Nicolas Farrugia

Ilyass Moummad, Nicolas Farrugia

Methods based on supervised learning using annotations in an end-to-end fashion have been the state-of-the-art for classification problems. However, they may be limited in their generalization capability, especially in the low data regime. In this study, we address this issue using supervised contrastive learning combined with available metadata to solve multiple pretext tasks that learn a good representation of data. We apply our approach on respiratory sound classification. This task is suited for this setting as demographic information such as sex and age are correlated with presence of lung diseases, and learning a system that implicitly encode this information may better detect anomalies. Supervised contrastive learning is a paradigm that learns similar representations to samples sharing the same class labels and dissimilar representations to samples with different class labels. The feature extractor learned using this paradigm extract useful features from the data, and we show that it outperforms cross-entropy in classifying respiratory anomalies in two different datasets. We also show that learning representations using only metadata, without class labels, obtains similar performance as using cross entropy with those labels only. In addition, when combining class labels with metadata using multiple supervised contrastive learning, an extension of supervised contrastive learning solving an additional task of grouping patients within the same sex and age group, more informative features are learned. This work suggests the potential of using multiple metadata sources in supervised contrastive settings, in particular in settings with class imbalance and few data. Our code is released at https://github.com/ilyassmoummad/scl_icbhi2017

Ilyass Moummad, Romain Serizel, Nicolas Farrugia

Bioacoustic sound event detection allows for better understanding of animal behavior and for better monitoring biodiversity using audio. Deep learning systems can help achieve this goal, however it is difficult to acquire sufficient annotated data to train these systems from scratch. To address this limitation, the Detection and Classification of Acoustic Scenes and Events (DCASE) community has recasted the problem within the framework of few-shot learning and organize an annual challenge for learning to detect animal sounds from only five annotated examples. In this work, we regularize supervised contrastive pre-training to learn features that can transfer well on new target tasks with animal sounds unseen during training, achieving a high F-score of 61.52%(0.48) when no feature adaptation is applied, and an F-score of 68.19%(0.75) when we further adapt the learned features for each new target task. This work aims to lower the entry bar to few-shot bioacoustic sound event detection by proposing a simple and yet effective framework for this task, by also providing open-source code.

Yassine El Ouahidi, Vincent Gripon, Bastien Pasdeloup et al.

We propose EEG-SimpleConv, a straightforward 1D convolutional neural network for Motor Imagery decoding in BCI. Our main motivation is to propose a simple and performing baseline to compare to, using only very standard ingredients from the literature. We evaluate its performance on four EEG Motor Imagery datasets, including simulated online setups, and compare it to recent Deep Learning and Machine Learning approaches. EEG-SimpleConv is at least as good or far more efficient than other approaches, showing strong knowledge-transfer capabilities across subjects, at the cost of a low inference time. We advocate that using off-the-shelf ingredients rather than coming with ad-hoc solutions can significantly help the adoption of Deep Learning approaches for BCI. We make the code of the models and the experiments accessible.

Yassine El Ouahidi, Lucas Drumetz, Giulia Lioi et al.

BCI Motor Imagery datasets usually are small and have different electrodes setups. When training a Deep Neural Network, one may want to capitalize on all these datasets to increase the amount of data available and hence obtain good generalization results. To this end, we introduce a spatial graph signal interpolation technique, that allows to interpolate efficiently multiple electrodes. We conduct a set of experiments with five BCI Motor Imagery datasets comparing the proposed interpolation with spherical splines interpolation. We believe that this work provides novel ideas on how to leverage graphs to interpolate electrodes and on how to homogenize multiple datasets.

Matteo Zambra, Dorian Cazau, Nicolas Farrugia et al.

Wind speed retrieval at sea surface is of primary importance for scientific and operational applications. Besides weather models, in-situ measurements and remote sensing technologies, especially satellite sensors, provide complementary means to monitor wind speed. As sea surface winds produce sounds that propagate underwater, underwater acoustics recordings can also deliver fine-grained wind-related information. Whereas model-driven schemes, especially data assimilation approaches, are the state-of-the-art schemes to address inverse problems in geoscience, machine learning techniques become more and more appealing to fully exploit the potential of observation datasets. Here, we introduce a deep learning approach for the retrieval of wind speed time series from underwater acoustics possibly complemented by other data sources such as weather model reanalyses. Our approach bridges data assimilation and learning-based frameworks to benefit both from prior physical knowledge and computational efficiency. Numerical experiments on real data demonstrate that we outperform the state-of-the-art data-driven methods with a relative gain up to 16% in terms of RMSE. Interestingly, these results support the relevance of the time dynamics of underwater acoustic data to better inform the time evolution of wind speed. They also show that multimodal data, here underwater acoustics data combined with ECMWF reanalysis data, may further improve the reconstruction performance, including the robustness with respect to missing underwater acoustics data.

