"Jian Zong",
"Tianrui Wang",
"Zhengding Luo",
"Meng Ge",
"Xiaobao Wang",
"Longbiao Wang",
"Jianwu Dang"
] | https://www.isca-archive.org/interspeech_2026/wang26t_interspeech.html | https://www.isca-archive.org/interspeech_2026/wang26t_interspeech.pdf | 10.21437/Interspeech.2026-1212 | 256-260 | @inproceedings{wang26t_interspeech,
title = {{MATA: A Training-Free Approach to Mitigate Cross-Modal Attention Imbalance in Large Audio Language Models}},
author = {Junyu Wang and Jian Zong and Tianrui Wang and Zhengding Luo and Meng Ge and Xiaobao Wang and Longbiao Wang and Jianwu Dang},
year = {2026... | Large Audio Language Models (LALMs) often suffer from audio-textual attention imbalance, prioritizing text over acoustic information during multi-modal fusion. This bias limits the utilization of acoustic cues and degrades audio reasoning performance. To mitigate this, we propose MATA, a novel training-free method that... | 2509.18816 | title_judge |
jin26b_interspeech | Learning to Rescale: On-the-Fly Sequence Length Adaptation in Non-Autoregressive Speech Synthesis | [
"Jiawei Jin",
"Ren Wang",
"Zhiyu Cui",
"Shun Lei",
"Yixuan Zhou",
"Zhiyong Wu"
] | https://www.isca-archive.org/interspeech_2026/jin26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/jin26b_interspeech.pdf | 10.21437/Interspeech.2026-1069 | 261-265 | @inproceedings{jin26b_interspeech,
title = {{Learning to Rescale: On-the-Fly Sequence Length Adaptation in Non-Autoregressive Speech Synthesis}},
author = {Jiawei Jin and Ren Wang and Zhiyu Cui and Shun Lei and Yixuan Zhou and Zhiyong Wu},
year = {2026},
booktitle = {{Interspeech 2026}},
pages ... | Non-autoregressive (NAR) text-to-speech (TTS) models excel in parallel inference and style consistency. However, their reliance on accurate character-level or global duration predictions remains a critical bottleneck, limiting both synthesis fidelity and naturalness. This paper proposes ElasticDLM (Elastic-length Diffu... | null | null |
zhu26e_interspeech | OmniVoice: Towards Omnilingual Zero-Shot Text-to-Speech with Diffusion Language Models | [
"Han Zhu",
"Lingxuan Ye",
"Wei Kang",
"Zengwei Yao",
"Liyong Guo",
"Fangjun Kuang",
"Zhifeng Han",
"Weiji Zhuang",
"Long Lin",
"Daniel Povey"
] | https://www.isca-archive.org/interspeech_2026/zhu26e_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhu26e_interspeech.pdf | 10.21437/Interspeech.2026-3256 | 266-271 | @inproceedings{zhu26e_interspeech,
title = {{OmniVoice: Towards Omnilingual Zero-Shot Text-to-Speech with Diffusion Language Models}},
author = {Han Zhu and Lingxuan Ye and Wei Kang and Zengwei Yao and Liyong Guo and Fangjun Kuang and Zhifeng Han and Weiji Zhuang and Long Lin and Daniel Povey},
year =... | We present OmniVoice, a massively multilingual zero-shot text-to-speech (TTS) model that scales to over 600 languages. At its core is a novel diffusion language model-style discrete non-autoregressive (NAR) architecture. Unlike conventional discrete NAR models that suffer from performance bottlenecks in complex two-sta... | 2604.00688 | title_snapshot |
lee26j_interspeech | WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models | [
"Hanna Lee",
"Tan Dat Nguyen",
"Jaehoon Kang",
"Kyuhong Shim"
] | https://www.isca-archive.org/interspeech_2026/lee26j_interspeech.html | https://www.isca-archive.org/interspeech_2026/lee26j_interspeech.pdf | 10.21437/Interspeech.2026-943 | 272-277 | @inproceedings{lee26j_interspeech,
title = {{WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models}},
author = {Hanna Lee and Tan Dat Nguyen and Jaehoon Kang and Kyuhong Shim},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {272--277},... | Recent decoder-only autoregressive text-to-speech (AR-TTS) models produce high-fidelity speech, but their memory and compute costs scale quadratically with sequence length due to full self-attention. In this paper, we propose WAND, Windowed Attention and Knowledge Distillation, a framework that adapts pretrained AR-TTS... | 2604.08558 | title_snapshot |
xie26c_interspeech | VoiceTTA: Enhancing Zero-Shot Text-to-Speech via Reinforcement Learning-Based Test-Time Adaptation | [
"Tianxin Xie",
"Chenxing Li",
"Dong Yu",
"Li Liu"
] | https://www.isca-archive.org/interspeech_2026/xie26c_interspeech.html | https://www.isca-archive.org/interspeech_2026/xie26c_interspeech.pdf | 10.21437/Interspeech.2026-1757 | 278-282 | @inproceedings{xie26c_interspeech,
title = {{VoiceTTA: Enhancing Zero-Shot Text-to-Speech via Reinforcement Learning-Based Test-Time Adaptation}},
author = {Tianxin Xie and Chenxing Li and Dong Yu and Li Liu},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {278--282},
doi = {... | Recently, zero-shot text-to-speech (TTS) has enabled high-fidelity and expressive speech synthesis, but it often fails to imitate unseen speaking styles from uncommon scenarios (e.g., crosstalk, dialects). Moreover, fine-tuning pretrained models requires large, high-quality datasets, limiting rapid personalization. We ... | 2606.26534 | title_snapshot |
choi26b_interspeech | ZeSTA: Zero-Shot TTS Augmentation with Domain-Conditioned Training for Data-Efficient Personalized Speech Synthesis | [
"Youngwon Choi",
"Jinwoo Oh",
"Hwayeon Kim",
"Hyeonyu Kim"
] | https://www.isca-archive.org/interspeech_2026/choi26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/choi26b_interspeech.pdf | 10.21437/Interspeech.2026-1269 | 283-288 | @inproceedings{choi26b_interspeech,
title = {{ZeSTA: Zero-Shot TTS Augmentation with Domain-Conditioned Training for Data-Efficient Personalized Speech Synthesis}},
author = {Youngwon Choi and Jinwoo Oh and Hwayeon Kim and Hyeonyu Kim},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = ... | We investigate the use of zero-shot text-to-speech (ZS-TTS) as a data augmentation source for low-resource personalized speech synthesis. While synthetic augmentation can provide linguistically rich and phonetically diverse speech, naively mixing large amounts of synthetic speech with limited real recordings often lead... | 2603.04219 | title_snapshot |
wang26f_interspeech | Dual-Space Constrained Face-Based Zero-Shot Text-to-Speech Synthesis | [
"Jianrong Wang",
"Shengjie Zhou",
"Ju Zhang",
"Dengcheng Hu",
"Qi Li"
] | https://www.isca-archive.org/interspeech_2026/wang26f_interspeech.html | https://www.isca-archive.org/interspeech_2026/wang26f_interspeech.pdf | 10.21437/Interspeech.2026-409 | 289-293 | @inproceedings{wang26f_interspeech,
title = {{Dual-Space Constrained Face-Based Zero-Shot Text-to-Speech Synthesis}},
author = {Jianrong Wang and Shengjie Zhou and Ju Zhang and Dengcheng Hu and Qi Li},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {289--293},
