Subgraph isomorphism is a fundamental problem in graph analysis that seeks to find all instances of a pattern graph within a larger data graph while preserving structural relationships. This NP-complete problem is central to domains such as biological network analysis, social network mining, and quantum circuit optimization. Traditional approaches rely on backtracking algorithms like VF2, which suffer from sequential bottlenecks that limit their ability to exploit modern parallel hardware. In this work, we introduce $Δ$-Motif, a GPU-accelerated subgraph isomorphism algorithm that reformulates the task through the lens of database operations. Our key insight is to represent both data and pattern graphs in tabular form, turning subgraph isomorphism into database primitives including joins, sorts, merges, and filters. $Δ$-Motif decomposes graphs into small building blocks called motifs and systematically combines them using scalable relational operations. By leveraging mature, optimized libraries from the NVIDIA RAPIDS ecosystem and Pandas framework, our solution achieves massive parallelism while remaining portable across systems supporting standard relational primitives. Benchmarks show that $Δ$-Motif outperforms established algorithms like VF2, achieving speedups of up to $595\times$ on GPUs. We further demonstrate its impact by applying it to quantum circuit compilation, addressing a critical bottleneck in quantum computing and enabling scaling to near- and medium-term devices. Our approach democratizes high-performance graph processing by exposing it through familiar database abstractions, eliminating the need for low-level programming while delivering exceptional computational efficiency.

Scientific workflows are critical to scientific data analysis and often involve computationally intensive processing of large datasets on compute clusters. As such, their execution tends to be long-running and resource-intensive, resulting in significant energy consumption and carbon emissions. While carbon-aware computing methods have received considerable attention in general cloud contexts, their application to scientific data analysis workflows remains a critical research gap. Our study addresses this oversight by showing how the delay tolerance, interruptibility, and scalability of scientific workflows can be leveraged for a significantly more sustainable execution model. In this study, we first quantify the problem of carbon emissions associated with running scientific workflows, and then demonstrate the transformative potential for carbon-aware workflow execution. We estimate the carbon footprint of seven real-world Nextflow workflows executed on diverse dedicated cluster and public cloud resources using high-resolution average and marginal grid carbon intensity data from open and commercial data providers. Furthermore, we conduct a systematic evaluation of the impact of carbon-aware temporal shifting, and the dynamic pausing and resuming of the workflow. Moreover, we investigate the impact of resource scaling at both workflow and workflow task levels. Finally, we report substantial potential reductions in overall carbon emissions, with temporal shifting capable of decreasing emissions by over 80%, and resource scaling by 67%.

We study linear regression models with clustered data, high-dimensional controls, and intricate exclusion restrictions. We propose a correctly centered internal instrument IV estimator that accommodates a broad class of exclusion restrictions and allows within-cluster dependence. The estimator admits a simple leave-out interpretation and is computationally tractable. We derive a central limit theorem for the associated quadratic form and propose a robust variance estimator. We also develop identification-robust inference procedures. Our framework extends dynamic panel methods to general clustered settings. We illustrate the approach in a large-scale fiscal intervention in rural Kenya, where spatial interference generates the exclusion-restriction pattern.

This paper investigates unequal error protection (UEP) in digital semantic communication, where semantically important bits require substantially higher reliability than less critical ones. To characterize this heterogeneity, we introduce a novel perspective that treats learned bit-flip probabilities of semantic bits as target error protection levels, thereby directly linking semantic importance to bit-level reliability. This formulation reveals that the required protection levels of the semantic bits may differ by several orders of magnitude, making short-block coding more advantageous than conventional long-block designs. Motivated by this, we develop two UEP frameworks that minimize total blocklength under heterogeneous reliability constraints. First, we propose a bit-level UEP framework based on repetition coding, providing an analytically tractable solution that precisely meets per-bit protection requirements. Second, to improve energy and blocklength efficiency, we design a block-level UEP framework in which the semantic bits are partitioned into short blocks with similar protection levels. Guided by finite blocklength capacity analysis, we derive a closed-form threshold condition for beneficial partitioning and develop a systematic algorithm for integrating modern channel codes. Simulation results on image transmission tasks demonstrate substantial gains in both task performance and transmission efficiency compared with conventional equal-protection schemes.

