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.

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.

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.

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.

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.

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

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.

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.

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.

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.

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

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.

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.

Recently, there has been a significant surge of interest in Self-Attention Networks (SANs) based on the Transformer architecture. This can be attributed to their notable ability for parallelization and their impressive performance across various Natural Language Processing applications. On the other hand, the utilization of large-scale, multi-purpose language models trained through self-supervision is progressively more prevalent, for tasks like speech recognition. In this context, the pre-trained model, which has been trained on extensive speech data, can be fine-tuned for particular downstream tasks like speaker verification. These massive models typically rely on SANs as their foundational architecture. Therefore, studying the potential capabilities and training challenges of such models is of utmost importance for the future generation of speaker verification systems. In this direction, we propose a speaker embedding extractor based on SANs to obtain a discriminative speaker representation given non-fixed length speech utterances. With the advancements suggested in this work, we could achieve up to 41% relative performance improvement in terms of EER compared to the naive SAN which was proposed in our previous work. Moreover, we empirically show the training instability in such architectures in terms of rank collapse and further investigate the potential solutions to alleviate this shortcoming.

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.

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.

Recent advances in Deep Learning (DL) for speaker recognition have improved the performance but are constrained to the need of labels for the background data, which is difficult in practice.In i-vector based speaker recognition, cosine (unsupervised) and PLDA (supervised) are the basic scoring techniques, with a big performance gap between the two.In this thesis we tried to fill this gap without using speaker labels in several ways.We applied Restricted Boltzmann Machine (RBM) vectors for the tasks of speaker clustering and tracking in TV broadcast shows.The experiments on AGORA database show that using this approach we gain a relative improvement of 12% and 11% for speaker clustering and tracking tasks, respectively.We also applied DL techniques in order to increase the discriminative power of i-vectors in speaker verification task, for which we have proposed the use of autoencoder in several ways, i.e., (1) as a pre-training for a Deep Neural Network (DNN), ( 2) as a nearest neighbor autoencoder for i-vectors, (3) as an average pooled nearest neighbor autoencoder.The experiments on VoxCeleb database show that we gain a relative improvement of 21%, 42% and 53%, using the three systems respectively.Finally we also proposed a self-supervised end-to-end speaker verification system.The architecture is based on a Convolutional Neural Network (CNN), trained as a siamese network with multiple branches.From the results we can see that our system shows comparable performance to a supervised baseline.

Artists have always travelled. But aboye all during their youth, searching to complete their apprenticeship. Howcver, our century has change the meaning of thar travel, doing it frequent, not only during formation years, but along all life. The internationalitation of artistic language imposes it

Most state-of-the-art Deep Learning systems for text-independent 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 (MHA) pooling, which extends our previous approach based on Self MHA. An additional self attention layer is added to the pooling layer that summarizes the context vectors produced by MHA 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 MHA, respectively. According to the obtained results, Double MHA 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.