Search

In particle physics, homogeneous calorimeters are used to measure the energy of particles as they interact with the detector material. Although not as precise as trackers or muon detectors, these calorimeters provide valuable insights into the properties of particles by analyzing their energy deposition patterns. Recent advances in material science, notably in nanomaterial scintillators with tunable emission bandwidths, have led to the proposal of the chromatic calorimetry concept. This proposed concept aims to track electromagnetic or hadronic shower progression within a module, enhancing particle identification and energy resolution by layering scintillators with different emission wavelengths. The idea is to use the emission spectra of the inorganic scintillators to reconstruct the shower progression. Our study validates this proposed concept using inorganic scintillators strategically stacked by decreasing emission wavelength. Using electrons and pions with up to 100 GeV, we achieved analytical discrimination and longitudinal shower measurement. This proof of concept underscores chromatic calorimetry's potential for broader applications.

We propose a difference-in-differences (DiD) method for a time-varying continuous treatment and multiple time periods. Our framework assesses the average treatment effect on the treated (ATET) when comparing two non-zero treatment doses. The identification is based on a conditional parallel trend assumption imposed on the mean potential outcome under the lower dose, given observed covariates and past treatment histories. We employ kernel-based ATET estimators for repeated cross-sections and panel data adopting the double/debiased machine learning framework to control for covariates and past treatment histories in a data-adaptive manner. We also demonstrate the asymptotic normality of our estimation approach under specific regularity conditions. In a simulation study, we find a compelling finite sample performance of undersmoothed versions of our estimators in setups with several thousand observations.

In recent years, researchers have been exploring ways to generalize Boltzmann machines (BMs) to quantum systems, leading to the development of variations such as fully-visible and restricted quantum Boltzmann machines (QBMs). Due to the non-commuting nature of their Hamiltonians, restricted QBMs face trainability issues, whereas fully-visible QBMs have emerged as a more tractable option, as recent results demonstrate their sample-efficient trainability. These results position fully-visible QBMs as a favorable choice, offering potential improvements over fully-visible BMs without suffering from the trainability issues associated with restricted QBMs. In this work, we show that low-dimensional, fully-visible QBMs can learn to generate distributions typically associated with higher-dimensional systems. We validate our findings through numerical experiments on both artificial datasets and real-world examples from the high energy physics problem of jet event generation. We find that non-commuting terms and Hamiltonian connectivity improve the learning capabilities of QBMs, providing flexible resources suitable for various hardware architectures. Furthermore, we provide strategies and future directions to maximize the learning capacity of fully-visible QBMs., Comment: 13 pages, 7 figures; supplementary material 5 pages, 5 figures

The study presents a novel approach for stochastic nonlinear model updating in structural dynamics, employing a Bayesian framework integrated with Markov Chain Monte Carlo (MCMC) sampling for parameter estimation by using an approximated likelihood function. The proposed methodology is applied to both numerical and experimental cases. The paper commences by introducing Bayesian inference and its constituents: the likelihood function, prior distribution, and posterior distribution. The resonant decay method is employed to extract backbone curves, which capture the non-linear behaviour of the system. A mathematical model based on a single degree of freedom (SDOF) system is formulated, and backbone curves are obtained from time response data. Subsequently, MCMC sampling is employed to estimate the parameters using both numerical and experimental data. The obtained results demonstrate the convergence of the Markov chain, present parameter trace plots, and provide estimates of posterior distributions of updated parameters along with their uncertainties. Experimental validation is performed on a cantilever beam system equipped with permanent magnets and electromagnets. The proposed methodology demonstrates promising results in estimating parameters of stochastic non-linear dynamical systems. Through the use of the proposed likelihood functions using backbone curves, the probability distributions of both linear and non-linear parameters are simultaneously identified. Based on this view, the necessity to segregate stochastic linear and non-linear model updating is eliminated.

