Search

Multivariate Time Series Classification (MTSC) is a ubiquitous problem in science and engineering, particularly in neuroscience, where most data acquisition modalities involve the simultaneous time-dependent recording of brain activity in multiple brain regions. In recent years, Random Convolutional Kernel models such as ROCKET and MiniRocket have emerged as highly effective time series classification algorithms, capable of achieving state-of-the-art accuracy results with low computational load. Despite their success, these types of models face two major challenges when employed in neuroscience: 1) they struggle to deal with high-dimensional data such as EEG and MEG, and 2) they are difficult to interpret. In this work, we present a novel ROCKET-based algorithm, named Detach-Rocket Ensemble, that is specifically designed to address these two problems in MTSC. Our algorithm leverages pruning to provide an integrated estimation of channel importance, and ensembles to achieve better accuracy and provide a label probability. Using a synthetic multivariate time series classification dataset in which we control the amount of information carried by each of the channels, we first show that our algorithm is able to correctly recover the channel importance for classification. Then, using two real-world datasets, a MEG dataset and an EEG dataset, we show that Detach-Rocket Ensemble is able to provide both interpretable channel relevance and competitive classification accuracy, even when applied directly to the raw brain data, without the need for feature engineering., Comment: To be published in European Conference on Machine Learning and Data Mining 2024, 20 pages, 7 figures, 2 tables

In the context of the CERN EP R&D on monolithic sensors and the ALICE ITS3 upgrade, the Tower Partners Semiconductor Co (TPSCo) 65 nm process has been qualified for use in high energy physics, and adopted for the ALICE ITS3 upgrade. An Analog Pixel Test Structure (APTS) featuring fast per pixel operational-amplifier-based buffering for a small matrix of four by four pixels, with a sensor with a small collection electrode and a very non-uniform electric field, was designed to allow detailed characterization of the pixel performance in this technology. Several variants of this chip with different pixel designs have been characterized with a (120 GeV/$c$) positive hadron beam. This result indicates that the APTS-OA prototype variants with the best performance achieve a time resolution of 63 ps with a detection efficiency exceeding 99% and a spatial resolution of 2 $\mu$m, highlighting the potential of TPSCo 65nm CMOS imaging technology for high-energy physics and other fields requiring precise time measurement, high detection efficiency, and excellent spatial resolution.

We embed the Penrose limit into the Weyl classical double copy. Thereby, we provide a lift of the double copy properties of plane wave spacetimes into black hole geometries and we open a novel avenue towards taking the classical double copy beyond statements about algebraically special backgrounds. In particular, the Penrose limit, viewed as the leading order Fermi coordinate expansion around a null geodesic, complements approaches leveraging asymptotic flatness such as the asymptotic Weyl double copy. Along the way, we show how our embedding of the Penrose limit within the Weyl double copy naturally fixes the functional ambiguity in the double copy for Petrov type N spacetimes. We also highlight the utility of a spinorial approach to the Penrose limit. In particular, we use this spinorial approach to derive a simple analytical expression for arbitrary Penrose limits of four-dimensional, vacuum type D spacetimes., Comment: 44+18 pages

Photoionized and electronically excited ethylene \ce{C2H4+} can undergo \ce{H}-loss, \ce{H2}-loss, and ethylene-ethylidene isomerization, where the latter entails a hydrogen migration. Recent pioneering experiments with few-femtosecond extreme ultraviolet pulses and complementary theoretical studies have shed light on the photodynamics of this prototypical organic cation. However, no theoretical investigation based on dynamics simulations reported to date has described the mechanisms and time scales of dissociation and isomerization. Herein, we simulate the coupled electron-nuclear dynamics of ethylene following vertical ionization and electronic excitation to its four lowest-lying cationic states. The electronic structure is treated at the CASSCF level, with an active space large enough to describe bond breaking and formation. The simulations indicate that dissociation and isomerization take place mainly on the cationic ground state and allow the probing of previous hypotheses concerning the correlation between the photochemical outcome and the traversed conical intersections. The results, moreover, support the long-standing view that \ce{H2}-loss may occur from the ethylidene form. However, the ethylene-ethylidene isomerization time predicted by the simulations is considerably longer than those previously inferred from indirect experimental measurements.