doi = {10.21437... | A human face conveys rich cues about speaker identity, enabling face-based zero-shot text-to-speech (TTS) for unseen speakers. However, in modular face-based TTS systems, the acoustic model is typically trained on speech-derived embeddings, while face-derived representations are introduced only at inference time, often... | null | null |
polok26_interspeech | Mind the Gap: Impact of Synthetic Conversational Data on Multi-Talker ASR and Speaker Diarization | [
"Alexander Polok",
"Ivan Medennikov",
"Honza Černocký",
"Shinji Watanabe",
"Lukáš Burget",
"Samuele Cornell"
] | https://www.isca-archive.org/interspeech_2026/polok26_interspeech.html | https://www.isca-archive.org/interspeech_2026/polok26_interspeech.pdf | 10.21437/Interspeech.2026-443 | 294-299 | @inproceedings{polok26_interspeech,
title = {{Mind the Gap: Impact of Synthetic Conversational Data on Multi-Talker ASR and Speaker Diarization}},
author = {Alexander Polok and Ivan Medennikov and Honza Černocký and Shinji Watanabe and Lukáš Burget and Samuele Cornell},
year = {2026},
booktitle = {{... | Recent breakthroughs in multi-talker ASR (MT-ASR) and speaker diarization (SD) rely on synthetic data to mitigate the scarcity of large-scale conversational recordings, yet the impact of specific simulation choices remains poorly understood. To mind the gap between simulated mixtures and real-world interactions, we pre... | 2605.15442 | title_snapshot |
polok26b_interspeech | Grounding Spoken LLMs in Multi-Speaker Audio via Diarization Conditioning | [
"Alexander Polok",
"Samuele Cornell",
"Sathvik Udupa",
"Honza Černocký",
"Shinji Watanabe",
"Lukáš Burget"
] | https://www.isca-archive.org/interspeech_2026/polok26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/polok26b_interspeech.pdf | 10.21437/Interspeech.2026-445 | 300-305 | @inproceedings{polok26b_interspeech,
title = {{Grounding Spoken LLMs in Multi-Speaker Audio via Diarization Conditioning}},
author = {Alexander Polok and Samuele Cornell and Sathvik Udupa and Honza Černocký and Shinji Watanabe and Lukáš Burget},
year = {2026},
booktitle = {{Interspeech 2026}},
pag... | We propose diarization-conditioned spoken language models (SLMs), a strategy for extending SLMs to far-field multi-talker audio. Rather than adapting the decoder via Serialized Output Training, which risks catastrophic forgetting, we condition the acoustic encoder on diarization masks to extract target-speaker represen... | 2606.18134 | title_snapshot |
guo26_interspeech | GLAD: Global-Local Aware Dynamic Mixture-of-Experts for Multi-Talker ASR | [
"Yujie Guo",
"Jiaming Zhou",
"Yuhang Jia",
"Shiwan Zhao",
"Yong Qin"
] | https://www.isca-archive.org/interspeech_2026/guo26_interspeech.html | https://www.isca-archive.org/interspeech_2026/guo26_interspeech.pdf | 10.21437/Interspeech.2026-1022 | 306-310 | @inproceedings{guo26_interspeech,
title = {{GLAD: Global-Local Aware Dynamic Mixture-of-Experts for Multi-Talker ASR}},
author = {Yujie Guo and Jiaming Zhou and Yuhang Jia and Shiwan Zhao and Yong Qin},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {306--310},
doi = {10.2143... | End-to-end multi-talker automatic speech recognition (MTASR) faces significant challenges in accurately transcribing overlapping speech. A critical bottleneck is that speaker-specific acoustic characteristics, which are essential for distinguishing overlapping speech, are often diluted in deep network layers. To addres... | 2509.13093 | title_snapshot |
kashiwagi26_interspeech | Speaker-Aware Hypothesis Clustering and Merging for Target-Speaker-free and Target-Speaker Multi-Talker ASR | [
"Yosuke Kashiwagi",
"Osamu Take",
"Hayato Futami",
"Emiru Tsunoo",
"Siddhant Arora",
"Shinji Watanabe"
] | https://www.isca-archive.org/interspeech_2026/kashiwagi26_interspeech.html | https://www.isca-archive.org/interspeech_2026/kashiwagi26_interspeech.pdf | 10.21437/Interspeech.2026-1604 | 311-315 | @inproceedings{kashiwagi26_interspeech,
title = {{Speaker-Aware Hypothesis Clustering and Merging for Target-Speaker-free and Target-Speaker Multi-Talker ASR}},
author = {Yosuke Kashiwagi and Osamu Take and Hayato Futami and Emiru Tsunoo and Siddhant Arora and Shinji Watanabe},
year = {2026},
bookti... | Multi-talker automatic speech recognition (ASR) has attracted increasing attention for overlapping speech scenarios. Hypothesis Clustering and Merging (HCM) achieves strong performance by clustering hypotheses in transcript space, but it does not explicitly consider speaker identity during clustering. As a result, HCM ... | null | null |
park26e_interspeech | Pushing the Boundaries of Streaming Multi-Speaker ASR: A Systematic Study of Architectural Trade-offs | [
"Taejin Park",
"Ivan Medennikov",
"Kunal Dhawan",
"Weiqing Wang",
"Jagadeesh Balam",
"Boris Ginsburg"
] | https://www.isca-archive.org/interspeech_2026/park26e_interspeech.html | https://www.isca-archive.org/interspeech_2026/park26e_interspeech.pdf | 10.21437/Interspeech.2026-2005 | 316-321 | @inproceedings{park26e_interspeech,
title = {{Pushing the Boundaries of Streaming Multi-Speaker ASR: A Systematic Study of Architectural Trade-offs}},
author = {Taejin Park and Ivan Medennikov and Kunal Dhawan and Weiqing Wang and Jagadeesh Balam and Boris Ginsburg},
year = {2026},
booktitle = {{Int... | Streaming multi-speaker ASR is a challenging task that must balance accuracy, latency, and efficiency while handling overlapping speech and maintaining coherent long-context modeling over extended conversations in an online fashion. We present a unified framework that categorizes streaming multi-speaker ASR into four a... | 2609.10265 | title_snapshot |
tawara26_interspeech | Who Spoke What When? Evaluating Spoken Language Models for Conversational ASR with Semantic and Overlap-Aware Metrics | [
"Naohiro Tawara",
"Samuele Cornell",
"Alexander Polok",
"Marc Delcroix",
"Lukáš Burget",
"Shinji Watanabe"
] | https://www.isca-archive.org/interspeech_2026/tawara26_interspeech.html | https://www.isca-archive.org/interspeech_2026/tawara26_interspeech.pdf | 10.21437/Interspeech.2026-2912 | 322-327 | @inproceedings{tawara26_interspeech,
title = {{Who Spoke What When? Evaluating Spoken Language Models for Conversational ASR with Semantic and Overlap-Aware Metrics}},
author = {Naohiro Tawara and Samuele Cornell and Alexander Polok and Marc Delcroix and Lukáš Burget and Shinji Watanabe},
year = {2026... | Conversational automatic speech recognition remains challeng-ing due to overlapping speech, far-field noise, and varying speaker counts. While recent LLM-based systems perform well on single-speaker benchmarks, their robustness in multi-speaker settings is unclear. We systematically compare LLM-based and modular pipeli... | 2603.22709 | title_snapshot |