The equations of classical mechanics can be used to model the time evolution of countless physical systems, from the astrophysical to the atomic scale. Accurate numerical integration requires small time steps, which limits the computational efficiency -- especially in cases such as molecular dynamics that span wildly different time scales. Using machine-learning (ML) algorithms to predict trajectories allows one to greatly extend the integration time step, at the cost of introducing artifacts such as lack of energy conservation and loss of equipartition between different degrees of freedom of a system. We propose learning data-driven structure-preserving (symplectic and time-reversible) maps to generate long time-step classical dynamics and show that this method is equivalent to learning the mechanical action of the system of interest. These models can be learned based on short reference trajectories, and be transferred across thermodynamic conditions and chemical composition. We show that an action-derived ML integrator eliminates the pathological behavior of non-structure-preserving ML predictors, and that the method can be applied iteratively, serving as a correction to computationally cheaper direct predictors.

We prove that for any reversible Finsler metric on S2, the number of prime closed geodesics grows quadratically with respect to length. The main tools are an improvement on Franks' theorem about the number of periodic points of area-preserving annulus maps, and the theory of cylindrical contact homology in the complement of a link.

We present $\texttt{PyBird-JAX}$, a differentiable, $\texttt{JAX}$-based implementation of $\texttt{PyBird}$, using internal neural network emulators to accelerate computationally costly operations for rapid large-scale structure (LSS) analysis. $\texttt{PyBird-JAX}$ computes one-loop EFTofLSS predictions for redshift-space galaxy power spectrum multipoles in 1.2 ms on a CPU and 0.2 ms on a GPU, achieving 3-4 orders of magnitude speed-up over $\texttt{PyBird}$. The emulators take a compact spline-based representation of the input linear power spectrum $P(k)$ as feature vectors, making the approach applicable to a wide range of cosmological models. We rigorously validate its accuracy against large-volume simulations and on BOSS data, including cosmologies not explicitly represented in the training set. Leveraging automatic differentiation, $\texttt{PyBird-JAX}$ supports Fisher forecasting, Taylor expansion of model predictions, gradient-based searches, and vectorised ensemble sampling. Interfaced with a variety of samplers and Boltzmann solvers, $\texttt{PyBird-JAX}$ provides a high-performance, end-to-end inference pipeline. Combined with a symbolic-$P(k)$ generator, a typical Stage-4 LSS MCMC converges in minutes on a GPU. Our results demonstrate that $\texttt{PyBird-JAX}$ delivers the precision and speed required for upcoming LSS surveys, opening the door to accelerated cosmological inference with minimal accuracy loss and no pretraining. In a companion paper [1], we put $\texttt{PyBird-JAX}$ to use in achieving LSS marginalised constraints free from volume projection effects through non-flat measures.

Purpose: Low-field MRI systems operate at single MHz-range frequencies, where signal losses are primarily dominated by thermal noise from the radio-frequency (RF) receive coils. Achieving operation close to this limit is essential for maximizing imaging performance and signal-to-noise ratio (SNR). However, electromagnetic interference (EMI) from cabling, electronics, and patient loading often degrades system performance. Our goal is to develop and validate a practical protocol that guides users in identifying and suppressing electromagnetic noise in low-field MRI systems, enabling operation near the thermal noise limit. Methods: We present a systematic, stepwise methodology that includes diagnostic measurements, hardware isolation strategies, and good practices for cabling and shielding. Each step is validated with corresponding noise measurements under increasingly complex system configurations, both unloaded and with a human subject present. Results: Noise levels were monitored through the incremental assembly of a low-field MRI system, revealing key sources of EMI and quantifying their impact. Final configurations achieved noise within 1.5x the theoretical thermal bound with a subject in the scanner. Image reconstructions illustrate the direct relationship between system noise and image quality. Conclusion: The proposed protocol enables low-field MRI systems to operate close to fundamental noise limits in realistic conditions. The framework also provides actionable guidance for the integration of additional system components, such as gradient drivers and automatic tuning networks, without compromising SNR.

