Search

Language models can achieve high accuracy on natural language tasks such as NLI, but performance suffers on manually created adversarial examples. We investigate the performance of a language model trained on the Stanford Natural Language Inference (SNLI) corpus on a manually created adversarial test set. We then improve the model's performance by fine tuning the model on a small, manually created adversarial training set, designed to help the language model to learn to differentiate between similar words and phrases in the data. We show an increase in accuracy on the adversarial test set (+ 13%) while still maintaining good performance on the original NLI task. We also show an increase in accuracy from 91.2% to 92.9% on the most similar contradictions in the SNLI test set (as judged by cosine similarity).

In the past few years, deep learning algorithms have been widely used for cardiac image segmentation. However, most of these architectures rely on convolutions that hardly model long-range dependencies, limiting their ability to extract contextual information. In order to tackle this issue, this article introduces the Swin Filtering Block network (SFB-net) which takes advantage of both conventional and swin transformer layers. The former are used to introduce spatial attention at the bottom of the network, while the latter are applied to focus on high level semantically rich features between the encoder and decoder. An average Dice score of 92.4 was achieved on the ACDC dataset. To the best of our knowledge, this result outperforms any other work on this dataset. The average Dice score of 87.99 obtained on the M\&M's dataset demonstrates that the proposed method generalizes well to data from different vendors and centres.

This paper introduces a new hybrid descriptor for 3D point matching and point cloud registration, combining local geometrical properties and learning-based feature propagation for each point's neighborhood structure description. The proposed architecture first extracts prior geometrical information by computing each point's planarity, anisotropy, and omnivariance using a Principal Components Analysis (PCA). This prior information is completed by a descriptor based on the normal vectors estimated thanks to constructing a neighborhood based on triangles. The final geometrical descriptor is propagated between the points using local graph convolutions and attention mechanisms. The new feature extractor is evaluated on ModelNet40, Bunny Stanford dataset, KITTI and MVP (Multi-View Partial)-RG for point cloud registration and shows interesting results, particularly on noisy and low overlapping point clouds.

Data produced by resting-state functional Magnetic Resonance Imaging are widely used to infer brain functional connectivity networks. Such networks correlate neural signals to connect brain regions, which consist in groups of dependent voxels. Previous work has focused on aggregating data across voxels within predefined regions. However, the presence of within-region correlations has noticeable impacts on inter-regional correlation detection, and thus edge identification. To alleviate them, we propose to leverage techniques from the large-scale correlation screening literature, and derive simple and practical characterizations of the mean number of correlation discoveries that flexibly incorporate intra-regional dependence structures. A connectivity network inference framework is then presented. First, inter-regional correlation distributions are estimated. Then, correlation thresholds that can be tailored to one's application are constructed for each edge. Finally, the proposed framework is implemented on synthetic and real-world datasets. This novel approach for handling arbitrary intra-regional correlation is shown to limit false positives while improving true positive rates.

Airborne hyperspectral images can be used to map the land cover in large urban areas, thanks to their very high spatial and spectral resolutions on a wide spectral domain. While the spectral dimension of hyperspectral images is highly informative of the chemical composition of the land surface, the use of state-of-the-art machine learning algorithms to map the land cover has been dramatically limited by the availability of training data. To cope with the scarcity of annotations, semi-supervised and self-supervised techniques have lately raised a lot of interest in the community. Yet, the publicly available hyperspectral data sets commonly used to benchmark machine learning models are not totally suited to evaluate their generalization performances due to one or several of the following properties: a limited geographical coverage (which does not reflect the spectral diversity in metropolitan areas), a small number of land cover classes and a lack of appropriate standard train / test splits for semi-supervised and self-supervised learning. Therefore, we release in this paper the Toulouse Hyperspectral Data Set that stands out from other data sets in the above-mentioned respects in order to meet key issues in spectral representation learning and classification over large-scale hyperspectral images with very few labeled pixels. Besides, we discuss and experiment self-supervised techniques for spectral representation learning, including the Masked Autoencoder, and establish a baseline for pixel-wise classification achieving 85% overall accuracy and 77% F1 score. The Toulouse Hyperspectral Data Set and our code are publicly available at https://www.toulouse-hyperspectral-data-set.com and https://www.github.com/Romain3Ch216/tlse-experiments, respectively., Comment: 17 pages, 13 figures

