arXiv subjects

Javier Hernando

Publications and source records attributed to Javier Hernando.

At least 19 recordsLinked to original sources

arXiv2026

Disentangling Speaker and Language Effects in Cross-Lingual Speaker Verification for Iberian Languages

Cross-lingual speaker verification (SV) systems typically exhibit performance degradation when enrollment and test utterances are spoken in different languages. However, standard evaluation protocols confound language mismatch with inter-speaker variability, as evaluation is generally performed with different speakers across languages. In this work, we introduce a bilingual same-speaker evaluation set for five Iberian languages, enabling analysis of cross-lingual SV under constant speaker identity. We apply this setup to a HuBERT-based SV system previously shown to exhibit strong language dependence, and analyze results using the Cross-Lingual Transfer Matrix (CLTM) to study pairwise cross-lingual transfer. Our results show that speaker-related variability accounts for part of the observed degradation, but language mismatch remains the main driver of cross-lingual performance loss. These findings provide a more precise characterization of language dependence in cross-lingual SV.

eess.AS↗

arXiv2026

Sequence-Aware Split Heuristic to Mitigate SM Underutilization in FlashAttention-3 Low-Head-Count Decoding

The standard FlashAttention-3 heuristic exhibits a GPU occupancy bottleneck in low-head-count decoding configurations because it disables sequence splitting based on sequence length alone, underutilizing the Streaming Multiprocessors of Hopper GPUs. Our proposed sequence-aware split policy mitigates this by allowing sequence-level parallelism in low-head-count regimes, improving hardware utilization to deliver roughly a 21 to 24% improvement in decoder kernel efficiency on metadata-enabled inference paths, with no observed regressions.

cs.AR↗

arXiv2026

How Attention Shapes Emotion: A Comparative Study of Attention Mechanisms for Speech Emotion Recognition

Speech Emotion Recognition (SER) plays a key role in advancing human-computer interaction. Attention mechanisms have become the dominant approach for modeling emotional speech due to their ability to capture long-range dependencies and emphasize salient information. However, standard self-attention suffers from quadratic computational and memory complexity, limiting its scalability. In this work, we present a systematic benchmark of optimized attention mechanisms for SER, including RetNet, LightNet, GSA, FoX, and KDA. Experiments on both MSP-Podcast benchmark versions show that while standard self-attention achieves the strongest recognition performance across test sets, efficient attention variants dramatically improve scalability, reducing inference latency and memory usage by up to an order of magnitude. These results highlight a critical trade-off between accuracy and efficiency, providing practical insights for designing scalable SER systems.

eess.AS↗

arXiv2026

Quantifying Cross-Lingual Transfer in Paralinguistic Speech Tasks

Paralinguistic speech tasks are often considered relatively language-agnostic, as they rely on extralinguistic acoustic cues rather than lexical content. However, prior studies report performance degradation under cross-lingual conditions, indicating non-negligible language dependence. Still, these studies typically focus on isolated language pairs or task-specific settings, limiting comparability and preventing a systematic assessment of task-level language dependence. We introduce the Cross-Lingual Transfer Matrix (CLTM), a systematic method to quantify cross-lingual interactions between pairs of languages within a given task. We apply the CLTM to two paralinguistic tasks, gender identification and speaker verification, using a multilingual HuBERT-based encoder, to analyze how donor-language data affects target-language performance during fine-tuning. Our results reveal distinct transfer patterns across tasks and languages, reflecting systematic, language-dependent effects.

eess.AS↗

arXiv2026

Bootstrapping Audiovisual Speech Recognition in Zero-AV-Resource Scenarios with Synthetic Visual Data