We investigate the response of a Ramsey interferometric quantum sensor based on a frustrated, three-spin system (a Kitaev trimer) to a classical time-dependent field (signal). The system eigenspectrum is symmetric about a critical point, $|\vec{b}| = 0$, with four of the spectral components varying approximately linearly with the magnetic field and four exhibiting a nonlinear dependence. Under the adiabatic approximation and for appropriate initial states, we show that the sensor's response to a zero-mean signal is such that below a threshold, $|\vec{b}| < b_\mathrm{th}$, the sensor does not respond to the signal, whereas above the threshold, the sensor acts as a detector that the signal has occurred. This thresholded response is approximately omnidirectional. Moreover, when deployed in an entangled multisensor configuration, the sensor achieves sensitivity at the Heisenberg limit. Such detectors could be useful both as stand-alone units for signal detection above a noise threshold and in two- or three-dimensional arrays, analogous to a quantum bubble chamber, for applications such as particle track detection and long-baseline telescopy.

Complex spin configurations in magnetic materials, ranging from collinear single-Q to noncoplanar multi-Q states, exhibit rich symmetry and chiral properties. However, their detailed characterization is often hindered by the limited spatial resolution of neutron diffraction techniques. Here we employ magnetic circular dichroism (MCD) and magnetic linear dichroism (MLD) to investigate the triangular lattice antiferromagnet Co$_{1/3}$TaS$_2$, revealing three-state (Z3) nematicity and also spin chirality across its multi-Q magnetic phases. At intermediate temperatures, the presence of MLD identifies nematicity arising from a single-Q stripe phase, while at high magnetic fields and low temperatures, a phase characterized solely by MCD emerges, signifying a purely chiral non-coplanar triple-Q state. Notably, at low temperatures and small fields, we discover a unique phase where both chirality and nematicity coexist. A theoretical analysis based on a continuous multi-Q manifold captures the emergence of these distinct magnetic phases, as a result of the interplay between four-spin interactions and weak magnetic anisotropy. Additionally, MCD and MLD microscopy spatially resolves the chiral and nematic domains. Our findings establish Co$_{1/3}$TaS$_2$ as a rare platform hosting diverse multi-Q states with distinct combinations of spin chirality and nematicity while demonstrating the effectiveness of polarized optical techniques in characterizing complex magnetic textures.

We formulate a ten-dimensional version of Kodaira-Spencer gravity on a Calabi-Yau five-fold that reproduces the classical Maurer-Cartan equation governing supersymmetric heterotic moduli. Quantising this theory's quadratic fluctuations, we show that its one-loop partition function simplifies to products of holomorphic Ray-Singer torsions and exhibits an anomaly that factorises as in $SO(32)$ and $E_8\times E_8$ supergravity. Based on this, we conjecture that this theory is the $SU(5)$-twisted version of ten-dimensional $N=1$ supergravity coupled to Yang-Mills and show that is related to the type I Kodaira-Spencer theory of Costello-Li via a non-local field redefinition. The counter-terms needed to cancel the anomaly and retain gauge invariance for the one-loop effective theory reconstruct the differential of a recently discovered double-extension complex. This complex has non-tensorial extension classes and its first cohomology counts the infinitesimal moduli of heterotic compactifications modulo order $α'^2$ corrections.

The aim of this work is to construct mock galaxy catalogues that accurately reproduce the redshift evolution of galaxy number density, clustering statistics, and baryonic properties, such as stellar mass for luminous red galaxies (LRGs) and absolute magnitude in the $r$-band for the bright galaxy sample (BGS), based on the first three years of observations from the Dark Energy Spectroscopic Instrument (DESI). To achieve this, we applied the subhalo abundance matching (SHAM) technique to the Uchuu $N$-body simulation, which follows the evolution of 2.1 trillion particles within a volume of $8\,h^{-3}\,\mathrm{Gpc}^{3}$, assuming a Planck base-$Λ$CDM cosmology. Using SHAM, we populated Uchuu subhalos with LRGs and BGS-BRIGHT ($r<19.5$) galaxies up to redshift $z=1.1$, assigning stellar masses to LRGs and luminosities to BGS galaxies (up to $M_{\rm r}\leq 20$). Furthermore, we analyzed the clustering dependence on stellar mass and luminosity for each tracer. Our results show that the Uchuu BGS-BRIGHT and LRG mocks accurately reproduce the observed redshift evolution of clustering, with better than 5\% agreement for separations of $1<r<20\,h^{-1}\,\mathrm{Mpc}$ and below 10\% for $0.1<r<1\,h^{-1}\,\mathrm{Mpc}$. For the Uchuu-LRG mock, we successfully captured the stellar mass dependence of clustering, while for the Uchuu-BGS mock, we replicated the clustering for various volume-limited subsamples. We also find good agreement between the data and mocks in the dependence of large-scale bias on luminosity for BGS-BRIGHT galaxies and on stellar mass for LRGs. Altogether, these results equip DESI with robust tools for generating high-fidelity lightcones for the remainder of the survey, thereby enhancing our understanding of the galaxy--halo connection.

