Search

Lagrangian decomposition (LD) is a relaxation method that provides a dual bound for constrained optimization problems by decomposing them into more manageable sub-problems. This bound can be used in branch-and-bound algorithms to prune the search space effectively. In brief, a vector of Lagrangian multipliers is associated with each sub-problem, and an iterative procedure (e.g., a sub-gradient optimization) adjusts these multipliers to find the tightest bound. Initially applied to integer programming, Lagrangian decomposition also had success in constraint programming due to its versatility and the fact that global constraints provide natural sub-problems. However, the non-linear and combinatorial nature of sub-problems in constraint programming makes it computationally intensive to optimize the Lagrangian multipliers with sub-gradient methods at each node of the tree search. This currently limits the practicality of LD as a general bounding mechanism for constraint programming. To address this challenge, we propose a self-supervised learning approach that leverages neural networks to generate multipliers directly, yielding tight bounds. This approach significantly reduces the number of sub-gradient optimization steps required, enhancing the pruning efficiency and reducing the execution time of constraint programming solvers. This contribution is one of the few that leverage learning to enhance bounding mechanisms on the dual side, a critical element in the design of combinatorial solvers. To our knowledge, this work presents the first generic method for learning valid dual bounds in constraint programming.

Modern out-of-order processors face speculative execution attacks. Despite various proposed software and hardware mitigations to prevent such attacks, new attacks keep arising from unknown vulnerabilities. Thus, a formal and rigorous evaluation of the ability of hardware designs to deal with speculative execution attacks is urgently desired. This paper proposes a formal verification technique called Contract Shadow Logic that can considerably improve RTL verification scalability while being applicable to different defense mechanisms. In this technique, we leverage computer architecture design insights to improve verification performance for checking security properties formulated as software-hardware contracts for secure speculation. Our verification scheme is accessible to computer architects and requires minimal formal-method expertise. We evaluate our technique on multiple RTL designs, including three out-of-order processors. The experimental results demonstrate that our technique exhibits a significant advantage in finding attacks on insecure designs and deriving complete proofs on secure designs, when compared to the baseline and two state-of-the-art verification schemes, LEAVE and UPEC., Comment: This paper has been accepted to ASPLOS 2025

Hardware development relies on simulations, particularly cycle-accurate RTL (Register Transfer Level) simulations, which consume significant time. As single-processor performance grows only slowly, conventional, single-threaded RTL simulation is becoming less practical for increasingly complex chips and systems. A solution is parallel RTL simulation, where ideally, simulators could run on thousands of parallel cores. However, existing simulators can only exploit tens of cores. This paper studies the challenges inherent in running parallel RTL simulation on a multi-thousand-core machine (the Graphcore IPU, a 1472-core machine). Simulation performance requires balancing three factors: synchronization, communication, and computation. We experimentally evaluate each metric and analyze how it affects parallel simulation speed, drawing on contrasts between the large-scale IPU and smaller but faster x86 systems. Using this analysis, we build Parendi, an RTL simulator for the IPU. It distributes RTL simulation across 5888 cores on 4 IPU sockets. Parendi runs large RTL designs up to 4x faster than a powerful, state-of-the-art x86 multicore system.

Hardware isolation primitives such as secure enclaves aim to protect sensitive programs, but remain vulnerable to transient execution attacks. Complete microarchitectural isolation is not a satisfactory defense mechanism as it leaves out public shared memory, critical for usability and application performance. Conversely, hardware-software co-designs for secure speculation can counter these attacks but are not yet practical, since they make assumptions on the speculation modes, the exposed microarchitectural state, and the software, which are all hard to support for the entire software stack. This paper advocates for processors to incorporate microarchitectural isolation primitives and mechanisms for controlled speculation, enabling different execution modes. These modes can restrict what is exposed to an attacker, effectively balancing performance and program-analysis complexity. We introduce two mechanisms to securely share memory between an enclave and an untrusted OS in an out-of-order processor. We show that our two modes are complementary, achieving speculative non-interference with a reasonable performance impact, while requiring minimal code annotation and simple program analysis doable by hand. Our prototype, Citadel, is a multicore processor running on an FPGA, booting untrusted Linux, and supporting comprehensive enclave capabilities, such as shared memory, and remote attestation. To our knowledge, Citadel is the first end-to-end enclave platform to run secure applications, such as cryptographic libraries or small private inference workloads, on a speculative out-of-order multicore processor while protecting against a significant class of side-channel attacks.