Audiovisual speech recognition (AVSR) combines acoustic and visual cues to improve transcription robustness under challenging conditions but remains out of reach for most under-resourced languages due to the lack of labeled video corpora for training. We propose a zero-AV-resource AVSR framework that relies on synthetic visual streams generated by lip-syncing static facial images with real audio. We first evaluate synthetic visual augmentation on Spanish benchmarks, then apply it to Catalan, a language with no annotated audiovisual corpora. We synthesize over 700 hours of talking-head video and fine-tune a pre-trained AV-HuBERT model. On a manually annotated Catalan benchmark, our model achieves near state-of-the-art performance with much fewer parameters and training data, outperforms an identically trained audio-only baseline, and preserves multimodal advantages in noise. Scalable synthetic video thus offers a viable substitute for real recordings in zero-AV-resource AVSR.

eess.AS↗

arXiv2025

Listening or Reading? Evaluating Speech Awareness in Chain-of-Thought Speech-to-Text Translation

Speech-to-Text Translation (S2TT) systems built from Automatic Speech Recognition (ASR) and Text-to-Text Translation (T2TT) modules face two major limitations: error propagation and the inability to exploit prosodic or other acoustic cues. Chain-of-Thought (CoT) prompting has recently been introduced, with the expectation that jointly accessing speech and transcription will overcome these issues. Analyzing CoT through attribution methods, robustness evaluations with corrupted transcripts, and prosody-awareness, we find that it largely mirrors cascaded behavior, relying mainly on transcripts while barely leveraging speech. Simple training interventions, such as adding Direct S2TT data or noisy transcript injection, enhance robustness and increase speech attribution. These findings challenge the assumed advantages of CoT and highlight the need for architectures that explicitly integrate acoustic information into translation.

cs.CL↗

arXiv2025

Revisiting Direct Speech-to-Text Translation with Speech LLMs: Better Scaling than CoT Prompting?

Recent work on Speech-to-Text Translation (S2TT) has focused on LLM-based models, introducing the increasingly adopted Chain-of-Thought (CoT) prompting, where the model is guided to first transcribe the speech and then translate it. CoT typically outperforms direct prompting primarily because it can exploit abundant Automatic Speech Recognition (ASR) and Text-to-Text Translation (T2TT) datasets to explicitly model its steps. In this paper, we systematically compare CoT and Direct prompting under increasing amounts of S2TT data. To this end, we pseudo-label an ASR corpus by translating its transcriptions into six European languages, and train LLM-based S2TT systems with both prompting strategies at different data scales. Our results show that Direct improves more consistently as the amount of data increases, suggesting that it may become a more effective approach as larger S2TT resources are created.

cs.CL↗

arXiv2025

Optimizing ASR for Catalan-Spanish Code-Switching: A Comparative Analysis of Methodologies

Code-switching (CS), the alternating use of two or more languages, challenges automatic speech recognition (ASR) due to scarce training data and linguistic similarities. The lack of dedicated CS datasets limits ASR performance, as most models rely on monolingual or mixed-language corpora that fail to reflect real-world CS patterns. This issue is critical in multilingual societies where CS occurs in informal and formal settings. A key example is Catalan-Spanish CS, widely used in media and parliamentary speeches. In this work, we improve ASR for Catalan-Spanish CS by exploring three strategies: (1) generating synthetic CS data, (2) concatenating monolingual audio, and (3) leveraging real CS data with language tokens. We extract CS data from Catalan speech corpora and fine-tune OpenAI's Whisper models, making them available on Hugging Face. Results show that combining a modest amount of synthetic CS data with the dominant language token yields the best transcription performance.

cs.CL↗

arXiv2025

Speech-to-Text Translation with Phoneme-Augmented CoT: Enhancing Cross-Lingual Transfer in Low-Resource Scenarios

We propose a Speech-to-Text Translation (S2TT) approach that integrates phoneme representations into a Chain-of-Thought (CoT) framework to improve translation in low-resource and zero-resource settings. By introducing phoneme recognition as an intermediate step, we enhance cross-lingual transfer, enabling translation even for languages with no labeled speech data. Our system builds on a multilingual LLM, which we extend to process speech and phonemes. Training follows a curriculum learning strategy that progressively introduces more complex tasks. Experiments on multilingual S2TT benchmarks show that phoneme-augmented CoT improves translation quality in low-resource conditions and enables zero-resource translation, while slightly impacting high-resource performance. Despite this trade-off, our findings demonstrate that phoneme-based CoT is a promising step toward making S2TT more accessible across diverse languages.

