Restoring communication for people who have lost the ability to speak or move after a brain injury is a major challenge. While intracranial implants now enable high-performing brain-computer-interfaces, non-invasive alternatives are still lagging behind. Here, we present Brain2Qwerty v2, a model that can decode the production of natural sentences solely from real-time magnetoencephalography (MEG) recordings. By collecting 22,000 sentences typed by nine subjects, each recorded for 10 hours, our model leverages character, word and sentence-level representations to achieve an average word error rate (WER) of 39

In this work we study the behavior of the forward-backward (FB) algorithm when the proximity operator is replaced by a sub-iterative procedure to approximate a Gaussian denoiser, in a Plug-and-Play (PnP) fashion. In particular, we consider both analysis and synthesis Gaussian denoisers within a dictionary framework, obtained by unrolling dual-FB iterations or FB iterations, respectively. We analyze the associated minimization problems as well as the asymptotic behavior of the resulting FB-PnP iterations. In particular, we show that the synthesis Gaussian denoising problem can be viewed as a proximity operator. For each case, analysis and synthesis, we show that the FB-PnP algorithms solve the same problem whether we use only one or an infinite number of sub-iteration to solve the denoising problem at each iteration. To this aim, we show that each "one sub-iteration" strategy within the FB-PnP can be interpreted as a primal-dual algorithm when a warm-restart strategy is used. We further present similar results when using a Moreau-Yosida smoothing of the global problem, for an arbitrary number of sub-iterations. Finally, we provide numerical simulations to illustrate our theoretical results. In particular we first consider a toy compressive sensing example, as well as an image restoration problem in a deep dictionary framework.

Event identification in continuous neural recordings is a critical task in neuroscience. Decoding in EEG is dominated by classifying windows aligned to known event onsets. However, while available in controlled experiments, such onsets are absent in continuous real-world monitoring. Here, we introduce DANCE, a deep learning pipeline that frames neural decoding as a set-prediction problem and jointly detects and classifies events directly from raw, unaligned signals. Evaluated separately on ten datasets curated from the literature with a wide variety of event types (ranging from milliseconds to minutes in duration), our model outperforms existing methods on a broad range of cognitive, clinical and BCI tasks. This single architecture establishes a new state of the art in the competitive task of seizure monitoring and matches the accuracy of onset-informed models for BCI tasks. Overall, our method marks a step towards end-to-end asynchronous neural decoding models

General anesthesia offers a rare opportunity to observe the human brain under a standardized, controlled perturbation. Yet intraoperative electroencephalography (EEG) is almost always reduced to a single proprietary depth index, collapsing a rich trajectory into one number and discarding how a brain moves between states. Here we ask whether the geometry of that trajectory, not merely the depth it reaches, carries clinically meaningful information. Using similarity-based self-supervised learning on raw, two-electrode frontal EEG, with no labels, we place each recording within a low-dimensional space in which anesthetic depth becomes one readable axis while the shape of a patient's path encodes additional structure. We validate the representation across two cohorts and two acquisition systems totaling more than 1,000 patients. Depth of anesthesia is predicted accurately (BIS mean absolute error = 3.2, R2 = 0.82), and in the sparse-montage setting our compact ( 68k parameter) model remains competitive with EEG foundation models orders of magnitude larger (4M-157M parameters), indicating that matching the representation to the recording dominates raw scale. The learned space organizes age along its own gradient, independent from depth, without supervision. The same space also aligns with interpretable anesthetic signatures like frontal alpha, slow-delta, and burst suppression, linking this data-driven representation to established neurophysiology. On an independent cohort with longitudinal follow-up, the geometry of the early trajectory separates 30- month cognitive and mortality outcomes complementary to age (AUROC 0.86). These results suggest that the path a brain traces through anesthesia is a label-efficient correlate of latent vulnerability, motivating prospective validation.

