Search

Diffusion MRI is a powerful tool that serves as a bridge between brain microstructure and cognition. Recent advancements in cognitive neuroscience have highlighted the persistent challenge of understanding how individual differences in brain structure influence behavior, especially in healthy people. While traditional linear models like Canonical Correlation Analysis (CCA) and Partial Least Squares (PLS) have been fundamental in this analysis, they face limitations, particularly with high-dimensional data analysis outside the training sample. To address these issues, we introduce a novel approach using deep learninga multivariate autoencoder model-to explore the complex non-linear relationships between brain microstructure and cognitive functions. The model's architecture involves separate encoder modules for brain structure and cognitive data, with a shared decoder, facilitating the analysis of multivariate patterns across these domains. Both encoders were trained simultaneously, before the decoder, to ensure a good latent representation that captures the phenomenon. Using data from the Human Connectome Project, our study centres on the insula's role in cognitive processes. Through rigorous validation, including 5 sample analyses for out-of-sample analysis, our results demonstrate that the multivariate autoencoder model outperforms traditional methods in capturing and generalizing correlations between brain and behavior beyond the training sample. These findings underscore the potential of deep learning models to enhance our understanding of brain-behavior relationships in cognitive neuroscience, offering more accurate and comprehensive insights despite the complexities inherent in neuroimaging studies.

The estimation of directed couplings between the nodes of a network from indirect measurements is a central methodological challenge in scientific fields such as neuroscience, systems biology and economics. Unfortunately, the problem is generally ill-posed due to the possible presence of unknown delays in the measurements. In this paper, we offer a solution of this problem by using a variational Bayes framework, where the uncertainty over the delays is marginalized in order to obtain conservative coupling estimates. To overcome the well-known overconfidence of classical variational methods, we use a hybrid-VI scheme where the (possibly flat or multimodal) posterior over the measurement parameters is estimated using a forward KL loss while the (nearly convex) conditional posterior over the couplings is estimated using the highly scalable gradient-based VI. In our ground-truth experiments, we show that the network provides reliable and conservative estimates of the couplings, greatly outperforming similar methods such as regression DCM.

In 1982, Cameron and Liebler investigated certain "special sets of lines" in PG(3,q), and gave several equivalent characterizations. Due to their interesting geometric and algebraic properties, these "Cameron-Liebler line classes" got a lot of attention. Several generalizations and variants have been considered in the literature, the main directions being a variation of the dimensions of the involved spaces, and studying the analogous situation in the subset lattice. An important tool is the interpretation of the objects as Boolean functions in the "Johnson" and "q-Johnson schemes". In this article, we develop a unified theory covering all these variations. Generalized versions of algebraic and geometric properties will be investigated, having a parallel in the notion of "designs" and "antidesigns" in association schemes, which is connected to Delsarte's concept of "design-orthogonality". This leads to a natural definition of the "degree" and the "weights" of functions in the ambient scheme, refining the existing definitions. We will study the effect of dualization and of elementary modifications of the ambient space on the degree and the weights. Moreover, a divisibility property of the sizes of Boolean functions of degree t will be proven.

Combinatorial designs have been studied for nearly 200 years. 50 years ago, Cameron, Delsarte, and Ray-Chaudhury started investigating their $q$-analogs, also known as subspace designs or designs over finite fields. Designs can be defined analogously in finite classical polar spaces, too. The definition includes the $m$-regular systems from projective geometry as the special case where the blocks are generators of the polar space. The first nontrivial such designs for $t > 1$ were found by De Bruyn and Vanhove in 2012, and some more designs appeared recently in the PhD thesis of Lansdown. In this article, we investigate the theory of classical and subspace designs for applicability to designs in polar spaces, explicitly allowing arbitrary block dimensions. In this way, we obtain divisibility conditions on the parameters, derived and residual designs, intersection numbers and an analog of Fisher's inequality. We classify the parameters of symmetric designs. Furthermore, we conduct a computer search to construct designs of strength $t=2$, resulting in designs for more than 140 previously unknown parameter sets in various classical polar spaces over $\mathbb{F}_2$ and $\mathbb{F}_3$.