cs.CL↗

arXiv2025

Breaking Language Barriers in Visual Language Models via Multilingual Textual Regularization

Rapid advancements in Visual Language Models (VLMs) have transformed multimodal understanding but are often constrained by generating English responses regardless of the input language. This phenomenon has been termed as Image-induced Fidelity Loss (IFL) and stems from limited multimodal multilingual training data. To address this, we propose a continuous multilingual integration strategy that injects text-only multilingual data during visual instruction tuning, preserving the language model's original multilingual capabilities. Extensive evaluations demonstrate that our approach significantly improves linguistic fidelity across languages without degradation in visual performance. We also explore model merging, which improves language fidelity but comes at the cost of visual performance. In contrast, our core method achieves robust multilingual alignment without trade-offs, offering a scalable and effective path to mitigating IFL for global VLM adoption.

cs.CV↗

arXiv2025

Mass-Editing Memory with Attention in Transformers: A cross-lingual exploration of knowledge

Recent research has explored methods for updating and modifying factual knowledge in large language models, often focusing on specific multi-layer perceptron blocks. This study expands on this work by examining the effectiveness of existing knowledge editing methods across languages and delving into the role of attention mechanisms in this process. Drawing from the insights gained, we propose Mass-Editing Memory with Attention in Transformers (MEMAT), a method that achieves significant improvements in all metrics while requiring minimal parameter modifications. MEMAT delivers a remarkable 10% increase in magnitude metrics, benefits languages not included in the training data and also demonstrates a high degree of portability. Our code and data are at https://github.com/dtamayo-nlp/MEMAT.

cs.CL↗

arXiv2025

Work-Efficient Parallel Non-Maximum Suppression Kernels

In the context of object detection, sliding-window classifiers and single-shot Convolutional Neural Network (CNN) meta-architectures typically yield multiple overlapping candidate windows with similar high scores around the true location of a particular object. Non-Maximum Suppression (NMS) is the process of selecting a single representative candidate within this cluster of detections, so as to obtain a unique detection per object appearing on a given picture. In this paper, we present a highly scalable NMS algorithm for embedded GPU architectures that is designed from scratch to handle workloads featuring thousands of simultaneous detections on a given picture. Our kernels are directly applicable to other sequential NMS algorithms such as FeatureNMS, Soft-NMS or AdaptiveNMS that share the inner workings of the classic greedy NMS method. The obtained performance results show that our parallel NMS algorithm is capable of clustering 1024 simultaneous detected objects per frame in roughly 1 ms on both NVIDIA Tegra X1 and NVIDIA Tegra X2 on-die GPUs, while taking 2 ms on NVIDIA Tegra K1. Furthermore, our proposed parallel greedy NMS algorithm yields a 14x-40x speed up when compared to state-of-the-art NMS methods that require learning a CNN from annotated data.

cs.CV↗

arXiv2025

Language Modelling for Speaker Diarization in Telephonic Interviews

The aim of this paper is to investigate the benefit of combining both language and acoustic modelling for speaker diarization. Although conventional systems only use acoustic features, in some scenarios linguistic data contain high discriminative speaker information, even more reliable than the acoustic ones. In this study we analyze how an appropriate fusion of both kind of features is able to obtain good results in these cases. The proposed system is based on an iterative algorithm where a LSTM network is used as a speaker classifier. The network is fed with character-level word embeddings and a GMM based acoustic score created with the output labels from previous iterations. The presented algorithm has been evaluated in a Call-Center database, which is composed of telephone interview audios. The combination of acoustic features and linguistic content shows a 84.29% improvement in terms of a word-level DER as compared to a HMM/VB baseline system. The results of this study confirms that linguistic content can be efficiently used for some speaker recognition tasks.

