搜索结果：Discover quantum science

Random quantum states have various applications in quantum information science. We discover a new ensemble of quantum states that serve as an ε-approximate state t-design while possessing extremely low entanglement, magic, and coherence. These resources can reach their theoretical lower bounds, Ω(log(t/ε)), which are also proven in this Letter. This implies that, for fixed t and ε, entanglement, magic, and coherence do not scale with the system size, i.e., O(1) with respect to the total number of qubits n. Moreover, we explicitly construct an ancilla-free shallow quantum circuit for generating such states by transforming k-qubit approximate state designs into n-qubit ones without increasing the support size. The depth of such a quantum circuit, O(t[logt]^{3}log n log[1/ε]), is the most efficient among existing algorithms without ancilla qubits. A class of quantum circuits proposed in our Letter offers reduced cost for classical simulation of random quantum states, leading to potential applications in quantum information processing. As a concrete example, we propose classical shadow tomography using an estimator with superpositions between only two states, from which almost all quantum states can be efficiently certified by requiring only O(1) measurements and classical postprocessing time.

Utilizing the general theory of open quantum systems to investigate the exact dynamical evolution of simple bilinear systems, we discover a mechanism of the dynamical genesis of quantum entanglement. We focus in detail on the exact quantum evolution dynamics of two photonic modes (or any two bosonic modes) coupled to each other through a linear interaction, as the simplest system of open quantum systems that we have investigated in the past two decades. Such a linear coupling alone fails to produce two-mode entanglement. We also start with an initially separable pure state of the two modes. By solving exactly the quantum equation of motion without relying on the probabilistic interpretation, we find that when the initial state of one mode is different from a coherent state (a minimum uncertainty wave packet with equal variance in the conjugate quadratures that corresponds to a well-defined classically "particle"), the causality in the time-evolution of each mode is internally violated. It also leads to the emergence of quantum entanglement between the two modes. The lack of causality is the nature of statistics. We discover that it is the internal violation of causality in the reduced (subsystem) dynamical evolution that results in the emergence of entanglement and statistic probability in quantum mechanics, even though the dynamical evolution of the whole system completely obeys the deterministic Schrödinger equation. This conclusion is valid for the quantum dynamics of more complicated composite systems. It may provide the fundamental mechanism of the dynamical genesis for both the entanglement and the statistical probability within the deterministic framework of quantum mechanics, which is the longest-standing problem that has not been fully understood since the birth of quantum mechanics.

Uncertainty principle establishes a remarkable lower bound to predict the measured outcome of two non-commuting observables. In this paper, based on the quantum renormalization-group method, we study the relation between quantum-memory-assisted entropic uncertainty relation (QMA-EUR) and quantum phase transition (QPT) in the spin XXZ model. The results shows that the entropic uncertainty and the lower bound have similar traits. In addition, we propose two schemes, one is based on quantum discord and classical correlation, the other Holevo quantity and mutual information, both can tighten the bound of EUR in the presence of quantum memory. The tighter the entropy uncertainty relationship is, the higher the accuracy of the predicted results will be. Moreover, we can obtain the optimal lower bound with the help of Holevo quantity and mutual information, which have the best optimization effect in this model. Additionally, we study QPT by virtue of EUR and after a certain number of iterations, finding that the value of QMA-EUR of the whole block-block state can form two saturated values, which are related to two different phases: spin-fluid phase and Néel phase. Afterwards, we discover that the QMA-EUR of the block-block state obeys the nonanalytic and scaling properties with entropic uncertainty relation exponent associated with correlation length. Our findings show that QMA-EUR deserves to be used as an effective tool in reflecting quantum criticality for more quantum many-body systems and may also shed light on many applications in quantum physics including the quantum key distribution, the detection of QPT and the evaluation of the capacity of quantum computation in critical systems.

Billions of organic molecules have been computationally generated, yet functional inorganic materials remain scarce due to limited data and structural complexity. Here we introduce Structural Constraint Integration in a GENerative model (SCIGEN), a framework that enforces geometric constraints, such as honeycomb and kagome lattices, within diffusion-based generative models to discover stable quantum materials candidates. SCIGEN enables conditional sampling from the original distribution, preserving output validity while guiding structural motifs. This approach generates ten million inorganic compounds with Archimedean and Lieb lattices, over 10% of which pass multistage stability screening. High-throughput density functional theory calculations on 26,000 candidates shows over 95% convergence and 53% structural stability. A graph neural network classifier detects magnetic ordering in 41% of relaxed structures. Furthermore, we synthesize and characterize two predicted materials, TiPd0.22Bi0.88 and Ti0.5Pd1.5Sb, which display paramagnetic and diamagnetic behaviour, respectively. Our results indicate that SCIGEN provides a scalable path for generating quantum materials guided by lattice geometry.