ye26_interspeech | Which Speech Representation Better Matches Text-Native Reasoning? A Study of Speech-Text Alignment on Frame Rate and Representation | [
"Zhen Ye",
"Xu Tan",
"Yiming Li",
"Guangyan Zhang",
"Chimin Chan",
"Haohe Liu",
"Zhengxi Liu",
"Hongzhan Lin",
"Zheqi Dai",
"Xinshen Zhang",
"Peiwen Sun",
"Qiuqiang Kong",
"Wei Xue"
] | https://www.isca-archive.org/interspeech_2026/ye26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ye26_interspeech.pdf | 10.21437/Interspeech.2026-21 | 328-336 | @inproceedings{ye26_interspeech,
title = {{Which Speech Representation Better Matches Text-Native Reasoning? A Study of Speech-Text Alignment on Frame Rate and Representation}},
author = {Zhen Ye and Xu Tan and Yiming Li and Guangyan Zhang and Chimin Chan and Haohe Liu and Zhengxi Liu and Hongzhan Lin and Zh... | Spoken dialogue models typically start from text LLM backbones, yet reasoning often degrades when conditioning on speech instead of text. We attribute part of this modality gap to a temporal-granularity mismatch: speech tokens are temporally redundant and far longer than text under matched semantics, diluting per-token... | 2606.12199 | title_snapshot |
sadok26_interspeech | InsideSSL: Understanding Self-Supervised Speech Representations using a Model-Centric Perspective | [
"Samir Sadok",
"Xavier Alameda-Pineda"
] | https://www.isca-archive.org/interspeech_2026/sadok26_interspeech.html | https://www.isca-archive.org/interspeech_2026/sadok26_interspeech.pdf | 10.21437/Interspeech.2026-733 | 337-346 | @inproceedings{sadok26_interspeech,
title = {{InsideSSL: Understanding Self-Supervised Speech Representations using a Model-Centric Perspective}},
author = {Samir Sadok and Xavier Alameda-Pineda},
year = {2026},
booktitle = {{Interspeech 2026 [Long Track]}},
pages = {337--346},
doi = {... | Self-supervised learning (SSL) models, such as Wav2Vec2, HuBERT, and WavLM, have become foundational across a wide range of speech and audio tasks. Despite their success, understanding their internal layer-wise dynamics remains an ongoing challenge. To address this, we propose a two-part model-centric framework called ... | 2607.06392 | title_snapshot |
mcauliffe26_interspeech | Montreal Forced Aligner and the state of speech-to-text alignment in 2026 | [
"Michael McAuliffe",
"Kaylynn Gunter",
"Michael Wagner",
"Morgan Sonderegger"
] | https://www.isca-archive.org/interspeech_2026/mcauliffe26_interspeech.html | https://www.isca-archive.org/interspeech_2026/mcauliffe26_interspeech.pdf | 10.21437/Interspeech.2026-2734 | 347-356 | @inproceedings{mcauliffe26_interspeech,
title = {{Montreal Forced Aligner and the state of speech-to-text alignment in 2026}},
author = {Michael McAuliffe and Kaylynn Gunter and Michael Wagner and Morgan Sonderegger},
year = {2026},
booktitle = {{Interspeech 2026 [Long Track]}},
pages = {347--... | The Montreal Forced Aligner (MFA) was released in 2016 and has since become the most widely used tool for forced alignment in research and industry. In the decade since, MFA has undergone substantial development, including expanded coverage across more languages and dialects using larger open-source datasets, harmonize... | 2606.18466 | title_snapshot |
arcosholzinger26_interspeech | GRIDS: Dimensionality-Aware Anomaly Detection in Learned Representations of Self-Supervised Speech Models | [
"Sandra Arcos-Holzinger",
"Sarah M. Erfani",
"James Bailey",
"Sanjeev Khudanpur"
] | https://www.isca-archive.org/interspeech_2026/arcosholzinger26_interspeech.html | https://www.isca-archive.org/interspeech_2026/arcosholzinger26_interspeech.pdf | 10.21437/Interspeech.2026-2719 | 357-366 | @inproceedings{arcosholzinger26_interspeech,
title = {{GRIDS: Dimensionality-Aware Anomaly Detection in Learned Representations of Self-Supervised Speech Models}},
author = {Sandra Arcos-Holzinger and Sarah M. Erfani and James Bailey and Sanjeev Khudanpur},
year = {2026},
booktitle = {{Interspeech 2... | Self-supervised speech models (S3Ms) achieve strong downstream performance, yet their learned representations remain poorly understood under natural and adversarial perturbations. Prior studies rely on representation similarity or global dimensionality, offering limited visibility into local geometric changes. We ask: ... | 2605.02715 | title_judge |
grinberg26_interspeech | ALARM: Audio–Language Alignment for Reasoning Models | [
"Petr Grinberg",
"Hassan Shahmohammadi"
] | https://www.isca-archive.org/interspeech_2026/grinberg26_interspeech.html | https://www.isca-archive.org/interspeech_2026/grinberg26_interspeech.pdf | 10.21437/Interspeech.2026-759 | 367-376 | @inproceedings{grinberg26_interspeech,
title = {{ALARM: Audio–Language Alignment for Reasoning Models}},
author = {Petr Grinberg and Hassan Shahmohammadi},
year = {2026},
booktitle = {{Interspeech 2026 [Long Track]}},
pages = {367--376},
doi = {10.21437/Interspeech.2026-759},
issn ... | Large audio language models (ALMs) extend LLMs with auditory understanding. A common approach freezes the LLM and trains only an adapter on self-generated targets. However, this fails for reasoning LLMs (RLMs) whose built-in chain-of-thought traces expose the textual surrogate input, yielding unnatural responses. We pr... | 2603.09556 | title_snapshot |
papi26_interspeech | Cross-Attention is Half Explanation in Speech-to-Text Models | [
"Sara Papi",
"Dennis Fucci",
"Marco Gaido",
"Matteo Negri",
"Luisa Bentivogli"
] | https://www.isca-archive.org/interspeech_2026/papi26_interspeech.html | https://www.isca-archive.org/interspeech_2026/papi26_interspeech.pdf | 10.21437/Interspeech.2026-40 | 377-386 | @inproceedings{papi26_interspeech,
title = {{Cross-Attention is Half Explanation in Speech-to-Text Models}},
author = {Sara Papi and Dennis Fucci and Marco Gaido and Matteo Negri and Luisa Bentivogli},
year = {2026},
booktitle = {{Interspeech 2026 [Long Track]}},
pages = {377--386},
doi ... | Cross-attention is widely used in speech-to-text (S2T) systems and often exploited for downstream applications such as timestamp prediction and speech-text alignment, under the assumption that it reflects input-output dependencies. While extensively debated in NLP, its explanatory role remains underexplored in the spee... | 2509.18010 | title_snapshot |
kang26_interspeech | Beyond Short Segments : Expanding Speaker Embeddings with Vector Archives | [
"Hyunku Kang",
"Minkyu Cho",
"Chanwoo Kim"
] | https://www.isca-archive.org/interspeech_2026/kang26_interspeech.html | https://www.isca-archive.org/interspeech_2026/kang26_interspeech.pdf | 10.21437/Interspeech.2026-3192 | 387-392 | @inproceedings{kang26_interspeech,
