Search

The Atomic Clock Ensemble in Space (ACES) mission is developing high performance clocks and links for space to test Einstein's theory of general relativity. From the International Space Station, the ACES payload will distribute a clock signal with fractional frequency stability and accuracy of 1E-16 establishing a worldwide network to compare clocks in space and on the ground. ACES will provide an absolute measurement of Einstein's gravitational redshift, it will search for time variations of fundamental constants, contribute to test topological dark matter models, and perform Standard Model Extension tests. Moreover, the ground clocks connected to the ACES network will be compared over different continents and used to measure geopotential differences at the clock locations. After solving some technical problems, the ACES flight model is now approaching its completion. System tests involving the laser-cooled Cs clock PHARAO, the active H-maser SHM and the on-board frequency comparator FCDP have measured the performance of the clock signal delivered by ACES. The ACES microwave link MWL is currently under test. The single-photon avalanche detector of the optical link ELT has been tested and will now be integrated in the ACES payload. The ACES mission concept, its scientific objectives, and the recent test results are discussed here together with the major milestones that will lead us to the ACES launch., Comment: 12 pages, 8 figures

It is usually assumed that the memory of any thermo-mechanical protocol applied to a glass can be erased by heating the material in the supercooled liquid. While this is true for thermally treated amorphous solids, we show that hydrostatic compression can irreversibly modify the atomic motion, thermodynamic state and structure of a prototypical metallic glass-former, in a way which depends on the degree of ergodicity reached by the material during compression. While enhanced kinetic and thermodynamic stability can be obtained by quenching the dense liquid, high-pressure annealing in the glass leads to thermal rejuvenation and complex structural rearrangements at the level of the short and medium range order. When heated above their glass transition temperature, these compressed glasses do not convert into the pristine supercooled liquid but rather transform into different systems, challenging the generally accepted idea of an equilibrium recovery controlled solely by the microscopic $\alpha$-relaxation process.

We present a general variational framework for the training of freeform nonlinearities in layered computational architectures subject to some slope constraints. The regularization that we add to the traditional training loss penalizes the second-order total variation of each trainable activation. The slope constraints allow us to impose properties such as 1-Lipschitz stability, firm non-expansiveness, and monotonicity/invertibility. These properties are crucial to ensure the proper functioning of certain classes of signal-processing algorithms (e.g., plug-and-play schemes, unrolled proximal gradient, invertible flows). We prove that the global optimum of the stated constrained-optimization problem is achieved with nonlinearities that are adaptive nonuniform linear splines. We then show how to solve the resulting function-optimization problem numerically by representing the nonlinearities in a suitable (nonuniform) B-spline basis. Finally, we illustrate the use of our framework with the data-driven design of (weakly) convex regularizers for the denoising of images and the resolution of inverse problems.

Text-to-SQL, the task of translating natural language questions into SQL queries, is part of various business processes. Its automation, which is an emerging challenge, will empower software practitioners to seamlessly interact with relational databases using natural language, thereby bridging the gap between business needs and software capabilities. In this paper, we consider Large Language Models (LLMs), which have achieved state of the art for various NLP tasks. Specifically, we benchmark Text-to-SQL performance, the evaluation methodologies, as well as input optimization (e.g., prompting). In light of the empirical observations that we have made, we propose two novel metrics that were designed to adequately measure the similarity between SQL queries. Overall, we share with the community various findings, notably on how to select the right LLM on Text-to-SQL tasks. We further demonstrate that a tree-based edit distance constitutes a reliable metric for assessing the similarity between generated SQL queries and the oracle for benchmarking Text2SQL approaches. This metric is important as it relieves researchers from the need to perform computationally expensive experiments such as executing generated queries as done in prior works. Our work implements financial domain use cases and, therefore contributes to the advancement of Text2SQL systems and their practical adoption in this domain., Comment: 10 pages, ICSE-SEIP 2024

We propose to learn non-convex regularizers with a prescribed upper bound on their weak-convexity modulus. Such regularizers give rise to variational denoisers that minimize a convex energy. They rely on few parameters (less than 15,000) and offer a signal-processing interpretation as they mimic handcrafted sparsity-promoting regularizers. Through numerical experiments, we show that such denoisers outperform convex-regularization methods as well as the popular BM3D denoiser. Additionally, the learned regularizer can be deployed to solve inverse problems with iterative schemes that provably converge. For both CT and MRI reconstruction, the regularizer generalizes well and offers an excellent tradeoff between performance, number of parameters, guarantees, and interpretability when compared to other data-driven approaches.