Practical diffusion sampling is a numerical approximation problem: under a fixed inference budget, one must simulate a reverse-time ODE or SDE using only a limited number of denoising steps, so discretization error is often the dominant source of error. Existing non-asymptotic analyses provide convergence guarantees, but are typically too loose and too insensitive to diffusion parameters to guide practical design: broad families of schedules receive the same rates, which depend on coarse worst-case quantities such as the dimension or the drift Lipschitz constant. We take a less ambitious but more informative route. In the exact-score setting, we derive first-order asymptotic expansions of the Euler-Maruyama weak and Fréchet discretization errors. These formulas hold for general smooth reverse diffusions and become fully explicit under Gaussian data. They show how discretization error adapts to the geometry of the data through the covariance spectrum, and how this geometry interacts with key diffusion parameters, including the diffusion schedules and the diffusion-term coefficient. This yields tractable objectives for geometry-aware parameter optimization. Finally, we show that the qualitative predictions of the Gaussian formulas remain robust across diffusion sampling problems with different geometries, including image generation on different datasets and image posterior sampling.

Modern machine learning progresses through empirical work, benchmarking new methods to evaluate relative performance. However, the statistical variability inherent to evaluation - exacerbated by the stochastic nature of many algorithms - often makes performance estimation unreliable due to the limited test samples available, leading to a validation crisis in which genuine advances are difficult to discern. In this work, we show that cross-validation improves markedly confidence when evaluating and comparing learning algorithm performances. We introduce the concept of sample gain, which quantifies the virtual data augmentation achieved by using multiple cross-validation splits to reduce benchmarking variance. Experiments on both synthetic and real-world datasets (histopathologic scans and NLP fine-tuning) demonstrate that multiple splits can substantially improve the reliability and stability of performance estimates, with diminishing returns often setting in later than expected. We also introduce a procedure to dynamically early-stop cross-validation by estimating from the first few folds if subsequent folds will bring large sample gains. Our findings highlight the value of pushing cross-validation on available samples to achieve robust and reliable benchmarking.

Physiological signal analysis often involves identifying events crucial to understanding biological dynamics. Many methods have been proposed to detect them, from handcrafted and supervised approaches to unsupervised techniques. All these methods tend to produce spurious events, particularly as they detect each event independently. This work introduces UNHaP (Unmix Noise from Hawkes Processes), a novel approach addressing the joint learning of temporal structures in events and the removal of spurious detections. By treating the event detection output as a mixture of structured Hawkes and unstructured Poisson events, UNHaP efficiently unmixes these processes and estimates their parameters. This approach significantly enhances event distribution characterization while minimizing false detection rates on simulated and real data.

Detecting rare events and anomalies in large-scale signals is essential in fields such as astronomy, physical simulations, and biomedical science. In many cases, this problem naturally decomposes into identifying common local patterns and detecting deviations that correspond to anomalies. Convolutional Dictionary Learning (CDL) is a powerful tool for modeling local structures, but its adoption for this task has been limited by computational demands and sensitivity to outliers. We introduce RoseCDL, a novel CDL algorithm designed for robust and scalable modeling of signal pattern distribution. RoseCDL leverages stochastic windowing for efficient training and incorporates inline outlier detection to enhance robustness. This enables unsupervised identification of anomalous and rare patterns in long signals based on the local reconstruction loss. Experiments on real-world datasets show that RoseCDL delivers improved detection accuracy and computational efficiency, making CDL practical for challenging detection tasks in large-scale signal analysis.

A central challenge in many areas of science and engineering is to identify model parameters that are consistent with prior knowledge and empirical data. Bayesian inference offers a principled framework for this task, but can be computationally prohibitive when models are defined by stochastic simulators. Simulation-based Inference (SBI) is a suite of methods developed to overcome this limitation, which has enabled scientific discoveries in fields such as particle physics, astrophysics, and neuroscience. The core idea of SBI is to train neural networks on data generated by a simulator, without requiring access to likelihood evaluations. Once trained, inference is amortized: The neural network can rapidly perform Bayesian inference on empirical observations without requiring additional training or simulations. In this tutorial, we provide a practical guide for practitioners aiming to apply SBI methods. We outline a structured SBI workflow and offer practical guidelines and diagnostic tools for every stage of the process – from setting up the simulator and prior, choosing and training inference networks, to performing inference and validating the results. We illustrate these steps through examples from astrophysics, psychophysics, and neuroscience. This tutorial empowers researchers to apply state-of-the-art SBI methods, facilitating efficient parameter inference for scientific discovery.