Previous work on additively-manufactured oxide dispersion strengthened alloys focused on experimental approaches, resulting in larger dispersoid sizes and lower number densities than can be achieved with conventional powder metallurgy. To improve the as-fabricated microstructure, this work integrates experiments with a thermodynamic and kinetic modeling framework to probe the limits of the dispersoid sizes and number densities that can be achieved with powder bed fusion-laser beam. Bulk samples of a Ni-20Cr $+$ 1 wt.% Y$_2$O$_3$ alloy are fabricated using a range of laser power and scanning velocity combinations. Scanning transmission electron microscopy characterization is performed to quantify the dispersoid size distributions across the processing space. The smallest mean dispersoid diameter (29 nm) is observed at 300 W and 1200 mm/s, with a number density of 1.0$\times$10$^{20}$ m$^{-3}$. The largest mean diameter (72 nm) is observed at 200 W and 200 mm/s, with a number density of 1.5$\times$10$^{19}$ m$^{-3}$. Scanning electron microscopy suggests that a considerable fraction of the oxide added to the feedstock is lost during processing, due to oxide agglomeration and the ejection of oxide-rich spatter from the melt pool. After accounting for these losses, the model predictions for the dispersoid diameter and number density align with the experimental trends. The results suggest that the mechanism that limits the final number density is collision coarsening of dispersoids in the melt pool. The modeling framework is leveraged to propose processing strategies to limit dispersoid size and increase number density., Comment: Main text: 37 pages, 20 figures

In today's digital world, cybercrime is responsible for significant damage to organizations, including financial losses, operational disruptions, or intellectual property theft. Cyberattacks often start with an email, the major means of corporate communication. Some rare, severely damaging email threats - known as spear phishing or Business Email Compromise - have emerged. However, the literature disagrees on their definition, impeding security vendors and researchers from mitigating targeted attacks. Therefore, we introduce targeted attacks. We describe targeted-attack-detection techniques as well as social-engineering methods used by fraudsters. Additionally, we present text-based attacks - with textual content as malicious payload - and compare non-targeted and targeted variants.

Given observed data and a probabilistic generative model, Bayesian inference searches for the distribution of the model's parameters that could have yielded the data. Inference is challenging for large population studies where millions of measurements are performed over a cohort of hundreds of subjects, resulting in a massive parameter space. This large cardinality renders off-the-shelf Variational Inference (VI) computationally impractical. In this work, we design structured VI families that efficiently tackle large population studies. Our main idea is to share the parameterization and learning across the different i.i.d. variables in a generative model, symbolized by the model's \textit{plates}. We name this concept \textit{plate amortization}. Contrary to off-the-shelf stochastic VI, which slows down inference, plate amortization results in orders of magnitude faster to train variational distributions. Applied to large-scale hierarchical problems, PAVI yields expressive, parsimoniously parameterized VI with an affordable training time. This faster convergence effectively unlocks inference in those large regimes. We illustrate the practical utility of PAVI through a challenging Neuroimaging example featuring 400 million latent parameters, demonstrating a significant step towards scalable and expressive Variational Inference.

In 1985, Janko and Tran Van Trung published an algorithm for constructing symmetric designs with prescribed automorphisms. This algorithm is based on the equations by Dembowski (1958) for tactical decompositions of point-block incidence matrices. In the sequel, the algorithm has been generalized and improved in many articles. In parallel, higher incidence matrices have been introduced by Wilson in 1982. They have proven useful for obtaining several restrictions on the existence of designs. For example, a short proof of the generalized Fisher's inequality makes use of these incidence matrices. In this paper, we introduce a unified approach to tactical decompositions and incidence matrices. It works for both combinatorial and subspace designs alike. As a result, we obtain a generalized Fisher's inequality for tactical decompositions of combinatorial and subspace designs. Moreover, our approach is explored for the construction of combinatorial and subspace designs of arbitrary strength.

Given some observed data and a probabilistic generative model, Bayesian inference aims at obtaining the distribution of a model's latent parameters that could have yielded the data. This task is challenging for large population studies where thousands of measurements are performed over a cohort of hundreds of subjects, resulting in a massive latent parameter space. This large cardinality renders off-the-shelf Variational Inference (VI) computationally impractical. In this work, we design structured VI families that can efficiently tackle large population studies. To this end, our main idea is to share the parameterization and learning across the different i.i.d. variables in a generative model -symbolized by the model's plates. We name this concept plate amortization, and illustrate the powerful synergies it entitles, resulting in expressive, parsimoniously parameterized and orders of magnitude faster to train large scale hierarchical variational distributions. We illustrate the practical utility of PAVI through a challenging Neuroimaging example featuring a million latent parameters, demonstrating a significant step towards scalable and expressive Variational Inference.