Creating dense and shallow nitrogen vacancy (NV) ensembles with good spin properties, is a prerequisite for developing diamond-based quantum sensors exhibiting better performance. Ion implantation is a key enabling tool for precisely controlling spatial localisation and density of NV colour centres in diamond. However, it suffers from a low creation yield, while higher ion fluences significantly damage the crystal lattice. In this work, we realize N2 ion implantation in the 30 to 40 keV range at high temperatures. At 800 C, NV ensemble photoluminescence emission is three to four times higher than room temperature implanted films, while narrow electron spin resonance linewidths of 1.5 MHz, comparable to well established implantation techniques are obtained. In addition, we found that ion fluences above 2E14 ions per cm2 can be used without graphitization of the diamond film, in contrast to room temperature implantation. This study opens promising perspectives in optimizing diamond films with implanted NV ensembles that could be integrated into quantum sensing devices., Comment: 10 pages, 1 table, 4 figures

During a conversation, individuals take turns speaking and engage in exchanges, which can occur smoothly or involve interruptions. Listeners have various ways of participating, such as displaying backchannels, signalling the aim to take a turn, waiting for the speaker to yield the floor, or even interrupting and taking over the conversation. These exchanges are commonplace in natural interactions. To create realistic and engaging interactions between human participants and embodied conversational agents (ECAs), it is crucial to equip virtual agents with the ability to manage these exchanges. This includes being able to initiate or respond to signals from the human user. In order to achieve this, we annotate, analyze and characterize these exchanges in human-human conversations. In this paper, we present an analysis of multimodal features, with a focus on prosodic features such as pitch (F0) and loudness, as well as facial expressions, to describe different types of exchanges.

Artificial neural networks are prone to being fooled by carefully perturbed inputs which cause an egregious misclassification. These \textit{adversarial} attacks have been the focus of extensive research. Likewise, there has been an abundance of research in ways to detect and defend against them. We introduce a novel approach of detection and interpretation of adversarial attacks from a graph perspective. For an input image, we compute an associated sparse graph using the layer-wise relevance propagation algorithm \cite{bach15}. Specifically, we only keep edges of the neural network with the highest relevance values. Three quantities are then computed from the graph which are then compared against those computed from the training set. The result of the comparison is a classification of the image as benign or adversarial. To make the comparison, two classification methods are introduced: 1) an explicit formula based on Wasserstein distance applied to the degree of node and 2) a logistic regression. Both classification methods produce strong results which lead us to believe that a graph-based interpretation of adversarial attacks is valuable., Comment: published in Neural Networks

Socially Interactive Agents (SIAs) are physical or virtual embodied agents that display similar behavior as human multimodal behavior. Modeling SIAs' non-verbal behavior, such as speech and facial gestures, has always been a challenging task, given that a SIA can take the role of a speaker or a listener. A SIA must emit appropriate behavior adapted to its own speech, its previous behaviors (intra-personal), and the User's behaviors (inter-personal) for both roles. We propose AMII, a novel approach to synthesize adaptive facial gestures for SIAs while interacting with Users and acting interchangeably as a speaker or as a listener. AMII is characterized by modality memory encoding schema - where modality corresponds to either speech or facial gestures - and makes use of attention mechanisms to capture the intra-personal and inter-personal relationships. We validate our approach by conducting objective evaluations and comparing it with the state-of-the-art approaches., Comment: 8 pages, 1 figure