The problem of Cortical Surface Reconstruction from magnetic resonance imaging has been traditionally addressed using lengthy pipelines of image processing techniques like FreeSurfer, CAT, or CIVET. These frameworks require very long runtimes deemed unfeasible for real-time applications and unpractical for large-scale studies. Recently, supervised deep learning approaches have been introduced to speed up this task cutting down the reconstruction time from hours to seconds. Using the state-of-the-art CorticalFlow model as a blueprint, this paper proposes three modifications to improve its accuracy and interoperability with existing surface analysis tools, while not sacrificing its fast inference time and low GPU memory consumption. First, we employ a more accurate ODE solver to reduce the diffeomorphic mapping approximation error. Second, we devise a routine to produce smoother template meshes avoiding mesh artifacts caused by sharp edges in CorticalFlow's convex-hull based template. Last, we recast pial surface prediction as the deformation of the predicted white surface leading to a one-to-one mapping between white and pial surface vertices. This mapping is essential to many existing surface analysis tools for cortical morphometry. We name the resulting method CorticalFlow$^{++}$. Using large-scale datasets, we demonstrate the proposed changes provide more geometric accuracy and surface regularity while keeping the reconstruction time and GPU memory requirements almost unchanged.

In this paper we introduce CorticalFlow, a new geometric deep-learning model that, given a 3-dimensional image, learns to deform a reference template towards a targeted object. To conserve the template mesh's topological properties, we train our model over a set of diffeomorphic transformations. This new implementation of a flow Ordinary Differential Equation (ODE) framework benefits from a small GPU memory footprint, allowing the generation of surfaces with several hundred thousand vertices. To reduce topological errors introduced by its discrete resolution, we derive numeric conditions which improve the manifoldness of the predicted triangle mesh. To exhibit the utility of CorticalFlow, we demonstrate its performance for the challenging task of brain cortical surface reconstruction. In contrast to current state-of-the-art, CorticalFlow produces superior surfaces while reducing the computation time from nine and a half minutes to one second. More significantly, CorticalFlow enforces the generation of anatomically plausible surfaces; the absence of which has been a major impediment restricting the clinical relevance of such surface reconstruction methods.

Instruction sets, from families like x86 and ARM, are at the center of many ambitious formal-methods projects. Many verification, synthesis, programming, and debugging tools rely on formal semantics of instruction sets, but different tools can use semantics in rather different ways. As a result, a central challenge for that community is how semantics should be written and what techniques should be used to connect them to new use cases. The best-known work applying single semantics across quite-different tools relies on domain-specific languages like Sail, where the language and its translation tools are specialized to the realm of instruction sets. We decided to explore a different approach, with semantics written in a carefully chosen subset of Haskell. This style does not depend on any new language translators, relying instead on parameterization of semantics over type-class instances. As a result, a semantics can be a first-class object within a logic, and application of a semantics for a new kind of tool can be a first-class operation in the logic, allowing sharing of theorems across applications. Our case study is for the open RISC-V instruction-set family, and we have used a single core semantics to support testing, interactive proof, and model checking of both software and hardware. We especially highlight an application of a first-class semantics within Coq that can be instantiated in different ways within one proof: simulation between variants where multiplication is implemented in hardware or in the machine code of a particular software trap handler.