On April 7th, 2022, after 16 months of tandem flight with its predecessor, Jason-3, the Copernicus Sentinel-6 Michael Freilich (MF) satellite became the regional reference mission in the sea level climate record. Its on-board altimeter, POSeidon-4, is the first radar altimeter allowing a simultaneous and continuous acquisition in Low 2023, a new version of Sentinel-6MF ground segment (processing baseline F08) brought major improvements for the LR mode, with the implementation of a numerical retracker in addition to the hist sent work covers the full assessment of this new LR numerical retracker over open ocean, spanning from the retracker's outputs to their contribution to the Global Mean Sea Level. Improvements with respect to MLE4 appeared mainly in terms of sea-state related effects, leading to a 60 % reduction of the Sentinel-6MF/Jason-3 Sea Surface Height Anomaly bias correlated to Significant Wave Height. Such result improves an already very good continuity between the two tandem phase is also precisely assessed. The small remaining discrepancies are attributed to different components of the system, such as the orbit, the radiometer wet troposphere correction, C-band pro feature of the Sentinel-6MF numerical retracker is the use of t Resolution (LR) as well as in High Resolution (HR) mode. In March orical Maximum Likelihood Estimator-4 (MLE4) retracker. The pre- missions. The agreement between Jason-3 and Sentinel-6MF over the cessing or an MLE4-based empirical adjustment. Another important he in-flight Point Target Response to mitigate instrumental changes and thus improve long term stability. The Global Mean Sea Level with respect to Jason-3, once the radiometer wet troposphere correction impact is removed. (c) 2024 COSPAR. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/ by-nc-nd/4.0/). analysis presented in this paper shows no significant trend difference

Sampling from an unknown distribution, accessible only through discrete samples, is a fundamental problem at the core of generative AI. The current state-of-the-art methods follow a two-step process: first, estimating the score function (the gradient of a smoothed log-distribution) and then applying a diffusion-based sampling algorithm – such as Langevin or Diffusion models. The resulting distribution's correctness can be impacted by four major factors: the generalization and optimization errors in score matching, and the discretization and minimal noise amplitude in the diffusion. In this paper, we make the sampling error explicit when using a diffusion sampler in the Gaussian setting. We provide a sharp analysis of the Wasserstein sampling error that arises from these four error sources. This allows us to rigorously track how the anisotropy of the data distribution (encoded by its power spectrum) interacts with key parameters of the end-to-end sampling method, including the number of initial samples, the stepsizes in both score matching and diffusion, and the noise amplitude. Notably, we show that the Wasserstein sampling error can be expressed as a kernel-type norm of the data power spectrum, where the specific kernel depends on the method parameters. This result provides a foundation for further analysis of the tradeoffs involved in optimizing sampling accuracy.

DeepInverse is an open-source PyTorch-based library for solving imaging inverse problems. The library covers all crucial steps in image reconstruction from the efficient implementation of forward operators (e.g., optics, MRI, tomography), to the definition and resolution of variational problems and the design and training of advanced neural network architectures. In this paper, we describe the main functionality of the library and discuss the main design choices.