title = {{Beyond Short Segments : Expanding Speaker Embeddings with Vector Archives}},
author = {Hyunku Kang and Minkyu Cho and Chanwoo Kim},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {387--392},
doi = {10.21437/Interspeech.2026-3192},
... | The performance of state-of-the-art speaker verification (SV) systems severely degrades on short utterances due to insufficient speaker-specific information. To address this critical challenge, we propose the Vector Archive Mapping ECAPA (VAM-ECAPA), a novel system designed to enhance feature extraction from short-dura... | 2609.25007 | title_snapshot |
kim26i_interspeech | Revisiting Label-Free Speaker Embedding Enhancement with vMF Profile Likelihood | [
"Seunghwan Kim",
"Jinyong Kim",
"Sooyoung Yang",
"Youngjin Ko",
"Myungjoo Kang"
] | https://www.isca-archive.org/interspeech_2026/kim26i_interspeech.html | https://www.isca-archive.org/interspeech_2026/kim26i_interspeech.pdf | 10.21437/Interspeech.2026-1146 | 393-397 | @inproceedings{kim26i_interspeech,
title = {{Revisiting Label-Free Speaker Embedding Enhancement with vMF Profile Likelihood}},
author = {Seunghwan Kim and Jinyong Kim and Sooyoung Yang and Youngjin Ko and Myungjoo Kang},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {393--397},
d... | Embedding enhancement improves speaker verification under acoustic mismatch without modifying a frozen backbone. Recent work has established a practical label-free setting for this task, but often adopts increasingly structured formulations. Here, the clean target is directly observed during training, making enhancemen... | null | null |
huang26m_interspeech | On the Robustness of Speaker Embeddings for Cross-Domain Speaker Retrieval | [
"Chuanqi Huang",
"Wei Xie",
"Xilu Wang"
] | https://www.isca-archive.org/interspeech_2026/huang26m_interspeech.html | https://www.isca-archive.org/interspeech_2026/huang26m_interspeech.pdf | 10.21437/Interspeech.2026-1796 | 398-402 | @inproceedings{huang26m_interspeech,
title = {{On the Robustness of Speaker Embeddings for Cross-Domain Speaker Retrieval}},
author = {Chuanqi Huang and Wei Xie and Xilu Wang},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {398--402},
doi = {10.21437/Interspeech.2026-1796},
... | Deploying speaker retrieval systems requires robust cross-domain embedding generalization. However, existing benchmarks focus on verification metrics, leaving ranking stability under retrieval constraints under-explored. This paper evaluates six pre-trained embedding models across multiple cross-domain scenarios. First... | null | null |
so26_interspeech | Toward Open-Set Speaker Attribute Prediction with Keyword-Appended LLM Embeddings | [
"Byoungjun So",
"Jaejun Lee",
"Kyogu Lee"
] | https://www.isca-archive.org/interspeech_2026/so26_interspeech.html | https://www.isca-archive.org/interspeech_2026/so26_interspeech.pdf | 10.21437/Interspeech.2026-3203 | 403-407 | @inproceedings{so26_interspeech,
title = {{Toward Open-Set Speaker Attribute Prediction with Keyword-Appended LLM Embeddings}},
author = {Byoungjun So and Jaejun Lee and Kyogu Lee},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {403--407},
doi = {10.21437/Interspeech.2026-32... | Understanding speaker attributes is crucial for voice-related applications, yet conventional approaches rely on fixed categorical labels, lacking semantic richness and zero-shot generalizability. We propose a novel framework for open-set speaker attribute prediction leveraging Large Language Model (LLM) embed-dings to ... | 2606.21979 | title_snapshot |
ulgen26_interspeech | Rethinking Speaker Embeddings for Speech Generation: Sub-Center Modeling for Capturing Intra-Speaker Diversity | [
"Ismail Rasim Ulgen",
"John Hansen",
"Carlos Busso",
"Berrak Sisman"
] | https://www.isca-archive.org/interspeech_2026/ulgen26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ulgen26_interspeech.pdf | 10.21437/Interspeech.2026-942 | 408-413 | @inproceedings{ulgen26_interspeech,
title = {{Rethinking Speaker Embeddings for Speech Generation: Sub-Center Modeling for Capturing Intra-Speaker Diversity}},
author = {Ismail Rasim Ulgen and John Hansen and Carlos Busso and Berrak Sisman},
year = {2026},
booktitle = {{Interspeech 2026}},
pages ... | Modeling speech variation is key to natural, expressive generation. Speaker embeddings are commonly used to condition personalized speech systems, but they are typically trained for speaker recognition, where intra-speaker variability is suppressed and inter-speaker separation is maximized. This objective leads to over... | 2407.04291 | title_snapshot |
gao26_interspeech | NoiseLoRA-SV: Hierarchical Noise-Conditioned Adaptation with Embedding Distillation for Robust Speaker Verification | [
"Dai Gao",
"Chen Jiang",
"Sizhe Liu",
"Peng Zhang"
] | https://www.isca-archive.org/interspeech_2026/gao26_interspeech.html | https://www.isca-archive.org/interspeech_2026/gao26_interspeech.pdf | 10.21437/Interspeech.2026-64 | 414-418 | @inproceedings{gao26_interspeech,
title = {{NoiseLoRA-SV: Hierarchical Noise-Conditioned Adaptation with Embedding Distillation for Robust Speaker Verification}},
author = {Dai Gao and Chen Jiang and Sizhe Liu and Peng Zhang},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {414--418}... | Current speaker verification (SV) models, including Low-Rank Adaptation (LoRA) variants, rely on static inference-time parameters and show limited robustness to non-stationary noise. We propose NoiseLoRA-SV, a dynamic framework that generates instance-adaptive weights on-the-fly during inference. Instead of full end-to... | null | null |
shang26_interspeech | Seed-Enh: Generative Speech Enhancement in Decoupled Semantic and Timbre Spaces | [
"Zengqiang Shang",
"Biao Liu",
"Yu Zhao",
"Pengyuan Zhang"
] | https://www.isca-archive.org/interspeech_2026/shang26_interspeech.html | https://www.isca-archive.org/interspeech_2026/shang26_interspeech.pdf | 10.21437/Interspeech.2026-200 | 419-423 | @inproceedings{shang26_interspeech,
title = {{Seed-Enh: Generative Speech Enhancement in Decoupled Semantic and Timbre Spaces}},
author = {Zengqiang Shang and Biao Liu and Yu Zhao and Pengyuan Zhang},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {419--423},
doi = {10.21437/... | Most existing speech enhancement operate directly in the acoustic space, where noise and speech components are inherently entangled, leading to artifacts such as high-frequency attenuation and background holes. In this paper, we propose Seed-Enh, a generative speech enhancement framework that performs enhancement in de... | null | null |