Thyroid carcinoma, a significant yet often controllable cancer, has seen a rise in cases, largely due to advancements in diagnostic methods. Differentiated thyroid cancer (DTC), which includes papillary and follicular varieties, is typically associated with a positive prognosis in academic circles. Nevertheless, there are still some individuals who may experience a recurrence. This study employs machine learning, particularly deep learning models, to predict the recurrence of DTC, with the goal of improving patient care through personalized treatment approaches. By analysing a dataset containing clinicopathological features of patients, the model achieved remarkable accuracy rates of 98% during training and 96% during testing. To improve the model's interpretability, we used techniques like LIME and Morris Sensitivity Analysis. These methods gave us valuable insights into how the model makes decisions. The results suggest that combining deep learning models with interpretability techniques can be extremely useful in quickly identifying the recurrence of thyroid cancer in patients. This can help in making informed therapeutic choices and customizing treatment approaches for individual patients.

Verification of infinite-state Markov chains is still a challenge despite several fruitful numerical or statistical approaches. For decisive Markov chains, there is a simple numerical algorithm that frames the reachability probability as accurately as required (however with an unknown complexity). On the other hand when applicable, statistical model checking is in most of the cases very efficient. Here we study the relation between these two approaches showing first that decisiveness is a necessary and sufficient condition for almost sure termination of statistical model checking. Afterwards we develop an approach with application to both methods that substitutes to a non decisive Markov chain a decisive Markov chain with the same reachability probability. This approach combines two key ingredients: abstraction and importance sampling (a technique that was formerly used for efficiency). We develop this approach on a generic formalism called layered Markov chain (LMC). Afterwards we perform an empirical study on probabilistic pushdown automata (an instance of LMC) to understand the complexity factors of the statistical and numerical algorithms. To the best of our knowledge, this prototype is the first implementation of the deterministic algorithm for decisive Markov chains and required us to solve several qualitative and numerical issues., Comment: 26 pages, 3 figures, to appear in proceeding of FSTTCS24

Automatic Speech Recognition (ASR) systems have achieved remarkable performance on widely used benchmarks such as LibriSpeech and Fleurs. However, these benchmarks do not adequately reflect the complexities of real-world conversational environments, where speech is often unstructured and contains disfluencies such as pauses, interruptions, and diverse accents. In this study, we introduce a multilingual conversational dataset, derived from TalkBank, consisting of unstructured phone conversation between adults. Our results show a significant performance drop across various state-of-the-art ASR models when tested in conversational settings. Furthermore, we observe a correlation between Word Error Rate and the presence of speech disfluencies, highlighting the critical need for more realistic, conversational ASR benchmarks.

Large language models (LLMs) have enabled a range of applications in zero-shot and few-shot learning settings, including the generation of synthetic datasets for training and testing. However, to reliably use these synthetic datasets, it is essential to understand how representative they are of real-world data. We investigate this by assessing the effectiveness of generating synthetic data through LLM and using it as a benchmark for various NLP tasks. Our experiments across six datasets, and three different tasks, show that while synthetic data can effectively capture performance of various methods for simpler tasks, such as intent classification, it falls short for more complex tasks like named entity recognition. Additionally, we propose a new metric called the bias factor, which evaluates the biases introduced when the same LLM is used to both generate benchmarking data and to perform the tasks. We find that smaller LLMs exhibit biases towards their own generated data, whereas larger models do not. Overall, our findings suggest that the effectiveness of synthetic data as a benchmark varies depending on the task, and that practitioners should rely on data generated from multiple larger models whenever possible.

Coherent Diffraction Imaging (CDI) is an experimental technique to get images of isolated structures by recording the light scattered off the sample. Thanks to the extremely bright and short coherent light pulses provided by X-ray Free Electron Lasers, CDI makes it possible to study nanostructures in the gas phase and get time-resolved snapshots of their ultrafast dynamics with unprecedented resolution. In principle, the sample density can be recovered from the scattered light field through a straightforward Fourier Transform operation. However, only the amplitude of the field is recorded, while the phase is lost during the measurement process and has to be retrieved by means of suitable, well-established, phase retrieval algorithms. We present the Memetic Phase Retrieval (MPR) method, an improved approach to the phase retrieval problem, which makes use of a combination of existing phase retrieval algorithms and evolutionary algorithms to mitigate the shortcomings of conventional approaches. We benchmark the method on experimental data acquired in two experimental campaigns at SwissFEL and European XFEL. Imaging results on isolated nanostructures reveal considerable stability of the algorithm's behavior on the input parameters, as well as the capability of identifying the solution in challenging conditions. A user-friendly implementation of the MPR method is released as open-source software, aiming at being a reference tool for the FEL imaging community., Comment: 37 pages, 13 figures. Software not yet released as open source. Software documentation under construction. Authors list up to date