The study of neurodegenerative diseases relies on the reconstruction and analysis of the brain cortex from magnetic resonance imaging (MRI). Traditional frameworks for this task like FreeSurfer demand lengthy runtimes, while its accelerated variant FastSurfer still relies on a voxel-wise segmentation which is limited by its resolution to capture narrow continuous objects as cortical surfaces. Having these limitations in mind, we propose DeepCSR, a 3D deep learning framework for cortical surface reconstruction from MRI. Towards this end, we train a neural network model with hypercolumn features to predict implicit surface representations for points in a brain template space. After training, the cortical surface at a desired level of detail is obtained by evaluating surface representations at specific coordinates, and subsequently applying a topology correction algorithm and an isosurface extraction method. Thanks to the continuous nature of this approach and the efficacy of its hypercolumn features scheme, DeepCSR efficiently reconstructs cortical surfaces at high resolution capturing fine details in the cortical folding. Moreover, DeepCSR is as accurate, more precise, and faster than the widely used FreeSurfer toolbox and its deep learning powered variant FastSurfer on reconstructing cortical surfaces from MRI which should facilitate large-scale medical studies and new healthcare applications., Comment: Accepted in 2021 IEEE Winter Conference on Applications of Computer Vision (WACV)

Brain morphometry from magnetic resonance imaging (MRI) is a consolidated biomarker for many neurodegenerative diseases. Recent advances in this domain indicate that deep convolutional neural networks can infer morphometric measurements within a few seconds. Nevertheless, the accuracy of the devised model for insightful bio-markers (mean curvature and thickness) remains unsatisfactory. In this paper, we propose a more accurate and efficient neural network model for brain morphometry named HerstonNet. More specifically, we develop a 3D ResNet-based neural network to learn rich features directly from MRI, design a multi-scale regression scheme by predicting morphometric measures at feature maps of different resolutions, and leverage a robust optimization method to avoid poor quality minima and reduce the prediction variance. As a result, HerstonNet improves the existing approach by 24.30% in terms of intraclass correlation coefficient (agreement measure) to FreeSurfer silver-standards while maintaining a competitive run-time.

This paper describes two locally-differential private algorithms for releasing user vectors such that the Jaccard similarity between these vectors can be efficiently estimated. The basic building block is the well known MinHash method. To achieve a privacy-utility trade-off, MinHash is extended in two ways using variants of Generalized Randomized Response and the Laplace Mechanism. A theoretical analysis provides bounds on the absolute error and experiments show the utility-privacy trade-off on synthetic and real-world data. The paper ends with a critical discussion of related work., Comment: Accepted at SISAP 2020

Alzheimer's disease is the most common neurodegenerative disease. The aim of this study is to infer structural changes in brain connectivity resulting from disease progression using cortical thickness measurements from a cohort of participants who were either healthy control, or with mild cognitive impairment, or Alzheimer's disease patients. For this purpose, we develop a novel approach for inference of multiple networks with related edge values across groups. Specifically, we infer a Gaussian graphical model for each group within a joint framework, where we rely on Bayesian hierarchical priors to link the precision matrix entries across groups. Our proposal differs from existing approaches in that it flexibly learns which groups have the most similar edge values, and accounts for the strength of connection (rather than only edge presence or absence) when sharing information across groups. Our results identify key alterations in structural connectivity which may reflect disruptions to the healthy brain, such as decreased connectivity within the occipital lobe with increasing disease severity. We also illustrate the proposed method through simulations, where we demonstrate its performance in structure learning and precision matrix estimation with respect to alternative approaches., Comment: Accepted to Biometrics January 2020

Longitudinal analysis has great potential to reveal developmental trajectories and monitor disease progression in medical imaging. This process relies on consistent and robust joint 4D segmentation. Traditional techniques are dependent on the similarity of images over time and the use of subject-specific priors to reduce random variation and improve the robustness and sensitivity of the overall longitudinal analysis. This is however slow and computationally intensive as subject-specific templates need to be rebuilt every time. The focus of this work to accelerate this analysis with the use of deep learning. The proposed approach is based on deep CNNs and incorporates semantic segmentation and provides a longitudinal relationship for the same subject. The proposed approach is based on deep CNNs and incorporates semantic segmentation and provides a longitudinal relationship for the same subject. The state of art using 3D patches as inputs to modified Unet provides results around ${0.91 \pm 0.5}$ Dice and using multi-view atlas in CNNs provide around the same results. In this work, different models are explored, each offers better accuracy and fast results while increasing the segmentation quality. These methods are evaluated on 135 scans from the EADC-ADNI Harmonized Hippocampus Protocol. Proposed CNN based segmentation approaches demonstrate how 2D segmentation using prior slices can provide similar results to 3D segmentation while maintaining good continuity in the 3D dimension and improved speed. Just using 2D modified sagittal slices provide us a better Dice and longitudinal analysis for a given subject. For the ADNI dataset, using the simple UNet CNN technique gives us ${0.84 \pm 0.5}$ and while using modified CNN techniques on the same input yields ${0.89 \pm 0.5}$. Rate of atrophy and RMS error are calculated for several test cases using various methods and analyzed.