The quantum loop and dimer models are archetypal correlated systems with local constraints. With natural foundations in statistical mechanics, they are of direct relevance to various important physical concepts and systems, such as topological order, lattice gauge theories, geometric frustrations, or more recently Rydberg array quantum simulators. However, how the thermal fluctuations interact with constraints has not been explored in the important class of nonbipartite geometries. Here we study, via unbiased quantum Monte Carlo simulations and field theoretical analysis, the finite-temperature phase diagram of the quantum loop model on the triangular lattice. We discover that the recently identified, "hidden" vison plaquette (VP) quantum crystal [X. Ran et al., Hidden orders and phase transitions for the fully packed quantum loop model on the triangular lattice, Commun. Phys. 7, 207 (2024)CMPYEL0868-316610.1038/s42005-024-01680-z] experiences a finite-temperature continuous transition, which smoothly connects to the (2+1)d Cubic* quantum critical point separating the VP and Z_{2} quantum spin liquid phases. This finite-temperature phase transition acquires a unique property of "remnants of fractionalization" at finite temperature, in that, both the cubic order parameter-the plaquette loop resonance-and its constituent-the vison field-exhibit independent criticality signatures. This phase transition is connected to a three-state Potts transition between the lattice nematic phase and the high-temperature disordered phase. We discuss the relevance of our results for current experiments on quantum simulation platforms.

The Christoffel symbol is an essential quantity in Einstein's general theory of relativity. We discover that an electric field can induce a nonlinear magnetization in quantum materials, described by a Christoffel symbol defined in the Hilbert space of quantum states (quantum Christoffel symbol). Quite different from the previous scenarios, this orbital magnetization does not need spin-orbit coupling and inversion symmetry breaking. Through symmetry analysis and first-principles calculations, we identify a number of point groups and 2D material candidates (e.g., BiF_{3}, ZnI_{2}, and Ru_{4}Se_{5}) that host this quantum Christoffel nonlinear magnetization. More importantly, this nonlinear magnetization allows the quantum Christoffel symbol to be probed by optical techniques such as magneto-optical Kerr spectroscopy or transport measurements such as tunneling magnetoresistance. This quantum Christoffel nonlinear magnetization gives a paradigm of how geometry dictates physics.

Conventional wisdom dictates that quantum effects become unimportant at high temperatures. In magnets, when the thermal energy exceeds interactions between atomic magnetic moments, the moments are usually uncorrelated, and classical paramagnetic behavior is observed. This thermal decoherence of quantum spin behaviors is a major hindrance to quantum information applications of spin systems. Remarkably, our neutron scattering experiments on Yb chains in an insulating perovskite crystal defy these conventional expectations. We find a sharply defined spectrum of spinons, fractional quantum excitations of spin-1/2 chains, to persist to temperatures much higher than the scale of the interactions between Yb magnetic moments. The observed sharpness of the spinon continuum's dispersive upper boundary indicates a spinon mean free path exceeding  ≈ 35 inter-atomic spacings at temperatures more than an order of magnitude above the interaction energy scale. We thus discover an important and highly unique quantum behavior, which expands the realm of quantumness to high temperatures where entropy-governed classical behaviors were previously believed to dominate. Our results have profound implications for spin systems in quantum information applications operating at finite temperatures and motivate new developments in quantum metrology.