The LHCb upgrade represents a major change of the experiment. The detectors have been almost completely renewed to allow running at an instantaneous luminosity five times larger than that of the previous running periods. Readout of all detectors into an all-software trigger is central to the new design, facilitating the reconstruction of events at the maximum LHC interaction rate, and their selection in real time. The experiment's tracking system has been completely upgraded with a new pixel vertex detector, a silicon tracker upstream of the dipole magnet and three scintillating fibre tracking stations downstream of the magnet. The whole photon detection system of the RICH detectors has been renewed and the readout electronics of the calorimeter and muon systems have been fully overhauled. The first stage of the all-software trigger is implemented on a GPU farm. The output of the trigger provides a combination of totally reconstructed physics objects, such as tracks and vertices, ready for final analysis, and of entire events which need further offline reprocessing. This scheme required a complete revision of the computing model and rewriting of the experiment's software., Comment: All figures and tables, along with any supplementary material and additional information, are available at http://lhcbproject.web.cern.ch/lhcbproject/Publications/LHCbProjectPublic/LHCb-DP-2022-002.html (LHCb public pages)

This paper introduces a new method for 3D point cloud registration based on deep learning. The architecture is composed of three distinct blocs: (i) an encoder composed of a convolutional graph-based descriptor that encodes the immediate neighbourhood of each point and an attention mechanism that encodes the variations of the surface normals. Such descriptors are refined by highlighting attention between the points of the same set and then between the points of the two sets. (ii) a matching process that estimates a matrix of correspondences using the Sinkhorn algorithm. (iii) Finally, the rigid transformation between the two point clouds is calculated by RANSAC using the Kc best scores from the correspondence matrix. We conduct experiments on the ModelNet40 dataset, and our proposed architecture shows very promising results, outperforming state-of-the-art methods in most of the simulated configurations, including partial overlap and data augmentation with Gaussian noise., Comment: 8 pages

A novel non-parametric estimator of the correlation between grouped measurements of a quantity is proposed in the presence of noise. This work is primarily motivated by functional brain network construction from fMRI data, where brain regions correspond to groups of spatial units, and correlation between region pairs defines the network. The challenge resides in the fact that both noise and intra-regional correlation lead to inconsistent inter-regional correlation estimation using classical approaches. While some existing methods handle either one of these issues, no non-parametric approaches tackle both simultaneously. To address this problem, we propose a trade-off between two procedures: correlating regional averages, which is not robust to intra-regional correlation; and averaging pairwise inter-regional correlations, which is not robust to noise. To that end, we project the data onto a space where Euclidean distance is used as a proxy for sample correlation. We then propose to leverage hierarchical clustering to gather together highly correlated variables within each region prior to inter-regional correlation estimation. We provide consistency results, and empirically show our approach surpasses several other popular methods in terms of quality. We also provide illustrations on real-world datasets that further demonstrate its effectiveness.

Solid state spins have demonstrated significant potential in quantum sensing with applications including fundamental science, medical diagnostics and navigation. The quantum sensing schemes showing best performance under ambient conditions all utilize microwave or radio-frequency driving, which poses a significant limitation for miniaturization, energy-efficiency and non-invasiveness of quantum sensors. We overcome this limitation by demonstrating a purely optical approach to coherent quantum sensing. Our scheme involves the $^{15}$N nuclear spin of the Nitrogen-Vacancy (NV) center in diamond as a sensing resource, and exploits NV spin dynamics in oblique magnetic fields near the NV's excited state level anti-crossing to optically pump the nuclear spin into a quantum superposition state. We demonstrate all-optical free-induction decay measurements - the key protocol for low-frequency quantum sensing - both on single spins and spin ensembles. Our results pave the way for highly compact quantum sensors to be employed for magnetometry or gyroscopy applications in challenging environments., Comment: 6 pages, 4 figures, plus supplementary material. Questions and comments are welcome