Sandqvis's semantics for classical logic without bivalence resolves the question of an anti-realist account of classical reasoning after Dummett. This paper applies the framework to the essential questions of metamathematics. The system intuitively handles $ω$-incompleteness, makes induction meaning-constitutive, and yields an elementary consistency proof for Peano Arithmetic using only ordinary induction on the natural numbers, with no appeal to transfinite ordinals or recognition-transcendent truth.

Identifying patterns of relations among the units of a complex system from measurements of their activities in time is a fundamental problem with many practical applications. Here, we introduce a method that detects dependencies of any order in multivariate time series data. The method first transforms a multivariate time series into a symbolic sequence, and then extract statistically significant strings of symbols through a Bayesian approach. Such motifs are finally modelled as the hyperedges of a hypergraph, allowing us to use network theory to study higher-order interactions in the original data. When applied to neural and social systems, our method reveals meaningful higher-order dependencies, highlighting their importance in both brain function and social behaviour.

Efficient tensor computation is a cornerstone of modern deep learning (DL) workloads, yet existing approaches struggle to achieve flexible and performant design and implementation of tensor layouts -- mappings between logical tensors and hardware resources. The increasing complexity of DL algorithms and hardware demands a generic and systematic approach to handling tensor layouts. In this work, we introduce Linear Layouts, a novel approach that models tensor layouts using linear algebra over $\mathbb{F}_2$. By representing tensor layouts as binary matrices acting on the bits of the hardware representation, our approach enables a generic layout definition -- as opposed to the classical case-by-case approach -- and allows for generic layout-to-layout conversions, eliminating the quadratic explosion that plagues existing solutions. We integrate linear layouts with Triton and demonstrate their effectiveness in optimizing individual Triton operators as well as kernels written in Triton. We also show that linear layouts reduce engineering effort in the compiler backend while fixing several bugs in Triton's legacy layout system.

We evaluate the role of quantum coherence as a thermodynamic resource in a noisy, Markovian, one-qubit heat engine. By consuming the coherence of noisy quantum states, we demonstrate that the engine can surpass the classical efficiency limit when operating according to a quantum Otto cycle. The engine's non-classical nature is demonstrated by its violation of the Leggett-Garg's temporal correlations inequality. Amplitude damping increases the extractable work under partial thermalization, thereby increasing the efficiency. In contrast, phase damping increases the extractable work under partial thermalization but reduces the efficiency. We implement the entire Otto cycle in a quantum circuit, simulating realistic amplitude and phase damping channels, as well as gate-level noise. We introduce an operational measure of the circuit's thermodynamic cost to establish a direct link between energy consumption and information processing in quantum heat engines.

Let us consider the set $Ω(\triangle ABC)$ of all tetrahedra $ABCD$ with a given non-degenerate base $ABC$ in $\mathbb{E}^3$ and $D$ lying outside the plane $ABC$. Let us denote by $Σ(\triangle ABC)$ the set $\left\{\Bigl(\cos \overlineα,\cos \overlineβ,\cos \overlineγ \Bigr)\in \mathbb{R}^3\,|\, ABCD \in Ω(\triangle ABC)\right\}$, where $\overlineα=\angle BDC$, $\overlineβ=\angle ADC$, and $\overlineγ=\angle ADB$. The paper is devoted to the problem of determining of the closure of $Σ(\triangle ABC)$ in $\mathbb{R}^3$ and its boundary.

Modern systems exhibit unprecedented complexity due to their increased scale, interconnectedness, and the heterogeneity of their digital and physical components. In response to scaling challenges, the system of systems paradigm proposes flexible aggregations of subsystems into a larger whole, while maintaining the independence of subsystems to various degrees. In response to the cyber-physical convergence, the digital twin paradigm proposes a tight coupling between digital and physical components through computational reflection and precise control. As these two paradigms address distinct parts of the overall challenge, combining the two promises more comprehensive methods to engineer what we call systems of twinned systems. The noticeably growing body of knowledge on systems of twinned systems calls for a review of the state of the art. In this work, we report on our systematic literature survey of systems of twinned systems. We screened over 2,500 potential studies, of which we included 80 and investigated them in detail. To converge system of systems and digital twins, we derive a classification framework for systems of twinned systems that is backward compatible with the currently accepted theories of system of systems and digital twins.