We initiate a general investigation into gravitational wave signatures of modifications to scattering of gravitational radiation from black holes. Such modifications may be present due to the quantum dynamics that makes black holes consistent with quantum mechanics, or in other models for departures from classical black hole behavior. We propose a parameterization of the corrections to scattering as a physically meaningful, model-independent, and practical bridge between theoretical and observational aspects of the problem; this parameterization can incorporate different models in the literature. We then describe how these corrections influence the gravitational wave signal, e.g. of a body orbiting a much more massive black hole. In particular, they generically change the rate of energy emission; this effect can be leveraged over many orbits of inspiral to enhance the sensitivity to small corrections, as has been noticed in simple models. We provide preliminary estimates of the sensitivity of future gravitational wave observations to these corrections, and outline further work to be done to connect both to a more fundamental theory of quantum black holes, and to realistic observational situations., Comment: 28 + 5 pages, 5 figures

The present paper introduces new sentiment data, MaCMS, for Magahi-Hindi-English (MHE) code-mixed language, where Magahi is a less-resourced minority language. This dataset is the first Magahi-Hindi-English code-mixed dataset for sentiment analysis tasks. Further, we also provide a linguistics analysis of the dataset to understand the structure of code-mixing and a statistical study to understand the language preferences of speakers with different polarities. With these analyses, we also train baseline models to evaluate the dataset's quality., Comment: Lrec-Colin 2024

Exploiting cognates for transfer learning in under-resourced languages is an exciting opportunity for language understanding tasks, including unsupervised machine translation, named entity recognition and information retrieval. Previous approaches mainly focused on supervised cognate detection tasks based on orthographic, phonetic or state-of-the-art contextual language models, which under-perform for most under-resourced languages. This paper proposes a novel language-agnostic weakly-supervised deep cognate detection framework for under-resourced languages using morphological knowledge from closely related languages. We train an encoder to gain morphological knowledge of a language and transfer the knowledge to perform unsupervised and weakly-supervised cognate detection tasks with and without the pivot language for the closely-related languages. While unsupervised, it overcomes the need for hand-crafted annotation of cognates. We performed experiments on different published cognate detection datasets across language families and observed not only significant improvement over the state-of-the-art but also our method outperformed the state-of-the-art supervised and unsupervised methods. Our model can be extended to a wide range of languages from any language family as it overcomes the requirement of the annotation of the cognate pairs for training. The code and dataset building scripts can be found at https://github.com/koustavagoswami/Weakly_supervised-Cognate_Detection, Comment: Accepted at EMNLP 2023 conference

Time Series Classification (TSC) is essential in fields like medicine, environmental science, and finance, enabling tasks such as disease diagnosis, anomaly detection, and stock price analysis. While machine learning models like Recurrent Neural Networks and InceptionTime are successful in numerous applications, they can face scalability issues due to computational requirements. Recently, ROCKET has emerged as an efficient alternative, achieving state-of-the-art performance and simplifying training by utilizing a large number of randomly generated features from the time series data. However, many of these features are redundant or non-informative, increasing computational load and compromising generalization. Here we introduce Sequential Feature Detachment (SFD) to identify and prune non-essential features in ROCKET-based models, such as ROCKET, MiniRocket, and MultiRocket. SFD estimates feature importance using model coefficients and can handle large feature sets without complex hyperparameter tuning. Testing on the UCR archive shows that SFD can produce models with better test accuracy using only 10\% of the original features. We named these pruned models Detach-ROCKET. We also present an end-to-end procedure for determining an optimal balance between the number of features and model accuracy. On the largest binary UCR dataset, Detach-ROCKET improves test accuracy by 0.6\% while reducing features by 98.9\%. By enabling a significant reduction in model size without sacrificing accuracy, our methodology improves computational efficiency and contributes to model interpretability. We believe that Detach-ROCKET will be a valuable tool for researchers and practitioners working with time series data, who can find a user-friendly implementation of the model at \url{https://github.com/gon-uri/detach_rocket}., Comment: 18 pages, 5 figures, 3 tables

We compute the spectrum of linearized gravitational excitations of black holes with substantial angular momentum in the presence of higher-derivative corrections to general relativity. We do so perturbatively to leading order in the higher-derivative couplings and up to order fourteen in the black hole angular momentum. This allows us to accurately predict quasinormal mode frequencies of black holes with spins up to about $70\%$ of the extremal value. For some higher-derivative corrections, we find that sizable rotation enhances the frequency shifts by almost an order of magnitude relative to the static case., Comment: v2: version accepted in PRD. Contains a new plot for the (330) modes and we added a new comment on the accuracy of our results. We make it clear that they are accurate enough to allow for precision tests of higher-derivative corrections in future experiments