Yassine El Ouahidi, Hugo Tessier, Giulia Lioi et al.

Graph Signal Processing is a promising framework to manipulate brain signals as it allows to encompass the spatial dependencies between the activity in regions of interest in the brain. In this work, we are interested in better understanding what are the graph frequencies that are the most useful to decode fMRI signals. To this end, we introduce a deep learning architecture and adapt a pruning methodology to automatically identify such frequencies. We experiment with various datasets, architectures and graphs, and show that low graph frequencies are consistently identified as the most important for fMRI decoding, with a stronger contribution for the functional graph over the structural one. We believe that this work provides novel insights on how graph-based methods can be deployed to increase fMRI decoding accuracy and interpretability.

Ilyass Moummad, Romain Serizel, Nicolas Farrugia

Self-supervised learning (SSL) in audio holds significant potential across various domains, particularly in situations where abundant, unlabeled data is readily available at no cost. This is pertinent in bioacoustics, where biologists routinely collect extensive sound datasets from the natural environment. In this study, we demonstrate that SSL is capable of acquiring meaningful representations of bird sounds from audio recordings without the need for annotations. Our experiments showcase that these learned representations exhibit the capacity to generalize to new bird species in few-shot learning (FSL) scenarios. Additionally, we show that selecting windows with high bird activation for self-supervised learning, using a pretrained audio neural network, significantly enhances the quality of the learned representations.

Yassine El Ouahidi, Giulia Lioi, Nicolas Farrugia et al.

In the context of Brain-Computer Interfaces, we propose an adaptive method that reaches offline performance level while being usable online without requiring supervision. Interestingly, our method does not require retraining the model, as it consists in using a frozen efficient deep learning backbone while continuously realigning data, both at input and latent spaces, based on streaming observations. We demonstrate its efficiency for Motor Imagery brain decoding from electroencephalography data, considering challenging cross-subject scenarios. For reproducibility, we share the code of our experiments.

Matteo Zambra, Nicolas Farrugia, Dorian Cazau et al.

Wind speed at sea surface is a key quantity for a variety of scientific applications and human activities. Due to the non-linearity of the phenomenon, a complete description of such variable is made infeasible on both the small scale and large spatial extents. Methods relying on Data Assimilation techniques, despite being the state-of-the-art for Numerical Weather Prediction, can not provide the reconstructions with a spatial resolution that can compete with satellite imagery. In this work we propose a framework based on Variational Data Assimilation and Deep Learning concepts. This framework is applied to recover rich-in-time, high-resolution information on sea surface wind speed. We design our experiments using synthetic wind data and different sampling schemes for high-resolution and low-resolution versions of original data to emulate the real-world scenario of spatio-temporally heterogeneous observations. Extensive numerical experiments are performed to assess systematically the impact of low and high-resolution wind fields and in-situ observations on the model reconstruction performance. We show that in-situ observations with richer temporal resolution represent an added value in terms of the model reconstruction performance. We show how a multi-modal approach, that explicitly informs the model about the heterogeneity of the available observations, can improve the reconstruction task by exploiting the complementary information in spatial and local point-wise data. To conclude, we propose an analysis to test the robustness of the chosen framework against phase delay and amplitude biases in low-resolution data and against interruptions of in-situ observations supply at evaluation time

Ilyass Moummad, Romain Serizel, Nicolas Farrugia

Deep learning has been widely used recently for sound event detection and classification. Its success is linked to the availability of sufficiently large datasets, possibly with corresponding annotations when supervised learning is considered. In bioacoustic applications, most tasks come with few labelled training data, because annotating long recordings is time consuming and costly. Therefore supervised learning is not the best suited approach to solve bioacoustic tasks. The bioacoustic community recasted the problem of sound event detection within the framework of few-shot learning, i.e. training a system with only few labeled examples. The few-shot bioacoustic sound event detection task in the DCASE challenge focuses on detecting events in long audio recordings given only five annotated examples for each class of interest. In this paper, we show that learning a rich feature extractor from scratch can be achieved by leveraging data augmentation using a supervised contrastive learning framework. We highlight the ability of this framework to transfer well for five-shot event detection on previously unseen classes in the training data. We obtain an F-score of 63.46\% on the validation set and 42.7\% on the test set, ranking second in the DCASE challenge. We provide an ablation study for the critical choices of data augmentation techniques as well as for the learning strategy applied on the training set.