A wireless wearable Electrical Impedance Tomography (EIT) system has been developed utilizing the AD5933 chip to achieve real-time imaging of lung respiration. The system employs a voltage excitation method tailored to human impedance characteristics, injecting current by applying a known voltage and measuring the resulting current through the body. Additionally, specific measures have been implemented to effectively suppress signal oscillations and leakage currents caused by parasitic capacitances. To enhance data acquisition speed, the system employs five parallel AD5933 units, with multiple techniques implemented to ensure high synchronization during simultaneous measurements. Performance testing shows that the system achieves a signal-to-noise ratio greater than 50 dB, a relative standard deviation below 0.3%, and a reciprocity error under 0.8%. Imaging experiments using a water tank phantom, human lungs during breathing, and a resting human calf further demonstrate that this portable EIT system can accurately measure biological tissues with high precision and low cost.

All-to-All(v) communication is a critical primitive in modern machine learning workloads, particularly mixture-of-experts (MoE) models. Unfortunately, efficient scheduling is challenging due to workload skew, heterogeneous two-tier fabrics, and incast congestion, compounded by the dynamic nature of MoE workloads, where traffic shifts every few hundred milliseconds. Existing schedulers are hardly scalable, incurring seconds to hours of synthesis time, making them impractical. We present FAST, an efficient All-to-All(v) scheduler. FAST addresses skew through intra-server rebalancing and enforces balanced, one-to-one scale-out transfers that avoid incast. Evaluated extensively on both NVIDIA H200 and AMD MI300X clusters, FAST consistently outperforms state-of-the-art solutions on skewed workloads while reducing synthesis time by orders of magnitude.

We present preconditioning techniques to solve linear systems of equations with a block two-by-two and three-by-three structure arising from finite element discretizations of the fictitious domain method with Lagrange multipliers. In particular, we propose two augmented Lagrangian-based preconditioners to accelerate the convergence of iterative solvers for such classes of linear systems. We consider two relevant examples to illustrate the performance of these preconditioners when used in conjunction with flexible GMRES: the Poisson and the Stokes fictitious domain problems. A spectral analysis is established for both exact and inexact versions of the preconditioners. We show the effectiveness of the proposed approach and the robustness of our preconditioning strategy through extensive numerical tests in both two and three dimensions.

We systematically derive the dissipationless quantum kinetic equation for a multi-band free fermionic system with U(1) symmetry. Using the Moyal product formalism, we fully band-diagonalize the dynamics. Expanding to the second order in gradients, which is beyond the semiclassical limit, we give a complete analysis of the band-resolved thermodynamics and transport properties, especially those arising from the quantum geometric tensor. We apply our framework to a Bloch band theory under electric fields near equilibrium and find the linear and nonlinear transport coefficients. We also obtain the dynamical density-density response functions in the metallic case, including quantum metric corrections. Our results and approach can be applied very generally to multi-band problems even in situations with spatially varying Hamiltonians and distributions.

The Radio Neutrino Observatory-Greenland (RNO-G, at Summit Station) experiment comprises an extensive fat-dipole antenna array deployed into ice boreholes over an eventual area of approximately 35 ${\rm km}^2$. Since the RNO-G experimental sensitivity depends on the radio-frequency properties of the firn, which are known to vary laterally on sub-km distance scales and vertically on sub-meter distance scales, a technique for quickly extracting information on firn ice properties with depth ($n(z)$) during drilling and deployment is desirable. Given that a dipole's resonant wavelength is fixed by geometry, the resonant frequency $f_{res}$ (measured as an S-parameter reflection coefficient [`$S_{11}$'] minimum) scales inversely with the local refractive index, allowing a translation of a depth-dependent $S_{11}$(z) profile into $n(z)$. $S_{11}$(z) data were initially taken in August, 2024 using a dipole lowered into a newly-drilled $98\pm 1$-mm diameter, 350-m deep borehole at Summit Station, Greenland, approximately 1 km from the site of the original GISP-2 core; improved measurements were subsequently made in May, 2025. We conclude that $S_{11}$(z) data can be used to estimate \RIP, on 50 cm vertical scales, at the per-cent level of accuracy required by experiments such as RN0-G.