koo26_interspeech | VeRe-Flow: Guiding Flow Matching toward Clean Speech via Velocity Contrastive Regularization and Representation Alignment for Noise-Robust Bandwidth Expansion | [
"Sujin Koo",
"Sangyoon Kim",
"Ji Sub Um",
"Hoirin Kim"
] | https://www.isca-archive.org/interspeech_2026/koo26_interspeech.html | https://www.isca-archive.org/interspeech_2026/koo26_interspeech.pdf | 10.21437/Interspeech.2026-712 | 424-428 | @inproceedings{koo26_interspeech,
title = {{VeRe-Flow: Guiding Flow Matching toward Clean Speech via Velocity Contrastive Regularization and Representation Alignment for Noise-Robust Bandwidth Expansion}},
author = {Sujin Koo and Sangyoon Kim and Ji Sub Um and Hoirin Kim},
year = {2026},
booktitle =... | Noise-robust bandwidth expansion aims to reconstruct high-fidelity wideband speech from noisy low-resolution inputs. While flow matching has shown strong performance in speech generation, accurately recovering clean speech from noisy inputs remains challenging due to the ambiguity of velocity estimation under noise. In... | 2606.29450 | title_snapshot |
li26l_interspeech | HFMSE: Harmonic-Guided Speech Enhancement with Flow Matching | [
"Jizhen Li",
"Weiping Tu",
"Yuhong Yang",
"Xinhong Li"
] | https://www.isca-archive.org/interspeech_2026/li26l_interspeech.html | https://www.isca-archive.org/interspeech_2026/li26l_interspeech.pdf | 10.21437/Interspeech.2026-722 | 429-433 | @inproceedings{li26l_interspeech,
title = {{HFMSE: Harmonic-Guided Speech Enhancement with Flow Matching}},
author = {Jizhen Li and Weiping Tu and Yuhong Yang and Xinhong Li},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {429--433},
doi = {10.21437/Interspeech.2026-722},
... | Generative speech enhancement methods have shown impressive performance by directly modeling clean speech distribution, yet their effectiveness critically depends on the reliability of conditional information. Mainstream conditioning strategy face two primary challenges: the shallow features extracted from noisy speech... | null | null |
gao26e_interspeech | PhASE-Flow: Phonetic-Conditioned Acoustic Flow Matching in SSL Representation Domain for Speech Enhancement | [
"Jun Gao",
"Xiaobin Rong",
"Yu Sun",
"Dahan Wang",
"Jing Lu"
] | https://www.isca-archive.org/interspeech_2026/gao26e_interspeech.html | https://www.isca-archive.org/interspeech_2026/gao26e_interspeech.pdf | 10.21437/Interspeech.2026-916 | 434-439 | @inproceedings{gao26e_interspeech,
title = {{PhASE-Flow: Phonetic-Conditioned Acoustic Flow Matching in SSL Representation Domain for Speech Enhancement}},
author = {Jun Gao and Xiaobin Rong and Yu Sun and Dahan Wang and Jing Lu},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {434--... | Flow matching (FM) enables high-fidelity generation, while self-supervised learning (SSL) speech models provide hierarchical representations spanning acoustic and phonetic levels. However, existing FM-based speech enhancement (SE) methods operate primarily in the spectral domain, treating SSL features only as external ... | 2606.17806 | title_snapshot |
ojha26_interspeech | Bridging Self-Supervised Learning and Speech Enhancement: A Wav2Vec2-Conditioned Framework | [
"Shuubham Ojha",
"Carol Espy-Wilson"
] | https://www.isca-archive.org/interspeech_2026/ojha26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ojha26_interspeech.pdf | 10.21437/Interspeech.2026-964 | 440-444 | @inproceedings{ojha26_interspeech,
title = {{Bridging Self-Supervised Learning and Speech Enhancement: A Wav2Vec2-Conditioned Framework}},
author = {Shuubham Ojha and Carol Espy-Wilson},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {440--444},
doi = {10.21437/Interspeech.20... | Diffusion models show potential for speech enhancement but lack linguistic guidance. We condition a diffusion-based model on wav2vec 2.0 features from noisy input, injected at the U-Net bottleneck via Feature-wise Linear Modulation (FiLM). Phonetic representations from wav2vec 2.0 features of degraded speech, anchor th... | 2606.22591 | title_snapshot |
liu26j_interspeech | Text-Annotated Noisy Speech as Supervision: A Dual-Learning Framework for Target-Domain Clean-Free Speech Enhancement | [
"Xin Liu",
"Shulin He",
"Xueliang Zhang"
] | https://www.isca-archive.org/interspeech_2026/liu26j_interspeech.html | https://www.isca-archive.org/interspeech_2026/liu26j_interspeech.pdf | 10.21437/Interspeech.2026-1259 | 445-448 | @inproceedings{liu26j_interspeech,
title = {{Text-Annotated Noisy Speech as Supervision: A Dual-Learning Framework for Target-Domain Clean-Free Speech Enhancement}},
author = {Xin Liu and Shulin He and Xueliang Zhang},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {445--448},
doi ... | Deep learning-based speech enhancement (SE) models are typically trained on synthetic pairs of clean speech and synthesized noisy speech, which often generalize poorly to real-world acoustic environments. In practical target environments, however, paired clean target speech aligned with noisy recordings is usually unav... | null | null |
zhu26b_interspeech | G-MaP-SE: Guided Speech Enhancement via GMM-Based Prior Matching | [
"Yike Zhu",
"Ziqian Wang",
"Zikai Liu",
"Xingchen Li",
"Zhuangqi Chen",
"Xianjun Xia",
"Chuanzeng Huang",
"Lei Xie"
] | https://www.isca-archive.org/interspeech_2026/zhu26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhu26b_interspeech.pdf | 10.21437/Interspeech.2026-2148 | 449-454 | @inproceedings{zhu26b_interspeech,
title = {{G-MaP-SE: Guided Speech Enhancement via GMM-Based Prior Matching}},
author = {Yike Zhu and Ziqian Wang and Zikai Liu and Xingchen Li and Zhuangqi Chen and Xianjun Xia and Chuanzeng Huang and Lei Xie},
year = {2026},
booktitle = {{Interspeech 2026}},
pag... | Using speaker embeddings as conditioning can strengthen speech enhancement, but most methods either require clean enrollment audio or rely on embeddings extracted from noisy speech, which are fragile under noise and domain shift. We propose G-MaP-SE, a guided enhancement framework that builds a clean-speech embedding p... | 2606.08580 | title_snapshot |
klement26_interspeech | Analysing Adversarial Priors for Data-driven Unsupervised Speech Enhancement | [
"Dominik Klement",
"Matthew Maciejewski",
"Sanjeev Khudanpur",
"Honza Černocký",
"Lukáš Burget"
] | https://www.isca-archive.org/interspeech_2026/klement26_interspeech.html | https://www.isca-archive.org/interspeech_2026/klement26_interspeech.pdf | 10.21437/Interspeech.2026-2511 | 455-459 | @inproceedings{klement26_interspeech,
title = {{Analysing Adversarial Priors for Data-driven Unsupervised Speech Enhancement}},
author = {Dominik Klement and Matthew Maciejewski and Sanjeev Khudanpur and Honza Černocký and Lukáš Burget},
year = {2026},
booktitle = {{Interspeech 2026}},