Myriam Bontonou, Nicolas Farrugia, Vincent Gripon

It is very common to face classification problems where the number of available labeled samples is small compared to their dimension. These conditions are likely to cause underdetermined settings, with high risk of overfitting. To improve the generalization ability of trained classifiers, common solutions include using priors about the data distribution. Among many options, data structure priors, often represented through graphs, are increasingly popular in the field. In this paper, we introduce a generic model where observed class signals are supposed to be deteriorated with two sources of noise, one independent of the underlying graph structure and isotropic, and the other colored by a known graph operator. Under this model, we derive an optimal methodology to classify such signals. Interestingly, this methodology includes a single parameter, making it particularly suitable for cases where available data is scarce. Using various real datasets, we showcase the ability of the proposed model to be implemented in real world scenarios, resulting in increased generalization accuracy compared to popular alternatives.

Myriam Bontonou, Giulia Lioi, Nicolas Farrugia et al.

Few-shot learning addresses problems for which a limited number of training examples are available. So far, the field has been mostly driven by applications in computer vision. Here, we are interested in adapting recently introduced few-shot methods to solve problems dealing with neuroimaging data, a promising application field. To this end, we create a neuroimaging benchmark dataset for few-shot learning and compare multiple learning paradigms, including meta-learning, as well as various backbone networks. Our experiments show that few-shot methods are able to efficiently decode brain signals using few examples, which paves the way for a number of applications in clinical and cognitive neuroscience, such as identifying biomarkers from brain scans or understanding the generalization of brain representations across a wide range of cognitive tasks.

Ghouthi Boukli Hacene, Vincent Gripon, Nicolas Farrugia et al.

In this paper, we tackle the problem of incrementally learning a classifier, one example at a time, directly on chip. To this end, we propose an efficient hardware implementation of a recently introduced incremental learning procedure that achieves state-of-the-art performance by combining transfer learning with majority votes and quantization techniques. The proposed design is able to accommodate for both new examples and new classes directly on the chip. We detail the hardware implementation of the method (implemented on FPGA target) and show it requires limited resources while providing a significant acceleration compared to using a CPU.

Myriam Bontonou, Carlos Lassance, Vincent Gripon et al.

Predicting the future of Graph-supported Time Series (GTS) is a key challenge in many domains, such as climate monitoring, finance or neuroimaging. Yet it is a highly difficult problem as it requires to account jointly for time and graph (spatial) dependencies. To simplify this process, it is common to use a two-step procedure in which spatial and time dependencies are dealt with separately. In this paper, we are interested in comparing various linear spatial representations, namely structure-based ones and data-driven ones, in terms of how they help predict the future of GTS. To that end, we perform experiments with various datasets including spontaneous brain activity and raw videos.

Ghouthi Boukli Hacene, Vincent Gripon, Nicolas Farrugia et al.

Deep learning-based methods have reached state of the art performances, relying on large quantity of available data and computational power. Such methods still remain highly inappropriate when facing a major open machine learning problem, which consists of learning incrementally new classes and examples over time. Combining the outstanding performances of Deep Neural Networks (DNNs) with the flexibility of incremental learning techniques is a promising venue of research. In this contribution, we introduce Transfer Incremental Learning using Data Augmentation (TILDA). TILDA is based on pre-trained DNNs as feature extractor, robust selection of feature vectors in subspaces using a nearest-class-mean based technique, majority votes and data augmentation at both the training and the prediction stages. Experiments on challenging vision datasets demonstrate the ability of the proposed method for low complexity incremental learning, while achieving significantly better accuracy than existing incremental counterparts.

Mathilde Ménoret, Nicolas Farrugia, Bastien Pasdeloup et al.

Graph Signal Processing (GSP) is a promising framework to analyze multi-dimensional neuroimaging datasets, while taking into account both the spatial and functional dependencies between brain signals. In the present work, we apply dimensionality reduction techniques based on graph representations of the brain to decode brain activity from real and simulated fMRI datasets. We introduce seven graphs obtained from a) geometric structure and/or b) functional connectivity between brain areas at rest, and compare them when performing dimension reduction for classification. We show that mixed graphs using both a) and b) offer the best performance. We also show that graph sampling methods perform better than classical dimension reduction including Principal Component Analysis (PCA) and Independent Component Analysis (ICA).