The $k$-nearest neighbor ($k$-NN) algorithm is one of the most popular methods for nonparametric classification. However, a relevant limitation concerns the definition of the number of neighbors $k$. This parameter exerts a direct impact on several properties of the classifier, such as the bias-variance tradeoff, smoothness of decision boundaries, robustness to noise, and class imbalance handling. In the present paper, we introduce a new adaptive $k$-nearest neighbours ($kK$-NN) algorithm that explores the local curvature at a sample to adaptively defining the neighborhood size. The rationale is that points with low curvature could have larger neighborhoods (locally, the tangent space approximates well the underlying data shape), whereas points with high curvature could have smaller neighborhoods (locally, the tangent space is a loose approximation). We estimate the local Gaussian curvature by computing an approximation to the local shape operator in terms of the local covariance matrix as well as the local Hessian matrix. Results on many real-world datasets indicate that the new $kK$-NN algorithm yields superior balanced accuracy compared to the established $k$-NN method and also another adaptive $k$-NN algorithm. This is particularly evident when the number of samples in the training data is limited, suggesting that the $kK$-NN is capable of learning more discriminant functions with less data considering many relevant cases., Comment: 18 pages, 4 figures

Automatic Essay Scoring (AES) is widely used to evaluate candidates for educational purposes. However, due to the lack of representative data, most existing AES systems are not robust, and their scoring predictions are biased towards the most represented data samples. In this study, we propose a model-agnostic phrase-level method to generate an adversarial essay set to address the biases and robustness of AES models. Specifically, we construct an attack test set comprising samples from the original test set and adversarially generated samples using our proposed method. To evaluate the effectiveness of the attack strategy and data augmentation, we conducted a comprehensive analysis utilizing various neural network scoring models. Experimental results show that the proposed approach significantly improves AES model performance in the presence of adversarial examples and scenarios without such attacks.

We study Dirac fermions at finite density coupled to a complex pairing field assumed to obey scalar field theory with quartic self-repulsion. The bulk of our work develops the mathematics that elucidates the propagation of fermionic excitations in such systems as independent spin (boosts) and charge (fermion number) degrees of freedom. A necessary ingredient is the presence of broken $U(1)$ symmetry in the pairing field and decoupling of its density and phase. In the fermion sector, these elements give rise to an emergent spin-dependent gauge coupling which binds in-vacuum spin and charge into elementary fermions, while driving proliferation of unbound spin and charge for finite condensation in the pairing field. Notably, the onset of spin-charge separation is signaled by $\mathcal{P} \mathcal{T}$-symmetry breaking and decoupling of spin components under Lorentz transformations. Our investigation concludes with two theorems that identify generic features of spin-charge separation in such systems., Comment: 14 pages, no figures

In this work we propose a universal basic mass density and a universal basic metric inside dark matter halos in the framework of Einstein equations providing an analytical ground for learning about dark matter. In a previous work the authors have proposed a simplified model for galaxies: a Schwarzschild black hole (that contains all the baryonic matter of the galaxy) immersed inside a dark matter halo. The solution solved the Einstein equations perturbatively and successfully gave the flat rotation curve and the baryonic Tully-Fisher relation, the two signatures of spiral galaxies. In this work we take the black hole mass (the baryonic mass) of the solution to be zero in order to focus our study on the dark matter halo exclusively. Among the results (1) is the prediction of a universal basic mass density of dark matter, $\rho=a_0/2\pi G r$, where $a_0\approx 2.8 \times 10^{-11} \ m/s^2$ is a universal constant whose value was deduced from observations, (2) we show that the mass and velocity curve of the dark halo agree excellently with observational data at intermediate distances where the baryonic matter contribution is negligible, (3) we show that the constant $a_0$ is the origin of the constant surface density of dark matter and the origin of the scale-radius/scale-density scaling relation of the Navarro-Frenk-White profile (4) we show how the inclusion of baryonic matter in our model solves the core-cusp problem., Comment: 15 pages, 4 figures

