Tootfinder

Benchmarking Training Paradigms, Dataset Composition, and Model Scaling for Child ASR in ESPnet
Anyu Ying, Natarajan Balaji Shankar, Chyi-Jiunn Lin, Mohan Shi, Pu Wang, Hye-jin Shim, Siddhant Arora, Hugo Van hamme, Abeer Alwan, Shinji Watanabe
https://arxiv.org/abs/2508.16576

Benchmarking Training Paradigms, Dataset Composition, and Model Scaling for Child ASR in ESPnet
Despite advancements in ASR, child speech recognition remains challenging due to acoustic variability and limited annotated data. While fine-tuning adult ASR models on child speech is common, comparisons with flat-start training remain underexplored. We compare flat-start training across multiple datasets, SSL representations (WavLM, XEUS), and decoder architectures. Our results show that SSL representations are biased toward adult speech, with flat-start training on child speech mitigating the…

i-LAVA: Insights on Low Latency Voice-2-Voice Architecture for Agents
Anupam Purwar, Aditya Choudhary
https://arxiv.org/abs/2509.20971 https://arxiv.org/pd…

i-LAVA: Insights on Low Latency Voice-2-Voice Architecture for Agents
We experiment with a low-latency, end-to-end voice-to-voice communication model to optimize it for real-time conversational applications. By analyzing components essential to voice to voice (V-2-V) system viz. automatic speech recognition (ASR), text-to-speech (TTS), and dialog management, our work analyzes how to reduce processing time while maintaining high-quality interactions to identify the levers for optimizing V-2-V system. Our work identifies that TTS component which generates life-like…

Danish Minister of Justice and chief architect of the current Chat Control proposal, Peter Hummelgaard:
"We must break with the totally erroneous perception that it is everyone's civil liberty to communicate on encrypted messaging services."
Ok, I agree but on one condition: "everyone" means EVERYONE. Police, military, government, politicians, lobbyists, 100% 7x24 transparency. No closed door meetings, no off the record, no sly meets in a pub. Everyone.
Anything else is sinister.

Implicit Augmentation from Distributional Symmetry in Turbulence Super-Resolution
Julia Balla, Jeremiah Bailey, Ali Backour, Elyssa Hofgard, Tommi Jaakkola, Tess Smidt, Ryley McConkey
https://arxiv.org/abs/2509.20683

Implicit Augmentation from Distributional Symmetry in Turbulence Super-Resolution
The immense computational cost of simulating turbulence has motivated the use of machine learning approaches for super-resolving turbulent flows. A central challenge is ensuring that learned models respect physical symmetries, such as rotational equivariance. We show that standard convolutional neural networks (CNNs) can partially acquire this symmetry without explicit augmentation or specialized architectures, as turbulence itself provides implicit rotational augmentation in both time and spac…

System-driven Interactive Design Support for Cloud Architecture: A Qualitative User Experience Study with Novice Engineers
Ryosuke Kohita, Akira Kasuga
https://arxiv.org/abs/2508.12385

System-driven Interactive Design Support for Cloud Architecture: A Qualitative User Experience Study with Novice Engineers
Cloud architecture design presents significant challenges due to the necessity of clarifying ambiguous requirements and systematically addressing complex trade-offs, especially for novice engineers with limited cloud experience. While recent advances in the use of AI tools have broadened available options, system-driven approaches that offer explicit guidance and step-by-step information management may be especially effective in supporting novices during the design process. This study qualitati…

A Systematic Review of FAIR-compliant Big Data Software Reference Architectures
Jo\~ao Pedro de Carvalho Castro, Maria J\'ulia Soares De Grandi, Cristina Dutra de Aguiar
https://arxiv.org/abs/2509.14370

A Systematic Review of FAIR-compliant Big Data Software Reference Architectures
To meet the standards of the Open Science movement, the FAIR Principles emphasize the importance of making scientific data Findable, Accessible, Interoperable, and Reusable. Yet, creating a repository that adheres to these principles presents significant challenges. Managing large volumes of diverse research data and metadata, often generated rapidly, requires a precise approach. This necessity has led to the development of Software Reference Architectures (SRAs) to guide the implementation pro…

HarmoniFuse: A Component-Selective and Prompt-Adaptive Framework for Multi-Task Speech Language Modeling
Yuke Si, Runyan Yang, Yingying Gao, Junlan Feng, Chao Deng, Shilei Zhang
https://arxiv.org/abs/2509.18570

HarmoniFuse: A Component-Selective and Prompt-Adaptive Framework for Multi-Task Speech Language Modeling
Recent advances in large language models have facilitated the development of unified speech language models (SLMs) capable of supporting multiple speech tasks within a shared architecture. However, tasks such as automatic speech recognition (ASR) and speech emotion recognition (SER) rely on distinct types of information: ASR primarily depends on linguistic content, whereas SER requires the integration of both linguistic and paralinguistic cues. Existing multitask SLMs typically adopt naive para…

Thinking in cocktail party: Chain-of-Thought and reinforcement learning for target speaker automatic speech recognition
Yiru Zhang, Hang Su, Lichun Fan, Zhenbo Luo, Jian Luan
https://arxiv.org/abs/2509.15612

Thinking in cocktail party: Chain-of-Thought and reinforcement learning for target speaker automatic speech recognition
Target Speaker Automatic Speech Recognition (TS-ASR) aims to transcribe the speech of a specified target speaker from multi-speaker mixtures in cocktail party scenarios. Recent advancement of Large Audio-Language Models (LALMs) has already brought some new insights to TS-ASR. However, significant room for optimization remains for the TS-ASR task within the LALMs architecture. While Chain of Thoughts (CoT) and Reinforcement Learning (RL) have proven effective in certain speech tasks, TS-ASR, whi…

From SALAMANDRA to SALAMANDRATA: BSC Submission for WMT25 General Machine Translation Shared Task
Javier Garcia Gilabert, Xixian Liao, Severino Da Dalt, Ella Bohman, Audrey Mash, Francesca De Luca Fornaciari, Irene Baucells, Joan Llop, Miguel Claramunt Argote, Carlos Escolano, Maite Melero
https://arxiv.org/abs/2508.12774

From SALAMANDRA to SALAMANDRATA: BSC Submission for WMT25 General Machine Translation Shared Task
In this paper, we present the SALAMANDRATA family of models, an improved iteration of SALAMANDRA LLMs (Gonzalez-Agirre et al., 2025) specifically trained to achieve strong performance in translation-related tasks for 38 European languages. SALAMANDRATA comes in two scales: 2B and 7B parameters. For both versions, we applied the same training recipe with a first step of continual pre-training on parallel data, and a second step of supervised fine-tuning on high-quality instructions. The BSC subm…

Analysis of Domain Shift across ASR Architectures via TTS-Enabled Separation of Target Domain and Acoustic Conditions
Tina Raissi, Nick Rossenbach, Ralf Schl\"uter
https://arxiv.org/abs/2508.09868

Analysis of Domain Shift across ASR Architectures via TTS-Enabled Separation of Target Domain and Acoustic Conditions
We analyze automatic speech recognition (ASR) modeling choices under domain mismatch, comparing classic modular and novel sequence-to-sequence (seq2seq) architectures. Across the different ASR architectures, we examine a spectrum of modeling choices, including label units, context length, and topology. To isolate language domain effects from acoustic variation, we synthesize target domain audio using a text-to-speech system trained on LibriSpeech. We incorporate target domain n-gram and neural …