Natural language processing of Low-Resource Languages (LRL) is often challenged by the lack of data. Therefore, achieving accurate machine translation (MT) in a low-resource environment is a real problem that requires practical solutions. Research in multilingual models have shown that some LRLs can be handled with such models. However, their large size and computational needs make their use in constrained environments (e.g., mobile/IoT devices or limited/old servers) impractical. In this paper, we address this problem by leveraging the power of large multilingual MT models using knowledge distillation. Knowledge distillation can transfer knowledge from a large and complex teacher model to a simpler and smaller student model without losing much in performance. We also make use of high-resource languages that are related or share the same linguistic root as the target LRL. For our evaluation, we consider Luxembourgish as the LRL that shares some roots and properties with German. We build multiple resource-efficient models based on German, knowledge distillation from the multilingual No Language Left Behind (NLLB) model, and pseudo-translation. We find that our efficient models are more than 30\% faster and perform only 4\% lower compared to the large state-of-the-art NLLB model., Comment: The associated model is published on HuggingFace: https://huggingface.co/etamin/Letz-Translate-OPUS-LB-EN The Dictionary used in this paper is available in Github: https://github.com/Etamin/Ltz_dictionary

The emergence of deep-learning-based methods to solve image-reconstruction problems has enabled a significant increase in reconstruction quality. Unfortunately, these new methods often lack reliability and explainability, and there is a growing interest to address these shortcomings while retaining the boost in performance. In this work, we tackle this issue by revisiting regularizers that are the sum of convex-ridge functions. The gradient of such regularizers is parameterized by a neural network that has a single hidden layer with increasing and learnable activation functions. This neural network is trained within a few minutes as a multistep Gaussian denoiser. The numerical experiments for denoising, CT, and MRI reconstruction show improvements over methods that offer similar reliability guarantees.

Lipschitz-constrained neural networks have several advantages over unconstrained ones and can be applied to a variety of problems, making them a topic of attention in the deep learning community. Unfortunately, it has been shown both theoretically and empirically that they perform poorly when equipped with ReLU activation functions. By contrast, neural networks with learnable 1-Lipschitz linear splines are known to be more expressive. In this paper, we show that such networks correspond to global optima of a constrained functional optimization problem that consists of the training of a neural network composed of 1-Lipschitz linear layers and 1-Lipschitz freeform activation functions with second-order total-variation regularization. Further, we propose an efficient method to train these neural networks. Our numerical experiments show that our trained networks compare favorably with existing 1-Lipschitz neural architectures.

Regression is one of the core problems tackled in supervised learning. Rectified linear unit (ReLU) neural networks generate continuous and piecewise-linear (CPWL) mappings and are the state-of-the-art approach for solving regression problems. In this paper, we propose an alternative method that leverages the expressivity of CPWL functions. In contrast to deep neural networks, our CPWL parameterization guarantees stability and is interpretable. Our approach relies on the partitioning of the domain of the CPWL function by a Delaunay triangulation. The function values at the vertices of the triangulation are our learnable parameters and identify the CPWL function uniquely. Formulating the learning scheme as a variational problem, we use the Hessian total variation (HTV) as regularizer to favor CPWL functions with few affine pieces. In this way, we control the complexity of our model through a single hyperparameter. By developing a computational framework to compute the HTV of any CPWL function parameterized by a triangulation, we discretize the learning problem as the generalized least absolute shrinkage and selection operator (LASSO). Our experiments validate the usage of our method in low-dimensional scenarios.

Many feedforward neural networks (NNs) generate continuous and piecewise-linear (CPWL) mappings. Specifically, they partition the input domain into regions on which the mapping is affine. The number of these so-called linear regions offers a natural metric to characterize the expressiveness of CPWL NNs. The precise determination of this quantity is often out of reach in practice, and bounds have been proposed for specific architectures, including for ReLU and Maxout NNs. In this work, we generalize these bounds to NNs with arbitrary and possibly multivariate CPWL activation functions. We first provide upper and lower bounds on the maximal number of linear regions of a CPWL NN given its depth, width, and the number of linear regions of its activation functions. Our results rely on the combinatorial structure of convex partitions and confirm the distinctive role of depth which, on its own, is able to exponentially increase the number of regions. We then introduce a complementary stochastic framework to estimate the average number of linear regions produced by a CPWL NN. Under reasonable assumptions, the expected density of linear regions along any 1D path is bounded by the product of depth, width, and a measure of activation complexity (up to a scaling factor). This yields an identical role to the three sources of expressiveness: no exponential growth with depth is observed anymore.