Nonlinear wave mixing in the X-ray range can provide valuable insights into the structural and electron dynamics of atomic and molecular systems on ultrafast time scales, with state- and site-selectivity and atomic resolution. This promising experimental toolbox was so far limited by requiring at least one near-visible laser, thus preventing core-shell two-dimensional X-ray spectroscopy. In this work, we demonstrate the generation of background-free all-X-ray four-wave mixing (XFWM) signals from a dilute gaseous sample (Ne). The measured and simulated two-dimensional spectral maps ($\omega_{\text{in}},\omega_{\text{out}}$) show multiple contributions involving the coherent response from core electrons. Notably, two-color resonant XFWM signals, essential for generalized multi-color schemes that allow to locally probe the electronic excitation of matter, are observed in neutral Ne. Moreover, stimulated Ne$^+$ emission in each of the propagating X-ray pulses leads to an increase of the temporal coherence in a narrow-bandwidth, which results in the coherent mixing of three X-ray lasers. Preliminary X-ray excitation experiments making use of multi-color time-delayed X-ray pulses demonstrate temporal resolution capability and show a time dependency consistent with a signal dominated by resonant XFWM processes. This first all-X-ray four-wave-mixing approach represents a major breakthrough towards multidimensional X-ray correlation spectroscopy and the general application of nonlinear all-X-ray wave-mixing.

Given an integer $m \geq 2$ and a sufficiently large $q$, we apply a variant of the Maynard--Tao sieve weight to establish the existence of an arithmetic progression with common difference $q$ for which the $m$-th least prime in such progression is $\ll_m q$, which is best possible. As we vary over progressions instead of fixing a particular one, the nature of our result differs from others in the literature. Furthermore, we generalize our result to dynamical systems. The quality of the result depends crucially on the first return time, which we illustrate in the case of Diophantine approximation., Comment: 15 pages

The need to immobilize low-level nuclear waste, in particular 137Cs-bearing waste, has led to a growing interest in geopolymer-based waste matrices, in addition to optimization attempts of cement matrix compositions for this specific application. Although the overall phase composition and structure of these matrices are well characterized, the binding sites of Cs in these materials have not been clearly identified. Recent studies have suggested that combining the sensitivity of solid-state Nuclear Magnetic Resonance (SSNMR) to the local atomic structure with other structural techniques provides insights into the mode of Cs binding and release. Density Functional Theory (DFT) can provide the connection between spectroscopic parameters and geometric properties. However, the reliability of DFT results strongly relies on the choice of a suitable exchange-correlation functional, which for 133Cs, the NMR surrogate for such studies, is not well-established. In this work we benchmark various functionals against their performance in predicting the geometry of various simple Cs compounds, their NMR quadrupolar coupling constants, and their chemical shift values. The results show rev-vdW-DF2 to be a leading, though not a flawless, candidate for these systems, and apparently for additional classes of Cs-containing systems., Comment: 19 pages, 9 figures, supporting information with 15 pages, 5 figures

Coronal mass ejections (CMEs) are magnetized plasma systems with highly complex magnetic topology and evolution. Methods developed to assess their magnetic configuration have primarily focused on reconstructing three-dimensional representations from one-dimensional time series measurements taken in situ using techniques based on the "highly twisted magnetic flux rope" approximations. However, the magnetic fields of CMEs is know to have more complicated geometries. Their structure can be quantified using measures of field line topology, which have been primarily used for solar physics research. In this work, we introduce a novel technique of directly quantifying the various form of magnetic helicity within a CME in the interplanetary space using synthetic in situ measurements. We use a relatively simple three-dimensional simulation of a CME initiated with a highly-twisted flux rope. We find that a significant portion of the magnetic helicity near 1~au is contained in writhe and mutual helicity rather than just in twist. We discuss the implications of this finding for fitting and reconstruction techniques., Comment: submitted to ApJ