We reduce the study of perturbations of rotating black holes in higher-derivative extensions of general relativity to a system of decoupled radial equations that stem from a set of universal Teukolsky equations. We detail a complete computational strategy to obtain these decoupled equations in general higher-derivative theories. We apply this to six-derivative gravity to compute the shifts in the quasinormal mode frequencies with respect to those of Kerr black holes in general relativity. At linear order in the angular momentum we reproduce earlier results obtained with a metric perturbation approach. In contrast with this earlier work, however, the method given here applies also to post-merger black holes with significant spin, which are of particular observational interest., Comment: 50 pages, 5 figures. v3: version published in PRD

We use Penrose limits to approximate quasinormal modes with large real frequencies. The Penrose limit associates a plane wave to a region of spacetime near a null geodesic. This plane wave can be argued to geometrically realize the geometrical optics approximation. Therefore, when applied to the bound null orbits around black holes, the Penrose limit can be used to study quasinormal modes. For instance, this Penrose limit point of view makes manifest the symmetry that emerges in the geometrical optics approximation of quasinormal modes in terms of isometries of the resulting plane wave spacetime. We apply the procedure to warped $\rm{AdS}_3$, Schwarzschild black holes and Kerr black holes. In the former we show explicitly how the symmetry algebra contracts to that of the limiting plane wave while in the latter we find the expected agreement with numerically computed quasinormal modes for large (real) frequencies., Comment: 27 pages, 9 figures

[Background] Photon strength functions are an important ingredient in calculations relevant for the nucleosynthesis of heavy elements. The relation to the photoabsorption cross section allows to experimentally constrain photon strength functions by investigating the photo-response of atomic nuclei. [Purpose] We determine the photoresponse of $^{66}$Zn in the energy region of 5.6 MeV to 9.9 MeV and analyze the contribution of the "elastic" decay channel back to the ground state. In addition, for the elastic channel electric and magnetic dipole transitions were separated. [Methods] Nuclear resonance fluorescence experiments were performed using a linearly-polarized quasi-monoenergetic photon beam at the High Intensity $\gamma$-ray Source. Photon beam energies from 5.6 to 9.9 MeV with an energy spread of about 3% were selected in steps of 200-300 keV. Two High Purity Germanium detectors were used for the subsequent $\gamma$-ray spectroscopy. [Results] Full photoabsorption cross sections are extracted from the data making use of the monoenergetic character of the photon beam. For the ground-state decay channel, the average contribution of electric and magnetic dipole strengths is disentangled. The average branching ratio back to the ground state is determined as well. [Conclusions] The new results indicate lower cross sections when compared to the values extracted from a former experiment using bremsstrahlung on $^{66}$Zn. In the latter, the average branching ratio to the ground state is estimated from statistical-model calculations in order to analyze the data. Corresponding estimates from statistical-model calculations underestimate this branching ratio compared to the values extracted from the present analysis, which would partly explain the high cross sections determined from the bremsstrahlung data.

The Laser Interferometer Space Antenna (LISA) has the potential to reveal wonders about the fundamental theory of nature at play in the extreme gravity regime, where the gravitational interaction is both strong and dynamical. In this white paper, the Fundamental Physics Working Group of the LISA Consortium summarizes the current topics in fundamental physics where LISA observations of GWs can be expected to provide key input. We provide the briefest of reviews to then delineate avenues for future research directions and to discuss connections between this working group, other working groups and the consortium work package teams. These connections must be developed for LISA to live up to its science potential in these areas., Comment: Accepted in: Living Reviews in Relativity

A characterisation of the Timepix4 pixel front-end with a strong focus on timing performance is presented. Externally generated test pulses were used to probe the per-pixel time-to-digital converter (TDC) and measure the time-bin sizes by precisely controlling the test-pulse arrival time in steps of 10 ps. The results indicate that the TDC can achieve a time resolution of 60 ps, provided that a calibration is performed to compensate for frequency variation in the voltage controlled oscillators of the pixel TDCs. The internal clock distribution system of Timepix4 was used to control the arrival time of internally generated analog test pulses in steps of about 20 ps. The analog test pulse mechanism injects a controlled amount of charge directly into the analog front-end (AFE) of the pixel, and was used to measure the time resolution as a function of signal charge, independently of the TDC. It was shown that for the default configuration, the AFE time resolution in the hole-collecting mode is limited to 105 ps. However, this can be improved up to about 60 ps by increasing the preamplifier bias-current at the cost of increased power dissipation. For the electron-collecting mode, an AFE time resolution of 47 ps was measured for a bare Timepix4 device at a signal charge of 21 ke. It was observed that additional input capacitance from a bonded sensor reduces this figure to 62 ps.