Resting state brain functional connectivity quantifies the similarity between brain regions, each of which consists of voxels at which dynamic signals are acquired via neuroimaging techniques such as blood-oxygen-level-dependent signals in functional magnetic resonance imaging. Pearson correlation and similar metrics have been adopted by neuroscientists to estimate inter-regional connectivity, usually after averaging of signals within regions. However, dependencies between signals within each region and the presence of noise could contaminate such inter-regional correlation estimates. We propose a mixed-effects model with a novel covariance structure that explicitly isolates the different sources of variability in the observed BOLD signals, including correlated regional signals, local spatiotemporal variability, and measurement error. Methods for tackling the computational challenges associated with restricted maximum likelihood estimation will be discussed. Large sample properties are discussed and used for uncertainty quantification. Simulation results demonstrate that the parameters of the proposed model parameters can be accurately estimated and is superior to the Pearson correlation of averages in the presence of spatiotemporal noise. The proposed model is also applied to a real data set of BOLD signals collected from rats to construct individual brain networks., Comment: 25 pages, 6 figures

The combination of machine learning models with physical models is a recent research path to learn robust data representations. In this paper, we introduce p$^3$VAE, a generative model that integrates a physical model which deterministically models some of the true underlying factors of variation in the data. To fully leverage our hybrid design, we enhance an existing semi-supervised optimization technique and introduce a new inference scheme that comes along meaningful uncertainty estimates. We apply p$^3$VAE to the pixel-wise classification of airborne hyperspectral images. Our experiments on simulated and real data demonstrate the benefits of our hybrid model against conventional machine learning models in terms of extrapolation capabilities and interpretability. In particular, we show that p$^3$VAE naturally has high disentanglement capabilities. Our code and data have been made publicly available at https://github.com/Romain3Ch216/p3VAE., Comment: 29 pages, 14 figures, submitted to Springer Machine Learning

Node role explainability in complex networks is very difficult, yet is crucial in different application domains such as social science, neurosciences or computer science. Many efforts have been made on the quantification of hubs revealing particular nodes in a network using a given structural property. Yet, in several applications, when multiple instances of networks are available and several structural properties appear to be relevant, the identification of node roles remains largely unexplored. Inspired by the node automorphically equivalence relation, we define an equivalence relation on graph nodes associated with any collection of nodal statistics (i.e. any functions on the node-set). This allows us to define new graph global measures: the power coefficient, and the orthogonality score to evaluate the parsimony and heterogeneity of a given nodal statistics collection. In addition, we introduce a new method based on structural patterns to compare graphs that have the same vertices set. This method assigns a value to a node to determine its role distinctiveness in a graph family. Extensive numerical results of our method are conducted on both generative graph models and real data concerning human brain functional connectivity. The differences in nodal statistics are shown to be dependent on the underlying graph structure. Comparisons between generative models and real networks combining two different nodal statistics reveal the complexity of human brain functional connectivity with differences at both global and nodal levels. Using a group of 200 healthy controls connectivity networks, our method computes high correspondence scores among the whole population, to detect homotopy, and finally quantify differences between comatose patients and healthy controls., Comment: 15 pages, 16 figures

We investigate the optical properties of silicon-vacancy (SiV) and germanium-vacancy (GeV) color centers in nanodiamonds under hydrostatic pressure up to 180 GPa. The nanodiamonds were synthetized by Si or Ge-doped plasma assisted chemical vapor deposition and, for our experiment, pressurized in a diamond anvil cell. Under hydrostatic pressure we observe blue-shifts of the SiV and GeV zero-phonon lines by 17 THz (70 meV) and 78 THz (320 meV), respectively. These measured pressure induced shifts are in good agreement with ab initio calculations that take into account the lattice compression based on the equation of state of diamond and that are extended to the case of the tin-vacancy (SnV) center. This work provides guidance on the use of group-IV-vacancy centers as quantum sensors under extreme pressures that will exploit their specific optical and spin properties induced by their intrinsic inversion-symmetric structure., Comment: 7 pages, 4 figures