This paper addresses the existence and uniqueness of solutions to Reflected Generalized Backward Stochastic Differential Equations (GRBSDEs) within a general filtration that supports a Brownian motion and an independent integer-valued random measure. Our study focuses on cases where the given data satisfy appropriate $\mathbb{L}^2$-integrability conditions and the coefficients satisfy a monotonicity assumption. Additionally, we establish a connection between the solution and an optimal control problem over the set of stopping times.

In this paper, we introduce a family of integral transforms, denoted by \(\mathcal{O}_α\), and constructed via kernel fusion of the fractional Fourier transform (FRFT) with angle \(α\notin π\mathbb{Z}\). We demonstrate that the \(\mathcal{O}_α\)-transformation constitutes a well-defined integral operator by establishing its basic operational properties. Besides, we survey various mathematical aspects of the uncertainty principles for the $\mathcal{O}_α$-transform, including Heisenberg's inequality, logarithmic uncertainty inequality, local uncertainty inequality, Hardy's inequality, Pitt's inequality, and Beurling-H{ö}rmander's theorem.

This paper classifies t-structures on the local derived category of a 3-fold flopping contraction, that are intermediate with respect to the heart of perverse coherent sheaves. Equivalently, this describes the complete lattice of torsion classes for the associated modification algebra. The intermediate hearts are (1) categories of coherent sheaves on birational models and tilts thereof in skyscrapers, (2) algebraic t-structures described in the homological minimal model programme, or (3) combinations of the above over appropriate open covers. An analogous classification is also proved for minimal (and partial) resolutions of Kleinian singularities, thus providing a description of all torsion pairs in the module categories of (contracted) affine preprojective algebras. The results have immediate applications to the classification of spherical modules and (semi)bricks, and are first steps towards describing all t-structures and spherical objects in derived categories of surfaces and 3-folds.

The aim of this paper is to prove a theorem of C.~Miranda on the Hölder regularity of convolution operators acting on the boundary of an open set in the limiting case in which the open set is of class $C^{1,1}$ and the densities are of class $C^{0,1}$. The convolution operators that we consider are generalizations of those that are associated to layer potential operators, which are a useful tool for the analysis of boundary value problems.

We present a method to construct high-order polynomial approximate invariants (AI) for non-integrable Hamiltonian dynamical systems, and apply it to modern ring-based particle accelerators. Taking advantage of a special property of one-turn transformation maps in the form of a square matrix, AIs can be constructed order-by-order iteratively. Evaluating AI with simulation data, we observe that AI's fluctuation is actually a measure of chaos. Through minimizing the fluctuations with control knobs in accelerators, the stable region of long-term motions could be enlarged.

The positional population count operation pospopcnt() counts for an array of w-bit words how often each of the w bits was set. Various applications in bioinformatics, database engineering, and digital processing exist. Building on earlier work by Klarqvist et al., we show how positional population counts can be rapidly computed using SIMD techniques with good performance from the first byte, approaching memory-bound speeds for input arrays of as little as 4 KiB. Improvements include an improved algorithm structure, better handling of unaligned and very short arrays, as well as faster bit-parallel accumulation of intermediate results. We provide a generic algorithm description as well as implementations for various SIMD instruction set extensions, including Intel AVX2, AVX-512, and ARM ASIMD, and discuss the adaption of our algorithm to other platforms.

We present an analysis on flavour anomalies in semileptonic rare $B$-meson decays using an effective field theory approach and assuming that new physics affects only one generation in the interaction basis and non-universal mixing effects are generated by the rotation to the mass basis. A global fit to experimental data is performed, focusing on LFU ratios $R_{D^{(*)}}$ and $R_{J/ψ}$ and branching ratios that exhibit tensions with Standard Model predictions on $B \rightarrow K^{(*)} ν\barν$ decays. In our analysis, we use a Machine Learning Montecarlo algorithm, a framework that emulates the highly non-Gaussian structure of the likelihood landscape with minimal training cost. This method enables the generation of high-resolution confidence regions and detailed correlation analyses. By comparing three different scenarios, we show that the one that introduces only mixing between the second and third quark generations and no mixing in the lepton sector, as well as independent coefficients for the singlet and triplet four fermion effective operators, provides the best fit to the experimental data. A comparison with previous results is performed. We highlight the key strengths of the Machine Learning framework in our analysis.