The silicification of DNA origami structures increases their mechanical and thermal stability and provides chemical protection. So far, it is unclear how silicification affects the internal structure of the DNA origami and whether the whole DNA framework is embedded or if silica just forms an outer shell. By using in situ small-angle X-ray scattering (SAXS), we show that the net-cationic silica precursor TMAPS induces substantial condensation of the DNA origami, which is further enhanced by the addition of TEOS at early reaction times to an almost 10 % size reduction. We identify the SAXS Porod invariant as a reliable, model-free parameter for the evaluation of the amount of silica formation at a given time. Contrast matching of the DNA double helix Lorentzian peak reveals that silica growth also occurs on the inner surfaces of the origami. The less polar silica forming within the origami structure, replacing more than 40 % of the internal hydration water causes a hydrophobic effect: origami condensation. In the maximally condensed state, thermal stabilization of the origami up to 60 {\deg}C could be observed. If the reaction is driven beyond this point, the overall size of the silicified origami increases again due to more and more silica deposition on the DNA origami. DNA origami objects with flat surfaces show a strong tendency towards aggregation during silicification, presumably driven by the very same entropic forces causing condensation. Our studies provide novel insights into the silicification reaction and hints for the formulation of optimized reaction protocols.

Researchers in neuroscience have a growing number of datasets available to study the brain, which is made possible by recent technological advances. Given the extent to which the brain has been studied, there is also available ontological knowledge encoding the current state of the art regarding its different areas, activation patterns, key words associated with studies, etc. Furthermore, there is an inherent uncertainty associated with brain scans arising from the mapping between voxels -- 3D pixels -- and actual points in different individual brains. Unfortunately, there is currently no unifying framework for accessing such collections of rich heterogeneous data under uncertainty, making it necessary for researchers to rely on ad hoc tools. In particular, one major weakness of current tools that attempt to address this kind of task is that only very limited propositional query languages have been developed. In this paper, we present NeuroLang, an ontology language with existential rules, probabilistic uncertainty, and built-in mechanisms to guarantee tractable query answering over very large datasets. After presenting the language and its general query answering architecture, we discuss real-world use cases showing how NeuroLang can be applied to practical scenarios for which current tools are inadequate.

Multivariate anomaly detection can be used to identify outages within large volumes of telemetry data for computing systems. However, developing an efficient anomaly detector that can provide users with relevant information is a challenging problem. We introduce our approach to hierarchical multivariate anomaly detection called DeCorus, a statistical multivariate anomaly detector which achieves linear complexity. It extends standard statistical techniques to improve their ability to find relevant anomalies within noisy signals and makes use of types of domain knowledge that system operators commonly possess to compute system-level anomaly scores. We describe the implementation of DeCorus an online log anomaly detection tool for network device syslog messages deployed at a cloud service provider. We use real-world data sets that consist of $1.5$ billion network device syslog messages and hundreds of incident tickets to characterize the performance of DeCorus and compare its ability to detect incidents with five alternative anomaly detectors. While DeCorus outperforms the other anomaly detectors, all of them are challenged by our data set. We share how DeCorus provides value in the field and how we plan to improve its incident detection accuracy., Comment: 11 pages, 4 figures, draft

Effective characterisation of the brain grey matter cytoarchitecture with quantitative sensitivity to soma density and volume remains an unsolved challenge in diffusion MRI (dMRI). Solving the problem of relating the dMRI signal with cytoarchitectural characteristics calls for the definition of a mathematical model that describes brain tissue via a handful of physiologically-relevant parameters and an algorithm for inverting the model. To address this issue, we propose a new forward model, specifically a new system of equations, requiring a few relatively sparse b-shells. We then apply modern tools from Bayesian analysis known as likelihood-free inference (LFI) to invert our proposed model. As opposed to other approaches from the literature, our algorithm yields not only an estimation of the parameter vector $\theta$ that best describes a given observed data point $x_0$, but also a full posterior distribution $p(\theta|x_0)$ over the parameter space. This enables a richer description of the model inversion, providing indicators such as credible intervals for the estimated parameters and a complete characterization of the parameter regions where the model may present indeterminacies. We approximate the posterior distribution using deep neural density estimators, known as normalizing flows, and fit them using a set of repeated simulations from the forward model. We validate our approach on simulations using dmipy and then apply the whole pipeline on two publicly available datasets.

Inferring which pathways are affected by a brain lesion is key for both pre and post-treatment planning. However, many disruptive lesions cause changes in the tissue that interrupt tractography algorithms. In such cases, aggregating information from healthy subjects can provide a solution to inferring the affected pathways. In this paper, we introduce a novel label fusion technique that leverages diffusion information to locate brain pathways. Through simulations and experiments in publicly available data we show that our method is able to correctly reconstruct brain pathways, even if they are affected by a focal lesion.