The equatorial symmetry of the Kerr black hole is generically broken in models of quantum gravity. Nevertheless, most phenomenological models start from the assumption of equatorial symmetry, and little attention has been given to the observability of this smoking gun signature of beyond-GR physics. Extreme mass-ratio inspirals (EMRIs), in particular, are known to sensitively probe supermassive black holes near their horizon; yet estimates for constraints on deviations from Kerr in space-based gravitational wave observations (e.g. with LISA) of such systems are currently based on equatorially symmetric models. We use modified "analytic kludge" waveforms to estimate how accurately LISA will be able to measure or constrain equatorial symmetry breaking, in the form of the lowest-lying odd parity multipole moments $S_2, M_3$. We find that the dimensionless multipole ratios such as $S_2/M^3$ will typically be detectable for LISA EMRIs with a measurement accuracy of $\Delta(S_2/M^3) \sim 1\%$; this will set a strong constraint on the breaking of equatorial symmetry., Comment: 16 pages + appendices; 2 figures and 4 tables

We study gravitational perturbations of slowly-rotating black holes in a general effective-field-theory extension of general relativity that includes up to eight-derivative terms. We show that two Schr\"odinger-like equations with spin-dependent effective potentials govern the odd - and even-parity master variables. These equations are coupled for parity-violating corrections, and this coupling affects the quasinormal modes even at linear order in the higher-derivative corrections, due to their isospectrality in general relativity. We provide results for the shifts in the fundamental quasinormal mode frequencies at linear order in the spin, which we expect to be valuable for high-precision phenomenology through future gravitational wave observations., Comment: 10 pages + appendices, 2 figures, double column. v2: references added, version sent to the journal

Although LEGO sets have entertained generations of children and adults, the challenge of designing customized builds matching the complexity of real-world or imagined scenes remains too great for the average enthusiast. In order to make this feat possible, we implement a system that generates a LEGO brick model from 2D images. We design a novel solution to this problem that uses an octree-structured autoencoder trained on 3D voxelized models to obtain a feasible latent representation for model reconstruction, and a separate network trained to predict this latent representation from 2D images. LEGO models are obtained by algorithmic conversion of the 3D voxelized model to bricks. We demonstrate first-of-its-kind conversion of photographs to 3D LEGO models. An octree architecture enables the flexibility to produce multiple resolutions to best fit a user's creative vision or design needs. In order to demonstrate the broad applicability of our system, we generate step-by-step building instructions and animations for LEGO models of objects and human faces. Finally, we test these automatically generated LEGO sets by constructing physical builds using real LEGO bricks., Comment: 9 pages, 10 figures

We present the findings of the LoResMT 2021 shared task which focuses on machine translation (MT) of COVID-19 data for both low-resource spoken and sign languages. The organization of this task was conducted as part of the fourth workshop on technologies for machine translation of low resource languages (LoResMT). Parallel corpora is presented and publicly available which includes the following directions: English$\leftrightarrow$Irish, English$\leftrightarrow$Marathi, and Taiwanese Sign language$\leftrightarrow$Traditional Chinese. Training data consists of 8112, 20933 and 128608 segments, respectively. There are additional monolingual data sets for Marathi and English that consist of 21901 segments. The results presented here are based on entries from a total of eight teams. Three teams submitted systems for English$\leftrightarrow$Irish while five teams submitted systems for English$\leftrightarrow$Marathi. Unfortunately, there were no systems submissions for the Taiwanese Sign language$\leftrightarrow$Traditional Chinese task. Maximum system performance was computed using BLEU and follow as 36.0 for English--Irish, 34.6 for Irish--English, 24.2 for English--Marathi, and 31.3 for Marathi--English., Comment: 10 pages

The widespread popularity of soccer across the globe has turned it into a multi-billion dollar industry. As a result, most professional clubs actively engage in talent identification and development programmes. Contemporary research has generally supported the use of executive functions - a class of neuropsychological processes responsible for cognitive behaviours - in predicting a soccer player's future success. However, studies on the developmental evolution of executive functions have yielded differing results in their structural form (such as inverted U-shapes, or otherwise). This article presents the first analysis of changes in the domain-generic and domain-specific executive functions based on longitudinal data measured on elite German soccer players. Results obtained from a latent variable model show that these executive functions experience noticeable growth from late childhood until pre-adolescence, but remain fairly stable in later growth stages. As a consequence, our results suggest that the role of executive functions in facilitating talent identification may have been overly emphasised.