Recent attacks have broken process isolation by exploiting microarchitectural side channels that allow indirect access to shared microarchitectural state. Enclaves strengthen the process abstraction to restore isolation guarantees. We propose MI6, an aggressive, speculative out-of-order processor capable of providing secure enclaves under a threat model that includes an untrusted OS and an attacker capable of mounting any software attack currently considered practical, including control flow speculation attacks. MI6 is inspired by Sanctum [16] and extends its isolation guarantee to more realistic memory hierarchies. It also introduces a purge instruction, which is used only when a secure process is scheduled, and implements it for a complex processor microarchitecture. We model the performance impact of enclaves in MI6 through FPGA emulation on AWS F1 FPGAs by running SPEC CINT2006 benchmarks on top of an untrusted Linux OS. Security comes at the cost of approximately 16.4% average slowdown for protected programs., Comment: 15 pages

In simulations and experiments, we study the drying of films containing mixtures of large and small colloidal particles in water. During drying, the mixture stratifies into a layer of the larger particles at the bottom with a layer of the smaller particles on top. We developed a model to show that a gradient in osmotic pressure, which develops dynamically during drying, is responsible for the segregation mechanism behind stratification., Comment: Accepted for publication in Physical Review Letters. Supplementary Material added as Appendix

If you color a table using k colors, and throw a needle randomly on it, for some proper definition, you get a certain probability that the endpoints will fall on different colors. How can one make this probability maximal? This problem is related to finite graphs having unit-length edges, and some bounds on the optimal probability are deduced., Comment: 11 pages, 5 figures

Point constellation recognition is a common problem with many pattern matching applications. Whilst useful in many contexts, this work is mainly motivated by fingerprint matching. Fingerprints are traditionally modelled as constellations of oriented points called minutiae. The fingerprint verifier's task consists in comparing two point constellations. The compared constellations may differ by rotation and translation or by much more involved transforms such as distortion or occlusion. This paper presents three new constellation matching algorithms. The first two methods generalize an algorithm by Bringer and Despiegel. Our third proposal creates a very interesting analogy between mechanical system simulation and the constellation recognition problem., Comment: 14 pages, short version submitted to SEC 2014

A 3-day international workshop on atmospheric monitoring and calibration for high-energy astroparticle detectors, with a view towards next-generation facilities. The atmosphere is an integral component of many high-energy astroparticle detectors. Imaging atmospheric Cherenkov telescopes and cosmic-ray extensive air shower detectors are the two instruments driving the rapidly evolving fields of very-high- and ultra-high-energy astrophysics. In these instruments, the atmosphere is used as a giant calorimeter where cosmic rays and gamma rays deposit their energy and initiate EASs; it is also the medium through which the resulting Cherenkov light propagates. Uncertainties in real-time atmospheric conditions and in the fixed atmospheric models typically dominate all other systematic errors. With the improved sensitivity of upgraded IACTs such as H.E.S.S.-II and MAGIC-II and future facilities like the Cherenkov Telescope Array (CTA) and JEM-EUSO, statistical uncertainties are expected to be significantly reduced, leaving the atmosphere as the limiting factor in the determination of astroparticle spectra. Varying weather conditions necessitate the development of suitable atmospheric monitoring to be integrated in the overall instrument calibration, including Monte Carlo simulations. With expertise distributed across multiple collaborations and scientific domains, an interdisciplinary workshop is being convened to advance progress on this critical and timely topic., Comment: Proceedings of the Atmospheric Monitoring for High-Energy Astroparticle Detectors (AtmoHEAD) Conference, Saclay (France), June 10-12, 2013

The H.E.S.S. experiment in Namibia, Africa, is designed to study the origin of high energy cosmic rays from 100 Gev to few tens of TeV, using the Cherenkov technique. To minimize the systematic errors on the derived fluxes of the measured sources, one has to calculate the impact of the atmospheric properties, namely the extinction parameter a. A LIDAR can provide this kind of information within the detectable energy range of the experiment. In this paper we report on the hardware components, operation and data taking of such a system installed on the HESS site for the last three years.