pages =... | Deep learning-based speech enhancement typically relies on paired data, creating a domain gap between training and deployment. Unsupervised methods based on generative adversarial networks (GANs) use unpaired data as source priors, but existing single-branch models often suffer from source leakage due to a dominant con... | null | null |
bejugam26_interspeech | UFL-GAN: A Multi-Discriminator GAN for Unsupervised Speech Enhancement | [
"Satvik Bejugam",
"Venkatesh Parvathala",
"Sri Rama Murty Kodukula"
] | https://www.isca-archive.org/interspeech_2026/bejugam26_interspeech.html | https://www.isca-archive.org/interspeech_2026/bejugam26_interspeech.pdf | 10.21437/Interspeech.2026-2565 | 460-464 | @inproceedings{bejugam26_interspeech,
title = {{UFL-GAN: A Multi-Discriminator GAN for Unsupervised Speech Enhancement}},
author = {Satvik Bejugam and Venkatesh Parvathala and Sri Rama Murty Kodukula},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {460--464},
doi = {10.21437... | Most deep learning based speech enhancement methods are usually trained in a supervised fashion, i.e., they typically rely on parallel corpora of noisy and clean speech pairs. This is often difficult to obtain in real-world scenarios, leading to the use of synthetic data. In this work, we propose a novel unsupervised s... | null | null |
hirose26_interspeech | Self-adaptive Gradient Conflict Mitigator for Continuous-Time Diffusion Models | [
"Takumi Hirose",
"Zhiyang Li",
"Nakamasa Inoue"
] | https://www.isca-archive.org/interspeech_2026/hirose26_interspeech.html | https://www.isca-archive.org/interspeech_2026/hirose26_interspeech.pdf | 10.21437/Interspeech.2026-3059 | 465-469 | @inproceedings{hirose26_interspeech,
title = {{Self-adaptive Gradient Conflict Mitigator for Continuous-Time Diffusion Models}},
author = {Takumi Hirose and Zhiyang Li and Nakamasa Inoue},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {465--469},
doi = {10.21437/Interspeech.... | Continuous-time diffusion models have demonstrated strong capabilities for modeling complex data distributions across various domains. However, these models often encounter gradient conflicts, where parameter updates at different timesteps interfere with each other, hindering effective training. To provide theoretical ... | null | null |
zhang26z_interspeech | Time-Unconditional Generative Speech Enhancement via Autonomous Rectified Flow | [
"Wen Zhang",
"Wenbin Jiang",
"Yang Zhang",
"Xiaofei Zhou"
] | https://www.isca-archive.org/interspeech_2026/zhang26z_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhang26z_interspeech.pdf | 10.21437/Interspeech.2026-1679 | 470-474 | @inproceedings{zhang26z_interspeech,
title = {{Time-Unconditional Generative Speech Enhancement via Autonomous Rectified Flow}},
author = {Wen Zhang and Wenbin Jiang and Yang Zhang and Xiaofei Zhou},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {470--474},
doi = {10.21437/I... | Most generative speech enhancement methods rely on explicit time-step embeddings for temporal conditioning. In this paper, we propose the Autonomous Rectified Flow framework, which challenges the necessity of such conditioning. Using a linear interpolation path, we show that the target vector field is inherently time-i... | 2606.20001 | title_snapshot |
ijjada26_interspeech | WaveNorm: A Low-Complexity Time-Domain Neural Adaptive Gain Control for Real-Time Speech Applications | [
"Deepika Ijjada",
"Charan Kumar Reddy B",
"Ashwini Hanaganti",
"Priyanka Devrao Jadhav",
"Varsha Uppalanchi",
"Balaji Padmanaban",
"Nivedita Chennupati",
"Karunakar Reddy Pucchakayala",
"Harish Rajamani",
"Naveen Ambati"
] | https://www.isca-archive.org/interspeech_2026/ijjada26_interspeech.html | https://www.isca-archive.org/interspeech_2026/ijjada26_interspeech.pdf | 10.21437/Interspeech.2026-2115 | 475-479 | @inproceedings{ijjada26_interspeech,
title = {{WaveNorm: A Low-Complexity Time-Domain Neural Adaptive Gain Control for Real-Time Speech Applications}},
author = {Deepika Ijjada and Charan Kumar Reddy B and Ashwini Hanaganti and Priyanka Devrao Jadhav and Varsha Uppalanchi and Balaji Padmanaban and Nivedita C... | Adaptive Gain Control (AGC) maintains consistent loudness in speech systems despite variations in speaker volume, noise, and recording conditions. Conventional AGC relies on fixed attack and release constants, limiting effectiveness in dynamic environments, which often causes clipping, delayed adaptation, noise amplifi... | null | null |
singh26b_interspeech | CHUCKLE - When Humans Teach AI to Learn Emotions the Easy Way | [
"Ankush Pratap Singh",
"Houwei Cao",
"Yong Liu"
] | https://www.isca-archive.org/interspeech_2026/singh26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/singh26b_interspeech.pdf | 10.21437/Interspeech.2026-1591 | 480-484 | @inproceedings{singh26b_interspeech,
title = {{CHUCKLE - When Humans Teach AI to Learn Emotions the Easy Way}},
author = {Ankush Pratap Singh and Houwei Cao and Yong Liu},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {480--484},
doi = {10.21437/Interspeech.2026-1591},
iss... | Curriculum learning (CL) structures training from simple to complex samples, facilitating progressive learning. However, existing CL approaches for emotion recognition often rely on heuristic, data-driven, or model-based definitions of sample difficulty, neglecting the difficulty for human perception, a critical factor... | 2510.09382 | title_snapshot |
jeon26d_interspeech | ParaPairAudioBench: Paralinguistic Pairwise Audio Benchmark for LALM-as-a-Judge | [
"Jisu Jeon",
"Seungyeon Jwa",
"Joosung Lee",
"Jinhyeon Kim",
"Woojin Chung",
"Hwiyeol Jo",
"Jeonghoon Kim",
"Jonghyun Choi",
"Soyoon Kim"
] | https://www.isca-archive.org/interspeech_2026/jeon26d_interspeech.html | https://www.isca-archive.org/interspeech_2026/jeon26d_interspeech.pdf | 10.21437/Interspeech.2026-2021 | 485-489 | @inproceedings{jeon26d_interspeech,
title = {{ParaPairAudioBench: Paralinguistic Pairwise Audio Benchmark for LALM-as-a-Judge}},
author = {Jisu Jeon and Seungyeon Jwa and Joosung Lee and Jinhyeon Kim and Woojin Chung and Hwiyeol Jo and Jeonghoon Kim and Jonghyun Choi and Soyoon Kim},
year = {2026},
... | Large Audio-Language Models (LALMs) have been widely used as judge models for the automatic evaluation of generated speech. However, prior approaches predominantly focus on holistic naturalness, leaving fine-grained paralinguistic distinctions underexplored. We introduce PARAPAIRAUDIOBENCH, a pairwise benchmark of 5,17... | 2606.24648 | title_snapshot |
spiesberger26_interspeech | Predicting Menstrual Cycle Phases from Speech: A Paralinguistic Approach | [
"Anika A. Spiesberger",