This paper addresses a new class of generalized Bolza problems governed by nonconvex integro-differential inclusions with endpoint constraints on trajectories, where the integral terms are given in the general (with time-dependent integrands in the dynamics) Volterra form. We pursue here a threefold goal. First we construct well-posed approximations of continuous-time integro-differential systems by their discrete-time counterparts with showing that any feasible solution to the original system can be strongly approximated in the $W^{1,2}$-norm topology by piecewise-linear extensions of feasible discrete trajectories. This allows us to verify in turn the strong convergence of discrete optimal solutions to a prescribed local minimizer for the original problem. Facing intrinsic nonsmoothness of original integro-differential problem and its discrete approximations, we employ appropriate tools of generalized differentiation in variational analysis to derive necessary optimality conditions for discrete-time problems (which is our second goal) and finally accomplish our third goal to obtain necessary conditions for the original continuous-time problems by passing to the limit from discrete approximations. In this way we establish, in particular, a novel necessary optimality condition of the Volterra type, which is the crucial result for dynamic optimization of integro-differential inclusions., Comment: 28 pages

Integro-differential sweeping processes with prox-regular sets in Hilbert spaces have been the subject of various recent studies. Diverse applications of such differential inclusions to complementarity problems, electrical circuits, frictionless contact, can be found in the literature. Here we provide a general theorem of existence of solution for such processes perturbed by a Lipschitz multimapping with nonconvex values., Comment: 19 pages

Consider the model where we can access a parity function through random uniform labeled examples in the presence of random classification noise. In this paper, we show that approximating the number of relevant variables in the parity function is as hard as properly learning parities. More specifically, let $\gamma:{\mathbb R}^+\to {\mathbb R}^+$, where $\gamma(x) \ge x$, be any strictly increasing function. In our first result, we show that from any polynomial-time algorithm that returns a $\gamma$-approximation, $D$ (i.e., $\gamma^{-1}(d(f)) \leq D \leq \gamma(d(f))$), of the number of relevant variables~$d(f)$ for any parity $f$, we can, in polynomial time, construct a solution to the long-standing open problem of polynomial-time learning $k(n)$-sparse parities (parities with $k(n)\le n$ relevant variables), where $k(n) = \omega_n(1)$. In our second result, we show that from any $T(n)$-time algorithm that, for any parity $f$, returns a $\gamma$-approximation of the number of relevant variables $d(f)$ of $f$, we can, in polynomial time, construct a $poly(\Gamma(n))T(\Gamma(n)^2)$-time algorithm that properly learns parities, where $\Gamma(x)=\gamma(\gamma(x))$. If $T(\Gamma(n)^2)=\exp({o(n/\log n)})$, this would resolve another long-standing open problem of properly learning parities in the presence of random classification noise in time $\exp({o(n/\log n)})$.

Investigating the dynamics of chromatin and the factors that are affecting it, has provided valuable insights into the organization and functionality of the genome in the cell nucleus. We control the expression of Lamin-A, an important organizer protein of the chromatin and nucleus structure. By simultaneously tracking tens of chromosomal loci (telomeres) in each nucleus, we find that the motion of chromosomal loci in Lamin-A depleted cells is both faster and more directed on a scale of a few micrometers, which coincides with the size of chromosome territories. Moreover, in the absence of Lamin-A we reveal the existence of correlations among neighboring telomeres. We show how these pairwise correlations are linked with the intermittent and persistent character of telomere trajectories, underscoring the importance of Lamin-A protein in chromosomal organization.

Let $x \ge 2$, let $N_x$ be an integer chosen uniformly at random from the set $\mathbb Z \cap [1, x]$, and let $(V_1, V_2, \ldots)$ be a Poisson--Dirichlet process of parameter $1$. We prove that there exists a coupling of these two random objects such that $$ \mathbb E \, \sum_{i \ge 1} |\log P_i- V_i\log x| \ll 1, $$ where the implied constant is absolute and $N_x = P_1P_2 \cdots$ is the unique factorization of $N_x$ into primes or ones with the $P_i$'s being non-increasing. This establishes a conjecture of Arratia (2002) arXiv:1305.0941 who proved that the left-hand side in the above estimate can be made $\ll \log\!\log x$. We also use this coupling to give a probabilistic proof of the Dirichlet law for the average distribution of the integer factorization into $k$ parts proved in 2023 by Leung arXiv:2206.14728 and we improve on its error term., Comment: 36 pages; minor corrections