Compilation of CTA contributions to the proceedings of the 33rd International Cosmic Ray Conference (ICRC2013), which took place in 2-9 July, 2013, in Rio de Janeiro, Brazil, Comment: Index of CTA conference proceedings at the ICRC2013, Rio de Janeiro (Brazil). v1: placeholder with no arXiv links yet, to be replaced once individual contributions have been all submitted. v2: final with arXiv links to all CTA contributions and full author list

We present a compositional compressible two-phase, liquid and gas, flow model for numerical simulations of hydrogen migration in deep geological radioactive waste repository. This model includes capillary effects and the gas diffusivity. The choice of the main variables in this model, Total or Dissolved Hydrogen Mass Concentration and Liquid Pressure, leads to a unique and consistent formulation of the gas phase appearance and disappearance. After introducing this model, we show computational evidences of its adequacy to simulate gas phase appearance and disappearance in different situations typical of underground radioactive waste repository.

We derive a compositional compressible two-phase, liquid and gas, flow model for numerical simulations of hydrogen migration in deep geological repository for radioactive waste. This model includes capillary effects and the gas high diffusivity. Moreover, it is written in variables (total hydrogen mass density and liquid pressure) chosen in order to be consistent with gas appearance or disappearance. We discuss the well possedness of this model and give some computational evidences of its adequacy to simulate gas generation in a water saturated repository.

Homogenization method is used to analyse the equivalent behavior of transient flow of a passive solute through highly heterogeneous porous media. The flow is governed by a coupled system which includes an elliptic equation and a linear convection-diffusion concentration equation with a diffusion term small with respect to the convection, i.e. a high Peclet number is considered. Asymptotic expansions lead to the definition of a macroscale transport model. Numerical computations to obtain the effective hydraulic conductivity and the macrodiffusivity tensor have been carried out via finite volume methods. Numerical experiments related to the simulations of solute transport have been performed comparing the heterogeneous medium to the corresponding effective medium. The results of the simulations are compared in terms of spatial moments, L2 errors and concentration contours. Results obtained from the simulations using the model obtained by the homogenization method show good agreement with the heterogeneous simulations., Comment: This paper has been withdrawn due to the copyrights

The goal of this paper is to give a mathematical model describing the global be haviour of the nuclear waste disposal process.The physical situation can be described as an array made of high number of alveoles inside of a low permeable layer (e.g. clay) included between two layers with slightly higher permeability (e.g. limestone). Radioactive elements are leaking from their containers over a period of time ]0,t_m[ . In a porous media (clay) there is a dilution effect (similar to diffusion) and convection due to some underground water flow. The real physical situation is simplified by replacing 5 alveoles by one macro alveole and by considering the convection velocity as a given field. These simplifications seam to be unimportant for understanding the global effects. The typical size of such macro alveole is of order : 100 m width, 1000 m length, 5 m hight. The distance between two alveoles is also of order 100 m. The size of law permeable layer is of order 150 m hight, and 3000 length. Since alveoles are small compared to the size of layer and their number is large direct numerical simulations using the {\em microscopic} model is unrealistic. On the other hand the ratio between domain length and the length of one alveole is small, of order 1/30, and allows an asymptotic study with respect to that small parameter, denoted by \epsilon, using the method of homogenization and boundary layers., Comment: One file, latex, corrected version, 12 pages

In this paper we study the filtration laws for the polymeric flow in a porous medium. We use the quasi-Newtonian models with share dependent viscosity obeying the power-law and the Carreau's law. Using the method of homogenization the coupled micro-macro homogenized law, governing the quasi-newtonian flow in a periodic model of a porous medium, was found. We decouple that law separating the micro from the macro scale. We write the macroscopic filtration law in the form of non-linear Darcy's law and we prove that the obtained law is well posed. We give the analytical as well as the numerical study of our model., Comment: 22 pages, 7 figures, Latex files, main file: main_file.tex which call all other files