eess.AS↗

arXiv2024

On the Use of Audio to Improve Dialogue Policies

With the significant progress of speech technologies, spoken goal-oriented dialogue systems are becoming increasingly popular. One of the main modules of a dialogue system is typically the dialogue policy, which is responsible for determining system actions. This component usually relies only on audio transcriptions, being strongly dependent on their quality and ignoring very important extralinguistic information embedded in the user's speech. In this paper, we propose new architectures to add audio information by combining speech and text embeddings using a Double Multi-Head Attention component. Our experiments show that audio embedding-aware dialogue policies outperform text-based ones, particularly in noisy transcription scenarios, and that how text and audio embeddings are combined is crucial to improve performance. We obtained a 9.8% relative improvement in the User Request Score compared to an only-text-based dialogue system on the DSTC2 dataset.

eess.AS↗

arXiv2024

BSC-UPC at EmoSPeech-IberLEF2024: Attention Pooling for Emotion Recognition

The domain of speech emotion recognition (SER) has persistently been a frontier within the landscape of machine learning. It is an active field that has been revolutionized in the last few decades and whose implementations are remarkable in multiple applications that could affect daily life. Consequently, the Iberian Languages Evaluation Forum (IberLEF) of 2024 held a competitive challenge to leverage the SER results with a Spanish corpus. This paper presents the approach followed with the goal of participating in this competition. The main architecture consists of different pre-trained speech and text models to extract features from both modalities, utilizing an attention pooling mechanism. The proposed system has achieved the first position in the challenge with an 86.69% in Macro F1-Score.

eess.AS↗

arXiv2024

Double Multi-Head Attention Multimodal System for Odyssey 2024 Speech Emotion Recognition Challenge

As computer-based applications are becoming more integrated into our daily lives, the importance of Speech Emotion Recognition (SER) has increased significantly. Promoting research with innovative approaches in SER, the Odyssey 2024 Speech Emotion Recognition Challenge was organized as part of the Odyssey 2024 Speaker and Language Recognition Workshop. In this paper we describe the Double Multi-Head Attention Multimodal System developed for this challenge. Pre-trained self-supervised models were used to extract informative acoustic and text features. An early fusion strategy was adopted, where a Multi-Head Attention layer transforms these mixed features into complementary contextualized representations. A second attention mechanism is then applied to pool these representations into an utterance-level vector. Our proposed system achieved the third position in the categorical task ranking with a 34.41% Macro-F1 score, where 31 teams participated in total.

eess.AS↗

arXiv2024

Speaker Characterization by means of Attention Pooling

State-of-the-art Deep Learning systems for speaker verification are commonly based on speaker embedding extractors. These architectures are usually composed of a feature extractor front-end together with a pooling layer to encode variable-length utterances into fixed-length speaker vectors. The authors have recently proposed the use of a Double Multi-Head Self-Attention pooling for speaker recognition, placed between a CNN-based front-end and a set of fully connected layers. This has shown to be an excellent approach to efficiently select the most relevant features captured by the front-end from the speech signal. In this paper we show excellent experimental results by adapting this architecture to other different speaker characterization tasks, such as emotion recognition, sex classification and COVID-19 detection.

eess.AS↗

arXiv2021

Double Multi-Head Attention for Speaker Verification

Most state-of-the-art Deep Learning systems for speaker verification are based on speaker embedding extractors. These architectures are commonly composed of a feature extractor front-end together with a pooling layer to encode variable-length utterances into fixed-length speaker vectors. In this paper we present Double Multi-Head Attention pooling, which extends our previous approach based on Self Multi-Head Attention. An additional self attention layer is added to the pooling layer that summarizes the context vectors produced by Multi-Head Attention into a unique speaker representation. This method enhances the pooling mechanism by giving weights to the information captured for each head and it results in creating more discriminative speaker embeddings. We have evaluated our approach with the VoxCeleb2 dataset. Our results show 6.09% and 5.23% relative improvement in terms of EER compared to Self Attention pooling and Self Multi-Head Attention, respectively. According to the obtained results, Double Multi-Head Attention has shown to be an excellent approach to efficiently select the most relevant features captured by the CNN-based front-ends from the speech signal.

eess.AS↗