Lipschitz-constrained neural networks have many applications in machine learning. Since designing and training expressive Lipschitz-constrained networks is very challenging, there is a need for improved methods and a better theoretical understanding. Unfortunately, it turns out that ReLU networks have provable disadvantages in this setting. Hence, we propose to use learnable spline activation functions with at least 3 linear regions instead. We prove that this choice is optimal among all component-wise $1$-Lipschitz activation functions in the sense that no other weight constrained architecture can approximate a larger class of functions. Additionally, this choice is at least as expressive as the recently introduced non component-wise Groupsort activation function for spectral-norm-constrained weights. Previously published numerical results support our theoretical findings.

Rectified-linear-unit (ReLU) neural networks, which play a prominent role in deep learning, generate continuous and piecewise-linear (CPWL) functions. While they provide a powerful parametric representation, the mapping between the parameter and function spaces lacks stability. In this paper, we investigate an alternative representation of CPWL functions that relies on local hat basis functions. It is predicated on the fact that any CPWL function can be specified by a triangulation and its values at the grid points. We give the necessary and sufficient condition on the triangulation (in any number of dimensions) for the hat functions to form a Riesz basis, which ensures that the link between the parameters and the corresponding CPWL function is stable and unique. In addition, we provide an estimate of the $\ell_2\rightarrow L_2$ condition number of this local representation. Finally, as a special case of our framework, we focus on a systematic parametrization of $\mathbb{R}^d$ with control points placed on a uniform grid. In particular, we choose hat basis functions that are shifted replicas of a single linear box spline. In this setting, we prove that our general estimate of the condition number is optimal. We also relate our local representation to a nonlocal one based on shifts of a causal ReLU-like function.

Ariel, the Atmospheric Remote-sensing Infrared Exoplanet Large-survey, was adopted as the fourth medium-class mission in ESA's Cosmic Vision programme to be launched in 2029. During its 4-year mission, Ariel will study what exoplanets are made of, how they formed and how they evolve, by surveying a diverse sample of about 1000 extrasolar planets, simultaneously in visible and infrared wavelengths. It is the first mission dedicated to measuring the chemical composition and thermal structures of hundreds of transiting exoplanets, enabling planetary science far beyond the boundaries of the Solar System. The payload consists of an off-axis Cassegrain telescope (primary mirror 1100 mm x 730 mm ellipse) and two separate instruments (FGS and AIRS) covering simultaneously 0.5-7.8 micron spectral range. The satellite is best placed into an L2 orbit to maximise the thermal stability and the field of regard. The payload module is passively cooled via a series of V-Groove radiators; the detectors for the AIRS are the only items that require active cooling via an active Ne JT cooler. The Ariel payload is developed by a consortium of more than 50 institutes from 16 ESA countries, which include the UK, France, Italy, Belgium, Poland, Spain, Austria, Denmark, Ireland, Portugal, Czech Republic, Hungary, the Netherlands, Sweden, Norway, Estonia, and a NASA contribution., Comment: Ariel Definition Study Report, 147 pages. Reviewed by ESA Science Advisory Structure in November 2020. Original document available at: https://www.cosmos.esa.int/documents/1783156/3267291/Ariel_RedBook_Nov2020.pdf/

Oxidative stress, which is one of the main harmful mechanisms of pathologies including is-chemic stroke, contributes to both neurons and endothelial cell damages, leading to vascular lesions. Although many antioxidants have been tested in preclinical studies, no treatment is currently available for stroke patients. Since cerium oxide nanoparticles (CNPs) exhibit remarkable antioxidant capacities, our objective is to develop an innovative coating to enhance CNPs biocompatibility without disrupting their antioxidant capacities or enhance their toxicity. This study reports the synthesis and characterization of functional polymers and their impact on the enzyme-like catalytic activity of CNPs. To study the toxicity and the antioxidant properties of CNPs for stroke and particularly endothelial damages, in vitro studies are conducted on a cerebral endothelial cell line (bEnd.3). Despite their internalization in bEnd.3 cells, coated CNPs are devoid of cytotoxicity. Microscopy studies report an intracellular localization of CNPs, more precisely in endosomes. All CNPs reduces glutamate-induced intracellular production of ROS in endothelial cells but one CNP significantly reduces both the production of mitochondrial super-oxide anion and DNA oxidation. In vivo studies report a lack of toxicity in mice. This study there-fore describes and identifies biocompatible CNPs with interesting antioxidant properties for ischemic stroke and related pathologies., Comment: 27 pages 9 figures