"Andreas Triantafyllopoulos",
"Melanie Weirich",
"Bjoern Schuller"
] | https://www.isca-archive.org/interspeech_2026/spiesberger26_interspeech.html | https://www.isca-archive.org/interspeech_2026/spiesberger26_interspeech.pdf | 10.21437/Interspeech.2026-1878 | 490-494 | @inproceedings{spiesberger26_interspeech,
title = {{Predicting Menstrual Cycle Phases from Speech: A Paralinguistic Approach}},
author = {Anika A. Spiesberger and Andreas Triantafyllopoulos and Melanie Weirich and Bjoern Schuller},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {490-... | Hormonal changes throughout the menstrual cycle can affect speech production. However, studies investigating differences in acoustic parameters have found inconsistent results. This could be because changes are likely subtle and multivariate. We therefore apply handcrafted and embedding-based features, combined with ma... | null | null |
anand26_interspeech | ParA-LLM: A Unified Approach to Paralinguistic and Acoustic Speech Understanding | [
"Nishit Anand",
"Jiaqi Su",
"Ke Chen",
"Yunyun Wang",
"Dinesh Manocha",
"Ramani Duraiswami",
"Rithesh Kumar",
"Zeyu Jin"
] | https://www.isca-archive.org/interspeech_2026/anand26_interspeech.html | https://www.isca-archive.org/interspeech_2026/anand26_interspeech.pdf | 10.21437/Interspeech.2026-3015 | 495-500 | @inproceedings{anand26_interspeech,
title = {{ParA-LLM: A Unified Approach to Paralinguistic and Acoustic Speech Understanding}},
author = {Nishit Anand and Jiaqi Su and Ke Chen and Yunyun Wang and Dinesh Manocha and Ramani Duraiswami and Rithesh Kumar and Zeyu Jin},
year = {2026},
booktitle = {{Int... | Recent advances in Audio LLMs have achieved human-level speech recognition, yet existing systems struggle to capture paralinguistic aspects such as speaker traits, expressive variations, and environmental acoustic conditions. To address this, we design a framework of 22 paralinguistic features and create a dataset of o... | 2609.22771 | title_snapshot |
tiwari26_interspeech | Say That Again: Visualizing Paralinguistic Cues with Prosody-Aware Diffusion | [
"Shyamji Tiwari"
] | https://www.isca-archive.org/interspeech_2026/tiwari26_interspeech.html | https://www.isca-archive.org/interspeech_2026/tiwari26_interspeech.pdf | 10.21437/Interspeech.2026-2102 | 501-506 | @inproceedings{tiwari26_interspeech,
title = {{Say That Again: Visualizing Paralinguistic Cues with Prosody-Aware Diffusion}},
author = {Shyamji Tiwari},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {501--506},
doi = {10.21437/Interspeech.2026-2102},
issn = {2958-179... | Text-to-image models ignore paralinguistic cues. Pitch, rate, and emotional inflection shape how listeners visualize a scene, yet audio-to-image methods treat speech as a single opaque vector. NovaDiffusion conditions diffusion-based synthesis on emotion-correlated prosodic features extracted alongside general audio re... | null | null |
dong26c_interspeech | Can Speech LLMs Approximate Human Ratings of Accentedness and Comprehensibility? Evidence from Correlational and Feature-Based Analyses | [
"Wenwei Dong",
"Catia Cucchiarini",
"Roeland van Hout",
"Helmer Strik"
] | https://www.isca-archive.org/interspeech_2026/dong26c_interspeech.html | https://www.isca-archive.org/interspeech_2026/dong26c_interspeech.pdf | 10.21437/Interspeech.2026-1991 | 507-511 | @inproceedings{dong26c_interspeech,
title = {{Can Speech LLMs Approximate Human Ratings of Accentedness and Comprehensibility? Evidence from Correlational and Feature-Based Analyses}},
author = {Wenwei Dong and Catia Cucchiarini and Roeland van Hout and Helmer Strik},
year = {2026},
booktitle = {{In... | Accentedness and comprehensibility scales are widely used to evaluate pronunciation development in second language learners. However, such assessments rely heavily on human rater evaluations. This study investigates whether a speech large language model (LLM) can approximate human judgments of accentedness and comprehe... | null | null |
gorman26_interspeech | Shifting Relational Paradigms for Affective Computing: Affective Resonance, Vitality Affects, and Vocal Interaction Fields | [
"Cy Gorman",
"Yihang Yao"
] | https://www.isca-archive.org/interspeech_2026/gorman26_interspeech.html | https://www.isca-archive.org/interspeech_2026/gorman26_interspeech.pdf | 10.21437/Interspeech.2026-2829 | 512-516 | @inproceedings{gorman26_interspeech,
title = {{Shifting Relational Paradigms for Affective Computing: Affective Resonance, Vitality Affects, and Vocal Interaction Fields}},
author = {Cy Gorman and Yihang Yao},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {512--516},
doi = {... | Affective computing has largely followed an individual-state paradigm, extracting discrete emotion labels or arousal/valence from isolated speakers. We argue this framing is incomplete for interaction. Drawing on affective resonance and vitality-contour accounts, we propose a relational framework in which the primary u... | 2609.09864 | title_snapshot |
liu26r_interspeech | P-SED : Asymmetric Prototype Metric Learning for Weakly Supervised Speech Emotion Diarization | [
"Yumeng Liu",
"Yukun Sun",
"Jian Peng",
"Lixu Sun",
"Nurmemet Yolwas"
] | https://www.isca-archive.org/interspeech_2026/liu26r_interspeech.html | https://www.isca-archive.org/interspeech_2026/liu26r_interspeech.pdf | 10.21437/Interspeech.2026-2388 | 517-521 | @inproceedings{liu26r_interspeech,
title = {{P-SED : Asymmetric Prototype Metric Learning for Weakly Supervised Speech Emotion Diarization}},
author = {Yumeng Liu and Yukun Sun and Jian Peng and Lixu Sun and Nurmemet Yolwas},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {517--521},... | Unlike Speech Emotion Recognition (SER), which predicts utterance-level labels, Speech Emotion Diarization (SED) models fine-grained temporal dynamics. However, learning reliable emotion boundaries from utterance-level labels remains challenging due to scarce frame-level annotations. We propose P-SED, a weakly supervis... | null | null |
liao26b_interspeech | High-Precision Prosodic Boundary Anchors from Acoustic Cues under Weak Supervision | [
"Hanyu Liao",
"Xiaoluan Liu"
] | https://www.isca-archive.org/interspeech_2026/liao26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/liao26b_interspeech.pdf | 10.21437/Interspeech.2026-209 | 522-526 | @inproceedings{liao26b_interspeech,
title = {{High-Precision Prosodic Boundary Anchors from Acoustic Cues under Weak Supervision}},
author = {Hanyu Liao and Xiaoluan Liu},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {522--526},
doi = {10.21437/Interspeech.2026-209},