Frequently, population studies feature pyramidally-organized data represented using Hierarchical Bayesian Models (HBM) enriched with plates. These models can become prohibitively large in settings such as neuroimaging, where a sample is composed of a functional MRI signal measured on 300 brain locations, across 4 measurement sessions, and 30 subjects, resulting in around 1 million latent parameters.Such high dimensionality hampers the usage of modern, expressive flow-based techniques.To infer parameter posterior distributions in this challenging class of problems, we designed a novel methodology that automatically produces a variational family dual to a target HBM. This variational family, represented as a neural network, consists in the combination of an attention-based hierarchical encoder feeding summary statistics to a set of normalizing flows. Our automatically-derived neural network exploits exchangeability in the plate-enriched HBM and factorizes its parameter space. The resulting architecture reduces by orders of magnitude its parameterization with respect to that of a typical flow-based representation, while maintaining expressivity.Our method performs inference on the specified HBM in an amortized setup: once trained, it can readily be applied to a new data sample to compute the parameters' full posterior.We demonstrate the capability and scalability of our method on simulated data, as well as a challenging high-dimensional brain parcellation experiment. We also open up several questions that lie at the intersection between normalizing flows, SBI, structured Variational Inference, and inference amortization.

Strongly walk regular graphs (SWRGs or $s$-SWRGs) form a natural generalization of strongly regular graphs (SRGs) where paths of length~2 are replaced by paths of length~$s$. They can be constructed as coset graphs of the duals of projective three-weight codes whose weights satisfy a certain equation. We provide classifications of the feasible parameters of these codes in the binary and ternary case for medium size code lengths. For the binary case, the divisibility of the weights of these codes is investigated and several general results are shown. It is known that an $s$-SWRG has at most 4 distinct eigenvalues $k > \theta_1 > \theta_2 > \theta_3$, and that the triple $(\theta_1, \theta_2, \theta_3)$ satisfies a certain homogeneous polynomial equation of degree $s - 2$ (Van Dam, Omidi, 2013). This equation defines a plane algebraic curve; we use methods from algorithmic arithmetic geometry to show that for $s = 5$ and $s = 7$, there are only the obvious solutions, and we conjecture this to remain true for all (odd) $s \ge 9$., Comment: 42 pages

With the growing number of published functional magnetic resonance imaging (fMRI) studies, meta-analysis databases and models have become an integral part of brain mapping research. Coordinate-based meta-analysis (CBMA) databases are built by automatically extracting both coordinates of reported peak activations and term associations using natural language processing (NLP) techniques. Solving term-based queries on these databases make it possible to obtain statistical maps of the brain related to specific cognitive processes. However, with tools like Neurosynth, only singleterm queries lead to statistically reliable results. When solving richer queries, too few studies from the database contribute to the statistical estimations. We design a probabilistic domain-specific language (DSL) standing on Datalog and one of its probabilistic extensions, CP-Logic, for expressing and solving rich logic-based queries. We encode a CBMA database into a probabilistic program. Using the joint distribution of its Bayesian network translation, we show that solutions of queries on this program compute the right probability distributions of voxel activations. We explain how recent lifted query processing algorithms make it possible to scale to the size of large neuroimaging data, where state of the art knowledge compilation (KC) techniques fail to solve queries fast enough for practical applications. Finally, we introduce a method for relating studies to terms probabilistically, leading to better solutions for conjunctive queries on smaller databases. We demonstrate results for two-term conjunctive queries, both on simulated meta-analysis databases and on the widely-used Neurosynth database.

We prove equality of analytic and topological $L^2$-torsion associated with an odd-dimensional finite volume hyperbolic manifold and a representation of the fundamental group which extends to the ambient Lie group. This generalizes a previous result due to L\"uck and Schick. Alternatively, this result can be regarded as the $L^2$-analogue of recent work by M\"uller and Rochon., Comment: 35 pages, 3 figures