High-resolution gamma-ray spectroscopy of 18N is performed with the Advanced GAmma Tracking Array AGATA, following deep-inelastic processes induced by an 18O beam on a 181Ta target. Six states are newly identified, which together with the three known excitations exhaust all negative-parity excited states expected in 18N below the neutron threshold. Spin and parities are proposed for all located states on the basis of decay branchings and comparison with large-scale shell-model calculations performed in the p-sd space, with the YSOX interaction. Of particular interest is the location of the 0^-_1 and 1^-_2 excitations, which provide strong constrains for cross-shell p-sd matrix elements based on realistic interactions, and help to simultaneously reproduce the ground and first-excited states in 16N and 18N, for the first time. Understanding the 18N structure may also have significant impact on neutron-capture cross-section calculations in r-process modeling including light neutron-rich nuclei.

Lifetimes of low-spin excited states in $^{98}$Zr were measured using the recoil-distance Doppler-shift technique and the Doppler-shift attenuation method. The nucleus of interest was populated in a $^{96}$Zr($^{18}$O,$^{16}$O)$^{98}$Zr two-neutron transfer reaction at the Cologne FN Tandem accelerator. Lifetimes of six low-spin excited states, of which four are unknown, were measured. The deduced $B(E2)$ values were compared with Monte Carlo shell model and interacting boson model with configuration mixing calculations. Both approaches reproduce well most of the data but leave challenging questions regarding the structure of some low lying states., Comment: 13 pages, 12 figures, 3 tables, accepted for publication in Physical Review C

The heaviest N=Z doubly-magic nucleus, $^{100}$Sn, and the neighboring nuclei offer unique opportunities to investigate the properties of nuclear interaction in extreme conditions. In particular, the Cd isotopes are expected to present features similar to those found in the Sn isotopic chain, since they have only two proton holes in the Z=50 shell. In this manuscript, the lifetime measurements of low-lying states in the even-mass $^{102-108}$Cd is presented. Thanks to the powerful detection capabilities of AGATA array and VAMOS++ spectrometer, the unusual employment of multi-nucleon transfer reactions permitted to investigate the first 2$^+$ and 4$^+$ states in all these nuclei, together with various deformed bands in $^{106}$Cd. The results were interpreted in the context of new state-of-the-art beyond-mean-field calculations, using the symmetry-conserving configuration-mixing approach. Despite the similarities in the electromagnetic properties of the low-lying states, there is a fundamental structural difference between the ground-state bands in the Z=48 and Z=50 isotopes. The comparison between experimental and theoretical results revealed a rotational character of the Cd nuclei, which have prolate-deformed ground states with $\beta_2 \approx 0.2$. At this deformation Z=48 becomes a closed-shell configuration, which is favored with respect to the spherical one., Comment: V1 : 14 pages, 12 figures. V2 : 16 pages, 11 figures. V3 : 16 pages, 11 figures (figure 11 corrected); triaxiality evaluated with "quadrupole sum rule". submitted to Phys.Rev.C

A novel Monte Carlo technique has been developed to determine lifetimes of excited states in the tens-to-hundreds femtoseconds range. The method is applied to low-energy heavy-ion binary reactions populating nuclei with complex velocity distributions. Its relevance is demonstrated in connection with the $^{18}$O(7.0 MeV/u) + $^{181}$Ta experiment, performed at GANIL with the AGATA+VAMOS+PARIS setup, to study neutron-rich O, C, N, ... nuclei. Excited states in $^{17}$O and $^{19}$O, with known lifetimes, are used to validate the method over the $\sim$20-400 fs lifetime-sensitivity range. Emphasis is given to the unprecedented position resolution provided by $\gamma$-tracking arrays, which turns out to be essential for reaching the required accuracy in Doppler-shift correction, at the basis of the detailed analysis of $\gamma$-ray lineshape and resulting state lifetime determination. The technique is anticipated to be an important tool for lifetime investigations in exotic neutron-rich nuclei, produced with intense ISOL-type beams.