issn... | Prosodic boundary detection has traditionally relied on manual annotations such as ToBI labels, which can be resource-intensive and are not always available in large speech corpora. This paper presents a weakly supervised framework that derives high-confidence prosodic boundary anchors from acoustic cues, without depen... | null | null |
hu26_interspeech | ArtNet: A JEPA-Like Articulatory Predictive Framework for Robust Zero-Shot Phoneme Recognition | [
"Zeqian Hu",
"Fuliang Weng",
"Shu Shang",
"Yaqian Zhou"
] | https://www.isca-archive.org/interspeech_2026/hu26_interspeech.html | https://www.isca-archive.org/interspeech_2026/hu26_interspeech.pdf | 10.21437/Interspeech.2026-304 | 527-531 | @inproceedings{hu26_interspeech,
title = {{ArtNet: A JEPA-Like Articulatory Predictive Framework for Robust Zero-Shot Phoneme Recognition}},
author = {Zeqian Hu and Fuliang Weng and Shu Shang and Yaqian Zhou},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {527--531},
doi = {... | Zero-shot cross-lingual phoneme recognition is often hindered by the fragility of direct acoustic-to-symbol mapping, which is susceptible to language-specific variations. Echoing joint-embedding predictive architecture (JEPA) work in vision, we propose ArtNet, a framework that explores a structured feature prediction t... | 2606.16595 | title_snapshot |
tanner26_interspeech | wav2VOT: automatic estimation of voice onset time, closure duration, and burst realisation with wav2vec2 | [
"James Tanner",
"Morgan Sonderegger",
"Jane Stuart-Smith",
"Tyler Kendall",
"Jeff Mielke"
] | https://www.isca-archive.org/interspeech_2026/tanner26_interspeech.html | https://www.isca-archive.org/interspeech_2026/tanner26_interspeech.pdf | 10.21437/Interspeech.2026-743 | 532-537 | @inproceedings{tanner26_interspeech,
title = {{wav2VOT: automatic estimation of voice onset time, closure duration, and burst realisation with wav2vec2}},
author = {James Tanner and Morgan Sonderegger and Jane Stuart-Smith and Tyler Kendall and Jeff Mielke},
year = {2026},
booktitle = {{Interspeech ... | While automatic tools for speech annotation are now commonplace within phonetic research pipelines, many tasks require substantial manual correction or training sets to perform accurately. Simultaneously, large speech models such as wav2vec2 have been shown to perform well at speech classification tasks, raising the qu... | 2606.28857 | title_snapshot |
tseng26_interspeech | Time-normalized spectrograms reveal segmental differences in English heterographic homophones | [
"Yu-Hsiang Tseng",
"Harald Baayen"
] | https://www.isca-archive.org/interspeech_2026/tseng26_interspeech.html | https://www.isca-archive.org/interspeech_2026/tseng26_interspeech.pdf | 10.21437/Interspeech.2026-793 | 538-543 | @inproceedings{tseng26_interspeech,
title = {{Time-normalized spectrograms reveal segmental differences in English heterographic homophones}},
author = {Yu-Hsiang Tseng and Harald Baayen},
year = {2026},
booktitle = {{Interspeech 2026}},
pages = {538--543},
doi = {10.21437/Interspeech.... | Homophones are words that are commonly assumed to sound the same but have different meanings. They are either homographic, sharing the same spelling, or heterographic, differing in spelling. This study focuses on heterographic homophones. Past studies have shown that the spoken word duration of such homophones is syste... | null | null |
gonzalez26d_interspeech | Minimum Token Thresholds and Stabilisation for Reliable Automatic Vowel Alignment: Empirical Study on TIMIT Vowels and MFA | [
"Simon Gonzalez",
"Jason Littlefield",
"Tao Hoang",
"Chloe Dean",
"Hayden Ooi",
"Myung Kim",
"Bradley Donnelly",
"Latchman Singh",
"Jennifer Biggs",
"Tim Cawley"
] | https://www.isca-archive.org/interspeech_2026/gonzalez26d_interspeech.html | https://www.isca-archive.org/interspeech_2026/gonzalez26d_interspeech.pdf | 10.21437/Interspeech.2026-934 | 544-548 | @inproceedings{gonzalez26d_interspeech,
title = {{Minimum Token Thresholds and Stabilisation for Reliable Automatic Vowel Alignment: Empirical Study on TIMIT Vowels and MFA}},
author = {Simon Gonzalez and Jason Littlefield and Tao Hoang and Chloe Dean and Hayden Ooi and Myung Kim and Bradley Donnelly and Lat... | Automatic forced alignment is standard in phonetic research, but the minimum data needed for reliable acoustic measurements remains unclear. This study examines how token quantity affects alignment reliability using the MFA on the TIMIT corpus, with manual phoneme boundaries as the gold standard. Focusing on vowels and... | null | null |
camara26b_interspeech | Word Lengthening as a Function of Utterance Position: A Multi-Corpus Study | [
"Mateo Cámara",
"José Luis Blanco",
"Juan Ignacio Godino-Llorente",
"Jeung-Yoon Choi",
"Stefanie Shattuck-Hufnagel"
] | https://www.isca-archive.org/interspeech_2026/camara26b_interspeech.html | https://www.isca-archive.org/interspeech_2026/camara26b_interspeech.pdf | 10.21437/Interspeech.2026-1379 | 549-553 | @inproceedings{camara26b_interspeech,
title = {{Word Lengthening as a Function of Utterance Position: A Multi-Corpus Study}},
author = {Mateo Cámara and José Luis Blanco and Juan Ignacio Godino-Llorente and Jeung-Yoon Choi and Stefanie Shattuck-Hufnagel},
year = {2026},
booktitle = {{Interspeech 202... | Efficient turn-taking requires interlocutors to predict turn endings within a few hundred milliseconds. Beyond syntactic and pragmatic completion, prosody (especially pre-boundary lengthening) supports projection. We test whether turn-final words are longer than mid-sentence words, whether this reflects prosodic modifi... | 2606.23232 | title_snapshot |
zhang26v_interspeech | Larynx segmentation in mid-sagittal speech production real-time MRI | [
"Yubin Zhang",
"Xuan Shi",
"Kevin Huang",
"Prakash Kumar",
"Kevin Lee",
"Louis Goldstein",
"Krishna Nayak",
"Shrikanth Narayanan"
] | https://www.isca-archive.org/interspeech_2026/zhang26v_interspeech.html | https://www.isca-archive.org/interspeech_2026/zhang26v_interspeech.pdf | 10.21437/Interspeech.2026-1402 | 554-559 | @inproceedings{zhang26v_interspeech,
title = {{Larynx segmentation in mid-sagittal speech production real-time MRI}},
author = {Yubin Zhang and Xuan Shi and Kevin Huang and Prakash Kumar and Kevin Lee and Louis Goldstein and Krishna Nayak and Shrikanth Narayanan},
year = {2026},
booktitle = {{Inters... | The spatiotemporal dynamics and coordination of laryngeal movements remain incompletely characterized. This study introduces a larynx segmentation and analysis pipeline for mid-sagittal speech production real-time MRI using Mask2Former, combining supervised learning with semi-supervised refinement. Our results suggest ... | null | null |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 100