Progress in AI is often demonstrated by new models claiming improved performance on tasks measuring model capabilities. Evaluating language models in particular is challenging, as small changes to how a model is evaluated on a task can lead to large changes in measured performance. There is no common standard setup, so different models are evaluated on the same tasks in different ways, leading to claims about which models perform best not being reproducible. We propose OLMES, a completely documented, practical, open standard for reproducible LLM evaluations. In developing this standard, we identify and review the varying factors in evaluation practices adopted by the community - such as details of prompt formatting, choice of in-context examples, probability normalizations, and task formulation. In particular, OLMES supports meaningful comparisons between smaller base models that require the unnatural "cloze" formulation of multiple-choice questions against larger models that can utilize the original formulation. OLMES includes well-considered recommendations guided by results from existing literature as well as new experiments investigating open questions.

This report presents a short summary of the activities of the "Standard Model" working group for the "Physics at TeV Colliders" workshop (Les Houches, France, 12-30 June, 2023)., Comment: Proceedings of the Standard Model Working Group of the 2023 Les Houches Workshop, Physics at TeV Colliders, Les Houches 12-30 June 2023. 48 pages

The shift from a linear to a circular economy has the potential to simultaneously reduce uncertainties of material supplies and waste generation. However, to date, the development of robotic and, more generally, autonomous systems have been rarely integrated into circular economy implementation strategies despite their potential to reduce the operational costs and the contamination risks from handling waste. In addition, the science of circularity still lacks the physical foundations needed to improve the accuracy and the repeatability of the models. Hence, in this paper, we merge deep-learning vision, compartmental dynamical thermodynamics, and robotic manipulation into a theoretically-coherent physics-based research framework to lay the foundations of circular flow designs of materials. The proposed framework tackles circularity by generalizing the design approach of the Rankine cycle enhanced with dynamical systems theory. This differs from state-of-the-art approaches to circular economy, which are mainly based on data analysis, e.g., material flow analysis (MFA). We begin by reviewing the literature of the three abovementioned research areas, then we introduce the proposed unified framework and we report the initial application of the framework to plastics systems along with initial simulation results of reinforcement-learning control of robotic waste sorting. This shows the framework applicability, generality, scalability, and the similarity and difference between the optimization of artificial neural systems and the proposed compartmental networks. Finally, we discuss the still not fully exploited opportunities for robotics in circular economy and the future challenges in the theory and practice of the proposed circularity framework., Comment: To be submitted

It was shown in [11] that for every origin-symmetric star body $K \subseteq \mathbb R^n$ of volume $1$, every even continuous probability density $f$ on $K$ and $1 \leq k \leq n-1$, there exists a subspace $F \subseteq \mathbb R^n$ of codimension $k$ such that \[ \int_{K \cap F} f \geq c^k (d_{\rm ovr}(K, \mathcal{BP}_k^n))^{-k} \] where $d_{\rm ovr}(K, \mathcal{BP}_k^n)$ is the outer volume ratio distance from $K$ to the class of generalized $k$-intersection bodies, and $c>0$ is a universal constant. The upper bound $d_{\rm ovr}(K, \mathcal{BP}_k^n) \leq c' \sqrt{n/k} \left(\log\left(\frac{en}k\right)\right)^{3/2}$ was established in [13] for every origin-symmetric convex body $K$. In this note we show that there exist an origin-symmetric convex body $K$ of volume $1$ and an even continuous probability density $f$ supported on $K$ such that for every subspace $F$ of codimension $k$, \[ \int_{K \cap F} f \leq \left( c \sqrt{\frac n{k \log(n)} } \right)^{-k}. \] As a consequence we obtain a lower bound for $d_{\rm ovr}(K, \mathcal{BP}_k^n)$ with $K$ a convex body, complementing the upper bound in \cite{koldobsky2011isomorphic}. This is \[c \sqrt{n/k} (\log(n))^{-1/2} \leq \sup_K d_{\rm ovr}(K, \mathcal{BP}_k^n) \leq c' \sqrt{n/k} \left(\log\left(\frac{en}k\right)\right)^{3/2}.\] The case $k=1$ was obtained previously in [5,6].