An important tool in the investigation of the early stages of neurodegenerative disease is the study of dissected living tissue from the brain of an animal model. Such investigations allow the physical structure of individual neurons and neural circuits to be probed alongside neuronal electrical activity, disruption of which can shed light on the mechanisms of emergence of disease. Existing techniques for recording activity rely on potentially damaging direct interaction with the sample, either mechanically as point electrical probes or via intense focused laser light combined with highly specific genetic modification and/or potentially toxic fluorescent dyes. In this work, we instead perform passive, microscopic-scale recording of electrical activity using a biocompatible quantum sensor based on colour centres in diamond. We record biomagnetic field induced by ionic currents in mouse corpus callosum axons without direct sample interaction, accurately recovering signals corresponding to action potential propagation while demonstrating in situ pharmacology during biomagnetic recording through tetrodotoxin inhibition of voltage gated sodium channels. Our results open a promising new avenue for the microscopic recording of neuronal signals, offering the prospect of high resolution imaging of electrical circuits in the living mammalian brain.

The detection of physiological activity at the microscopic level is key for understanding the function of biosystems and relating this to physical structure. Current sensing methods often rely on invasive probes to stimulate and detect activity, bearing the risk of inducing damage in the target system. In recent years, a new type of biosensor based on color centers in diamond offers the possibility to passively, noninvasively sense and image living biological systems. Here, we use such a sensor for the \textit{in-vitro} recording of the local magnetic field generated by tightly focused, high intensity pulsed laser optogenetic neuromuscular stimulation of the extensor digitorum longus muscles. Recordings captured a compound action potential response and a slow signal component which we seek to explain using a detailed model of the biological system. We show that our sensor is capable of recording localized neuromuscular activity from the laser stimulation site without photovoltaic or fluorescence artifacts associated with alternative techniques. Our work represents an important step towards selective induction of localized neurobiological activity while performing passive sensing and imaging with diamond sensors, motivating further research into mapping of neural activity and intra-cellular processes.

The production of negative ions is of significant interest for applications including mass spectrometry, materials surface processing, and neutral beam injection for magnetic confined fusion. Neutral beam injection sources maximise negative ion production through the use of surface production processes and low work function metals, which introduce complex engineering. Investigating materials and techniques to avoid the use of low work function metals is of interest to broaden the application of negative ion sources and simplify future devices. In this study, we use pulsed sample biasing to investigate the surface production of negative ions from nitrogen doped diamond. The use of a pulsed bias allows for the study of insulating samples in a preserved surface state at temperatures between 150$^{\circ}$C and 700$^{\circ}$C in a 2 Pa, 130 W, (n$_e$ $\sim$ $10^9$ cm$^{-3}$, T$_e$ $\sim$ 0.6 eV) inductively coupled deuterium plasma. The negative ion yield during the application of a pulsed negative bias is measured using a mass spectrometer and found to be approximately 20% higher for nitrogen doped diamond compared to non-doped diamond. It is also shown that the pulsed sample bias has a lower peak negative ion yield compared to a continuous sample bias, which suggests that the formation of an optimum ratio of defects on its surface can be favourable for negative ion production., Comment: This is the Accepted Manuscript version of an article accepted for publication in Journal of Physics D: Applied physics. IOP Publishing Ltd is not responsible for any errors or omissions in this version of the manuscript or any version derived from it. This Accepted Manuscript is published under a CC BY licence. The Version of Record is available online at

In the general setting of long-memory multivariate time series, the long-memory characteristics are defined by two components. The long-memory parameters describe the autocorrelation of each time series. And the long-run covariance measures the coupling between time series, with general phase parameters. It is of interest to estimate the long-memory, long-run covariance and general phase parameters of time series generated by this wide class of models although they are not necessarily Gaussian nor stationary. This estimation is thus not directly possible using real wavelets decomposition or Fourier analysis. Our purpose is to define an inference approach based on a representation using quasi-analytic wavelets. We first show that the covariance of the wavelet coefficients provides an adequate estimator of the covariance structure including the phase term. Consistent estimators based on a local Whittle approximation are then proposed. Simulations highlight a satisfactory behavior of the estimation on finite samples on linear time series and on multivariate fractional Brownian motions. An application on a real neuroscience dataset is presented, where long-memory and brain connectivity are inferred.