This work focuses on the problem of detecting fake audio clips. To improve current audio spoofing detection models, we propose a selection of multiple audio augmentations spe-cially designed to resemble audio spoofing attacks. These augmentations are experimentally found to be very useful and using them achieves a notable performance of 2.8% EER on the ASVspoof 2019 challenge evaluation set. Unlike the widely employed acoustic features, in this paper we explore the use of Mel-spectrogram image features and employ vari-ous audio codecs to achieve robustness to codec and transmission channel variability present in the ASVspoof2021 Evalu-ation set. To better handle spectral information, crucial to de-tect spoofing, we adopt the WaveletCNN and VGG16 archi-tectures which outperform all baselines. Finally, we find that robustness of countermeasure systems degrades dramatically when provided with speech samples degraded through VoIP network transmission or mismatching audio compression.

With the increasing development of advanced unmanned aerial vehicles (UAVs), communication between operators and these intelligent systems is becoming more stressful. For the safety of UAV flights, automatic psychological stress detection is becoming a key research topic for successful missions. Stress can be reliably estimated via some biological markers which are not appropriate in many cases of human-machine-interaction setups. In this paper, we propose a non-intrusive deep learning-based stress level estimation approach. The goal is to identify the region where the operator's emotional state projects in the space defined by the latent dimensional emotions of arousal and valence since the stress region is well delimited in this space. The proposed multimodal approach uses sequential temporal CNN and LSTM with an Attention Weighted Average layer in the vision modality. As a second modality, we investigate local and global descriptors such as Mel-frequency cepstral coefficients, i-vector embeddings as well as Fisher-vector encodings. The multimodal-fusion approach uses a strategy referred to as "late-fusion" that involves the combination of unimodal model outputs as inputs of the decision engine. Since we have to deal with more naturalistic behavior in operator-machine interaction contexts, the One minute Gradual Emotion Challenge dataset was used for predictive model validation.

Over the recent years, various self-supervised embedding learning methods for deep speaker verification were proposed. The performance of the self-supervised learning framework highly depends on the data augmentation technique, but due to the sensitive nature of speaker information within the speech signal, most speaker embedding training relies on simple augmentations such as additive noise or simulated reverberation. Thus while the conventional self-supervised speaker embedding systems can yield minimum within-utterance variability, their capability to generalize to out-of-set utterance is limited. In order to alleviate this problem, we investigate the utilization of the instance mix (i-mix) regularization for training a self-supervised speaker embedding system. Moreover, we propose a new mixup strategy that applies i-mix on the latent space, instead of the raw acoustic feature domain. In this paper, the i-mix and the proposed l-mix strategy were incorporated into the self-supervised angular prototypical and softmax-based objective functions and were evaluated on the VoxCeleb dataset. From the experimental results, we observe that the self-supervised embedding network can benefit greatly from the i-mix and l-mix strategies in terms of training stability and speaker verification performance.

Multimodal emotion recognition has immense potential for the comprehensive assessment of human emotions, utilizing multiple modalities that often exhibit complementary relationships. In video-based emotion recognition, audio and visual modalities have emerged as prominent contact-free channels, widely explored in existing literature. Current approaches typically employ cross-modal attention mechanisms between audio and visual modalities, assuming a constant state of complementarity. However, this assumption may not always hold true, as non-complementary relationships can also manifest, undermining the efficacy of cross-modal feature integration and thereby diminishing the quality of audio-visual feature representations. To tackle this problem, we introduce a novel Incongruity-Aware Cross-Attention (IACA) model, capable of harnessing the benefits of robust complementary relationships while efficiently managing non-complementary scenarios. Specifically, our approach incorporates a two-stage gating mechanism designed to adaptively select semantic features, thereby effectively capturing the inter-modal associations. Additionally, the proposed model demonstrates an ability to mitigate the adverse effects of severely corrupted or missing modalities. We rigorously evaluate the performance of the proposed model through extensive experiments conducted on the challenging RECOLA and Aff-Wild2 datasets. The results underscore the efficacy of our approach, as it outperforms state-of-the-art methods by adeptly capturing inter-modal relationships and minimizing the influence of missing or heavily corrupted modalities. Furthermore, we show that the proposed model is compatible with various cross-modal attention variants, consistently improving performance on both datasets.