For every convex body $K \subset \mathbb R^n$ and $\delta \in (0,1)$, the $\delta$-convolution body of $K$ is the set of $x \in \mathbb R^n$ for which $\left|K \cap (K+x)\right|_n \geq \delta \left|K\right|_n$. We show that for $n=2$ and any $\delta \in (0,1)$, ellipsoids do not maximize the volume of the $\delta$-convolution body of $K$, when $K$ runs over all convex bodies of a fixed volume. This behavior is somehow unexpected and contradicts the limit case $\delta \to 1^-$, which is governed by the Petty projection inequality.

Can we modify the training data distribution to encourage the underlying optimization method toward finding solutions with superior generalization performance on in-distribution data? In this work, we approach this question for the first time by comparing the inductive bias of gradient descent (GD) with that of sharpness-aware minimization (SAM). By studying a two-layer CNN, we rigorously prove that SAM learns different features more uniformly, particularly in early epochs. That is, SAM is less susceptible to simplicity bias compared to GD. We also show that examples containing features that are learned early are separable from the rest based on the model's output. Based on this observation, we propose a method that (i) clusters examples based on the network output early in training, (ii) identifies a cluster of examples with similar network output, and (iii) upsamples the rest of examples only once to alleviate the simplicity bias. We show empirically that USEFUL effectively improves the generalization performance on the original data distribution when training with various gradient methods, including (S)GD and SAM. Notably, we demonstrate that our method can be combined with SAM variants and existing data augmentation strategies to achieve, to the best of our knowledge, state-of-the-art performance for training ResNet18 on CIFAR10, STL10, CINIC10, Tiny-ImageNet; ResNet34 on CIFAR100; and VGG19 and DenseNet121 on CIFAR10., Comment: 43 pages, 15 figures, 9 tables

In common law jurisdictions, legal practitioners rely on precedents to construct arguments, in line with the doctrine of \emph{stare decisis}. As the number of cases grow over the years, prior case retrieval (PCR) has garnered significant attention. Besides lacking real-world scale, existing PCR datasets do not simulate a realistic setting, because their queries use complete case documents while only masking references to prior cases. The query is thereby exposed to legal reasoning not yet available when constructing an argument for an undecided case as well as spurious patterns left behind by citation masks, potentially short-circuiting a comprehensive understanding of case facts and legal principles. To address these limitations, we introduce a PCR dataset based on judgements from the European Court of Human Rights (ECtHR), which explicitly separate facts from arguments and exhibit precedential practices, aiding us to develop this PCR dataset to foster systems' comprehensive understanding. We benchmark different lexical and dense retrieval approaches with various negative sampling strategies, adapting them to deal with long text sequences using hierarchical variants. We found that difficulty-based negative sampling strategies were not effective for the PCR task, highlighting the need for investigation into domain-specific difficulty criteria. Furthermore, we observe performance of the dense models degrade with time and calls for further research into temporal adaptation of retrieval models. Additionally, we assess the influence of different views , Halsbury's and Goodhart's, in practice in ECtHR jurisdiction using PCR task., Comment: Accepted to LREC-COLING 2024

We studied the anelastic aftereffect of a flexure being used in a Kibble balance, where the flexure is subjected to a large excursion in velocity mode after which a high-precision force comparison is performed. We investigated the effect of a constant and a sinusoidal excursion on the force comparison. We explored theoretically and experimentally a simple erasing procedure, i.e., bending the flexure in the opposite direction for a given amplitude and time. We found that the erasing procedure reduced the time-dependent force by about 30%. The investigation was performed with an analytical model and verified experimentally with our new Kibble balance at the National Institute of Standards and Technology employing flexures made from precipitation-hardened Copper Beryllium alloy C17200. Our experimental determination of the modulus defect of the flexure yields 1.2E-4. This result is about a factor of two higher than previously reported from experiments. We additionally found a static shift of the flexure's internal equilibrium after a change in the stress and strain state. These static shifts, although measurable, are small and deemed uncritical for our Kibble balance application at present. During this investigation, we discovered magic flexures that promise to have very little anelastic relaxation. In these magic flexures, the mechanism causing anelastic relaxation is compensated for by properly shaping and loading a flexure with a non-constant cross-section in the region of bending., Comment: 18 pages 22 figure, accepted byMetrologia