Satellite radar altimetry has been used for over 30 years to measure sea surface height (SSH) variations to build records of essential climate variables like the mean sea level (MSL) for a robust assessment of climate change. A key step of the data processing toward this goal is the retracking, that is, the statistical analysis of the radar waveforms to estimate the geophysical parameters. For a robust and optimal estimation, retracking algorithms should account for a reliable waveform model, the time-varying instrumental point target response (PTR), and an accurate description of the waveform noise. However, this is not the case for the operational retracking solutions implemented in the ground segments, which specifically make use of an unweighted estimator, therefore not accounting for the waveform speckle noise. In this article, we present a novel computationally efficient retracking solution for Ku-band low resolution mode (LRM) data that accounts for both the in-flight PTR and a realistic waveform noise through a weighted estimator defined to be statistically equivalent to a maximum likelihood estimator. We consider two waveform models: the Adaptive model and the numerical Brown model, for a consistent comparison with operational retracking solutions. We focus on the current reference mission, Sentinel-6 MF, for which an accurate noise characterization is crucial to account for the pulse-to-pulse correlations resulting from the higher pulse repetition frequency (PRF) compared to conventional configurations. We demonstrate that the novel retracking solution is optimal, providing parameter uncertainties compatible with the Cramer-Rao bounds of minimum variance, and unbiased, while a bias up to 1 cm in the epoch estimation and suboptimality for all parameters is found for the unweighted solutions. We validated the algorithm on realistic simulations and applied it to one cycle of Sentinel-6MF LR 20-Hz data, demonstrating significant improvements in the precision of estimated parameter: similar to 60% for significant wave height (SWH), similar to 12% for the epoch, and similar to 50% for sigma(0) compared to current operational solutions. We report for the first time, geophysical parameter uncertainties consistently computed at 20 Hz as output of the retracker. We also introduce an innovative Bayesian approach for analyzing waveform data to complement current solutions, providing a robust method for the estimation of the parameter uncertainties and correlations. Finally, we perform a comprehensive analysis of parameter correlations on simulations and data, compared to theoretical expectations based on Fisher matrix analysis, demonstrating the importance of optimality for a correct estimation. Specifically, we show that the unweighted, suboptimal retracking solution significantly underestimates the epoch-SWH correlation by a factor of similar to 1.5 in the correlation coefficients, compared to the optimal solution and theoretical expectations. This could significantly impact the estimation of corrections such as sea state bias (SSB) and high-frequency adjustment (HFA), warranting further assessment. Overall, optimal retracking solutions should be considered to derive robust long-term sea-level records for climate research using past, current, and future altimetry missions.

In a wide range of spatio-temporal datasets, from sociology to seismology, self-exciting dynamics are often observed, characterized by event triggering and clustering across both space and time. Space-time Hawkes processes provide a powerful framework to model such phenomena. This paper introduces a flexible parametric inference method to estimate the underlying kernel parameters involved in the intensity function of a space-time Hawkes process based on such data. Our approach combines three core components: 1) kernels with finite support, 2) discretization of the space-time domain, and 3) efficient (possibly approximate) precomputations. The inference method we propose then relies on a gradient-based solver that offers both computational efficiency and strong statistical performance. Alongside a detailed presentation of the algorithmic framework, we present numerical experiments on synthetic and real spatio-temporal data, offering solid empirical evidence of the validity and applicability of the proposed methodology.

Selecting an appropriate prior to compensate for information loss due to the measurement operator is a fundamental challenge in imaging inverse problems. Implicit priors based on denoising neural networks have become central to widely-used frameworks such as Plug-and-Play (PnP) algorithms. In this work, we introduce Fixed-points of Restoration (FiRe) priors as a new framework for expanding the notion of priors in PnP to general restoration models beyond traditional denoising models. The key insight behind FiRe is that smooth images emerge as fixed points of the composition of a degradation operator with the corresponding restoration model. This enables us to derive an explicit formula for our implicit prior by quantifying invariance of images under this composite operation. Adopting this fixed-point perspective, we show how various restoration networks can effectively serve as priors for solving inverse problems. The FiRe framework further enables ensemble-like combinations of multiple restoration models as well as acquisition-informed restoration networks, all within a unified optimization approach. Experimental results validate the effectiveness of FiRe across various inverse problems, establishing a new paradigm for incorporating pretrained restoration models into PnP-like algorithms. Code available at https://github.com/matthieutrs/fire.

Unsupervised Domain Adaptation (DA) consists of adapting a model trained on a labeled source domain to perform well on an unlabeled target domain with some data distribution shift. While many methods have been proposed in the literature, fair and realistic evaluation remains an open question, particularly due to methodological difficulties in selecting hyperparameters in the unsupervised setting. With SKADA-bench, we propose a framework to evaluate DA methods on diverse modalities, beyond computer vision task that have been largely explored in the literature. We present a complete and fair evaluation of existing shallow algorithms, including reweighting, mapping, and subspace alignment. Realistic hyperparameter selection is performed with nested cross-validation and various unsupervised model selection scores, on both simulated datasets with controlled shifts and real-world datasets across diverse modalities, such as images, text, biomedical, and tabular data. Our benchmark highlights the importance of realistic validation and provides practical guidance for real-life applications, with key insights into the choice and impact of model selection approaches. SKADA-bench is open-source, reproducible, and can be easily extended with novel DA methods, datasets, and model selection criteria without requiring re-evaluating competitors. SKADA-bench is available on Github at https://github.com/scikit-adaptation/skada-bench.