Ensembling is now recognized as an effective approach for increasing the predictive performance and calibration of deep networks. We introduce a new approach, Parameter Ensembling by Perturbation (PEP), that constructs an ensemble of parameter values as random perturbations of the optimal parameter set from training by a Gaussian with a single variance parameter. The variance is chosen to maximize the log-likelihood of the ensemble average ($\mathbb{L}$) on the validation data set. Empirically, and perhaps surprisingly, $\mathbb{L}$ has a well-defined maximum as the variance grows from zero (which corresponds to the baseline model). Conveniently, calibration level of predictions also tends to grow favorably until the peak of $\mathbb{L}$ is reached. In most experiments, PEP provides a small improvement in performance, and, in some cases, a substantial improvement in empirical calibration. We show that this "PEP effect" (the gain in log-likelihood) is related to the mean curvature of the likelihood function and the empirical Fisher information. Experiments on ImageNet pre-trained networks including ResNet, DenseNet, and Inception showed improved calibration and likelihood. We further observed a mild improvement in classification accuracy on these networks. Experiments on classification benchmarks such as MNIST and CIFAR-10 showed improved calibration and likelihood, as well as the relationship between the PEP effect and overfitting; this demonstrates that PEP can be used to probe the level of overfitting that occurred during training. In general, no special training procedure or network architecture is needed, and in the case of pre-trained networks, no additional training is needed., Comment: NeurIPS 2020

This paper proposes an extension of the finite cell method (FCM) to V-rep models, a novel geometric framework for volumetric representations. This combination of an embedded domain approach (FCM) and a new modeling framework (V-rep) forms the basis for an efficient and accurate simulation of mechanical artifacts, which are not only characterized by complex shapes but also by their non-standard interior structure. These objects gain more and more interest in the context of the new design opportunities opened by additive manufacturing, mainly when graded or micro-structured material is applied. Two different types of functionally graded materials (FGM) are considered: The first one, multi-material FGM, is described using the V-rep models' inherent property to assign different properties throughout the interior of a domain. The second, single-material FGM -- which is heterogeneously micro-structured -- characterizes the effective material behavior of representative volume elements by homogenization and performs large-scale simulations using the embedded domain approach., Comment: Preprint submitted to Advanced Modeling and Simulation in Engineering Sciences

We generalize a Cheeger-M\"uller type theorem for flat, unitary bundles on infinite covering spaces over manifolds-with-boundary, proven by Burghelea, Friedlander and Kappeller arXiv:dg-ga/9510010 [math.DG]. Employing recent anomaly results by Br\"uning, Ma and Zhang, we prove an analogous statement for a general flat bundle that is only required to have a unimodular restriction to the boundary., Comment: 34 pages

The fundamental changes in power supply and increasing decentralization require more active grid operation and an increased integration of ICT at all power system actors. This trend raises complexity and increasingly leads to interactions between primary grid operation and ICT as well as different power system actors. For example, virtual power plants control various assets in the distribution grid via ICT to jointly market existing flexibilities. Failures of ICT or targeted attacks can thus have serious effects on security of supply and system stability. This paper presents a holistic approach to providing methods specifically for actors in the power system for prevention, detection, and reaction to ICT attacks and failures. The focus of our measures are solutions for ICT monitoring, systems for the detection of ICT attacks and intrusions in the process network, and the provision of actionable guidelines as well as a practice environment for the response to potential ICT security incidents., Comment: 6 pages, 4 figures, to be published in Proceedings of the 2020 6th IEEE International Energy Conference (ENERGYCon)

Population imaging markedly increased the size of functional-imaging datasets, shedding new light on the neural basis of inter-individual differences. Analyzing these large data entails new scalability challenges, computational and statistical. For this reason, brain images are typically summarized in a few signals, for instance reducing voxel-level measures with brain atlases or functional modes. A good choice of the corresponding brain networks is important, as most data analyses start from these reduced signals. We contribute finely-resolved atlases of functional modes, comprising from 64 to 1024 networks. These dictionaries of functional modes (DiFuMo) are trained on millions of fMRI functional brain volumes of total size 2.4TB, spanned over 27 studies and many research groups. We demonstrate the benefits of extracting reduced signals on our fine-grain atlases for many classic functional data analysis pipelines: stimuli decoding from 12,334 brain responses, standard GLM analysis of fMRI across sessions and individuals, extraction of resting-state functional-connectomes biomarkers for 2,500 individuals, data compression and meta-analysis over more than 15,000 statistical maps. In each of these analysis scenarii, we compare the performance of our functional atlases with that of other popular references, and to a simple voxel-level analysis. Results highlight the importance of using high-dimensional "soft" functional atlases, to represent and analyse brain activity while capturing its functional gradients. Analyses on high-dimensional modes achieve similar statistical performance as at the voxel level, but with much reduced computational cost and higher interpretability. In addition to making them available, we provide meaningful names for these modes, based on their anatomical location. It will facilitate reporting of results.