Generative artificial intelligence (AI) offers numerous opportunities for research and innovation, but its commercialization has raised concerns about the transparency and safety of frontier AI models. Most models lack the necessary components for full understanding, auditing, and reproducibility, and some model producers use restrictive licenses whilst claiming that their models are "open source". To address these concerns, we introduce the Model Openness Framework (MOF), a three-tiered ranked classification system that rates machine learning models based on their completeness and openness, following open science principles. For each MOF class, we specify code, data, and documentation components of the model development lifecycle that must be released and under which open licenses. In addition, the Model Openness Tool (MOT) provides a user-friendly reference implementation to evaluate the openness and completeness of models against the MOF classification system. Together, the MOF and MOT provide timely practical guidance for (i) model producers to enhance the openness and completeness of their publicly-released models, and (ii) model consumers to identify open models and their constituent components that can be permissively used, studied, modified, and redistributed. Through the MOF, we seek to establish completeness and openness as core tenets of responsible AI research and development, and to promote best practices in the burgeoning open AI ecosystem., Comment: 28 pages, 4 figures, 2 tables

Robot art presents an opportunity to both showcase and advance state-of-the-art robotics through the challenging task of creating art. Creating large-scale artworks in particular engages the public in a way that small-scale works cannot, and the distinct qualities of brush strokes contribute to an organic and human-like quality. Combining the large scale of murals with the strokes of the brush medium presents an especially impactful result, but also introduces unique challenges in maintaining precise, dextrous motion control of the brush across such a large workspace. In this work, we present the first robot to our knowledge that can paint architectural-scale murals with a brush. We create a hybrid robot consisting of a cable-driven parallel robot and 4 degree of freedom (DoF) serial manipulator to paint a 27m by 3.7m mural on windows spanning 2-stories of a building. We discuss our approach to achieving both the scale and accuracy required for brush-painting a mural through a combination of novel mechanical design elements, coordinated planning and control, and on-site calibration algorithms with experimental validations., Comment: 8 pages IEEE conference format, submitted to IROS 2024

Detection of moving sources over complicated background is important for several reasons. First is measuring the astrophysical motion of the source. Second is that such motion resulting from atmospheric scintillation, color refraction, or astrophysical reasons is a major source of false alarms for image subtraction methods. We extend the Zackay, Ofek, and Gal-Yam image subtraction formalism to deal with moving sources. The new method, named translient (translational transient) detector, applies hypothesis testing between the hypothesis that the source is stationary and that the source is moving. It can be used to detect source motion or to distinguish between stellar variability and motion. For moving source detection, we show the superiority of translient over the proper image subtraction, using the improvement in the receiver-operating characteristic curve. We show that in the small translation limit, Translient is an optimal detector of point source motion in any direction. Furthermore, it is numerically stable, fast to calculate, and presented in a closed form. Efficient transient detection requires both the proper image subtraction statistics and the translient statistics: when the translient statistic is higher, then the subtraction artifact is likely due to motion. We test our algorithm both on simulated data and on real images obtained by the Large Array Survey Telescope (LAST). We demonstrate the ability of translient to distinguish between motion and variability, which has the potential to reduce the number of false alarms in transients detection. We provide the translient implementation in Python and MATLAB., Comment: 13 pages, 11 figures, submitted to AJ

Smiles and laughs detection systems have attracted a lot of attention in the past decade contributing to the improvement of human-agent interaction systems. But very few considered these expressions as distinct, although no prior work clearly proves them to belong to the same category or not. In this work, we present a deep learning-based multimodal smile and laugh classification system, considering them as two different entities. We compare the use of audio and vision-based models as well as a fusion approach. We show that, as expected, the fusion leads to a better generalization on unseen data. We also present an in-depth analysis of the behavior of these models on the smiles and laughs intensity levels. The analyses on the intensity levels show that the relationship between smiles and laughs might not be as simple as a binary one or even grouping them in a single category, and so, a more complex approach should be taken when dealing with them. We also tackle the problem of limited resources by showing that transfer learning allows the models to improve the detection of confusing intensity levels.