Implicit deep learning has recently gained popularity with applications ranging from meta-learning to Deep Equilibrium Networks (DEQs). In its general formulation, it relies on expressing some components of deep learning pipelines implicitly, typically via a root equation called the inner problem. In practice, the solution of the inner problem is approximated during training with an iterative procedure, usually with a fixed number of inner iterations. During inference, the inner problem needs to be solved with new data. A popular belief is that increasing the number of inner iterations compared to the one used during training yields better performance. In this paper, we question such an assumption and provide a detailed theoretical analysis in a simple setting. We demonstrate that overparametrization plays a key role: increasing the number of iterations at test time cannot improve performance for overparametrized networks. We validate our theory on an array of implicit deep-learning problems. DEQs, which are typically overparametrized, do not benefit from increasing the number of iterations at inference while meta-learning, which is typically not overparametrized, benefits from it.

This study presents the first azimuth cutoff analysis in Synthetic Aperture Radar (SAR) altimetry, aiming to assess its applicability in characterizing sea-state dynamics. In SAR imaging, the azimuth cutoff serves as a proxy for the shortest waves, in terms of wavelength, that can be detected by the satellite under certain wind and wave conditions. The magnitude of this parameter is closely related to the wave orbital velocity variance, a key parameter for characterizing wind-wave systems. We exploit wave modulations exhibited in the tail of fully-focused SAR waveforms and extract the azimuth cutoff from the radar signal through the analysis of its along-track autocorrelation function. We showcase the capability of Sentinel-6A in deriving these two parameters based on analyses in the spatial and wavenumber domains, accompanied by a discussion of the limitations. We use Level-1A high-resolution Sentinel-6A data from one repeat cycle (10 days) globally to verify our findings against wave modeled data. In the spatial domain analysis, the estimation of azimuth cutoff involves fitting a Gaussian function to the along-track autocorrelation function. Results reveal pronounced dependencies on wind speed and significant wave height, factors primarily determining the magnitude of the velocity variance. In extreme sea states, the parameters are underestimated by the altimeter, while in relatively calm sea states and in the presence of swells, a substantial overestimation trend is observed. We introduce an alternative approach to extract the azimuth cutoff by identifying the fall-off wavenumber in the wavenumber domain. Results indicate effective mitigation of swell-induced errors, with some additional sensitivity to extreme sea states compared to the spatial domain approach.

Identifying characteristic patterns in time series, such as heartbeats or brain responses to a stimulus, is critical to understanding the physical or physiological phenomena monitored with sensors. Convolutional sparse coding (CSC) methods, which aim to approximate signals by a sparse combination of short signal templates (also called atoms), are well-suited for this task. However, enforcing sparsity leads to non-convex and untractable optimization problems. This article proposes finding the optimal solution to the original and non-convex CSC problem when the atoms do not overlap. Specifically, we show that the reconstruction error satisfies a simple recursive relationship in this setting, which leads to an efficient detection algorithm. We prove that our method correctly estimates the number of patterns and their localization, up to a detection margin that depends on a certain measure of the signal-to-noise ratio. In a thorough empirical study, with simulated and real-world physiological data sets, our method is shown to be more accurate than existing algorithms at detecting the patterns' onsets.

Physiological signal analysis often involves identifying events crucial to understanding biological dynamics. Traditional methods rely on handcrafted procedures or supervised learning, presenting challenges such as expert dependence, lack of robustness, and the need for extensive labeled data. Data-driven methods like Convolutional Dictionary Learning (CDL) offer an alternative but tend to produce spurious detections. This work introduces UNHaP (Unmix Noise from Hawkes Processes), a novel approach addressing the joint learning of temporal structures in events and the removal of spurious detections. Leveraging marked Hawkes processes, UNHaP distinguishes between events of interest and spurious ones. By treating the event detection output as a mixture of structured and unstructured events, UNHaP efficiently unmixes these processes and estimates their parameters. This approach significantly enhances the understanding of event distributions while minimizing false detection rates.