In this work, we introduce a new qubit mapping strategy for the Variational Quantum Eigensolver (VQE) applied to nuclear shell model calculations, where each Slater Determinant (SD) is mapped to a qubit, rather than assigning qubits to individual single-particle states. While this approach may increase the total number of qubits required in some cases, it enables the construction of simpler quantum circuits that are more compatible with current noisy intermediate-scale quantum (NISQ) devices. We apply this method to seven nuclei: Four lithium isotopes [Formula: see text]Li from the p-shell, [Formula: see text]F from the sd-shell, and two heavier nuclei ([Formula: see text]Po, and [Formula: see text]Pb). We run circuits representing their ground states on a noisy simulator (IBM's FakeFez backend) and quantum hardware ([Formula: see text]). For heavier nuclei, we demonstrate the feasibility of simulating [Formula: see text]Po and [Formula: see text]Pb as 22- and 29-qubit systems, respectively. Additionally, we employ Zero-Noise Extrapolation (ZNE) via two-qubit gate folding to mitigate errors in both simulated and hardware-executed results. Post-mitigation, the best results show less than 4 % deviation from shell model binding energy predictions across all nuclei studied. This SD-based qubit mapping proves particularly effective for lighter nuclei and two-nucleon systems, offering a promising route for near-term quantum simulations in nuclear physics.

Recent advances in artificial intelligence (AI) are revolutionizing materials science by unlocking unprecedented capabilities in designing novel compounds and accurately predicting their properties. Among these, graph-based machine learning (ML) algorithms have garnered significant attention for their ability to capture complex atomic interactions and use them as effective descriptors. In this study, we integrated state-of-the-art generative AI (Gen AI) and ML techniques with quantum mechanical calculations to discover novel next-generation electrolytes for alkali metal batteries. We developed a Generative Adversarial Network (GAN) framework incorporating a graph-based generator and discriminator models to generate novel electrolyte candidates. The GAN model was trained on a subset of approximately 1 million molecules from the GDB-11 database, which enabled the generation of 30,000 unique and chemically valid molecules. Concurrently, a Message Passing Neural Network (MPNN) model was trained for property prediction by utilizing the QM9 dataset. Using the trained MPNN model, we predicted the properties of the newly generated molecules and screened the candidates based on the criteria of negative standard enthalpy of formation and a wide HOMO-LUMO gap. First-principles density functional theory (DFT) calculations were conducted for additional screening and to evaluate key thermodynamic and electrochemical properties, including standard enthalpy of formation, oxidation potential, and reduction potentials. Finally, a set of 26 promising candidates was acquired with outstanding electrochemical characteristics. Our findings demonstrate the potential of AI-driven approaches to discover high-performance, stable, and efficient electrolytes as promising alternatives to conventional organic electrolytes for next-generation energy storage systems.

The electronic-structure informatics (ESI) descriptor set was applied to discover novel α-glucosidase inhibitors from a natural product (NP) database. The in silico screening was carried out through regression modelling for inhibitory activity (pIC50) using XGBoost with the ESI descriptor set. The optimized model achieved a test R2 of 0.85, demonstrating its high predictive accuracy. To explore potent NPs for α-glucosidase inhibition, in silico screening of 2623 NPs was performed. Already known NP-α-glucosidase inhibitors such as theasinensin A, chebulagic acid, and casuarictin were "re-identified" through the screening. It also revealed structurally novel NP compounds with moderate inhibitory activity and new scaffolds different from those of the known inhibitors. A series of docking simulations on the newly discovered compounds revealed that their binding scores are higher than a marketed drug, acarbose. These results demonstrate the applicability and uniqueness of the ESI descriptor set in "scaffold hopping" using NP databases.Scientific contributionThis study shows that the electronic-structure informatics (ESI) descriptor set supports effective scaffold hopping for discovering α-glucosidase inhibitors from natural product (NP) libraries beyond conventional structure-based searches. By combining quantum-chemistry-derived ESI descriptors with machine learning, we identify structurally novel NP inhibitor candidates, which exhibit competitive predicted activity despite low similarity to known chemotypes. This work demonstrates the value of electronic-structure information in in silico screening for identifying chemically diverse candidates.

Quantum-dot optoelectronics, pivotal for lighting, lasing and photovoltaics, rely on nanocrystalline oxide electron-injection layer. Here, we discover that the prevalent surface magnesium-modified zinc oxide electron-injection layer possesses poor n-type attributes, leading to the suboptimal and encapsulation-resin-sensitive performance of quantum-dot light-emitting diodes. A heavily n-doped nanocrystalline electron-injection layer-exhibiting ohmic transport with 1000 times higher electron conductivity and improved hole blockage-is developed via a simple reductive treatment. The resulting sub-bandgap-driven quantum-dot light-emitting diodes exhibit optimal efficiency and extraordinarily-high brightness, surpassing current benchmarks by at least 2.6-fold, and reaching levels suitable for quantum-dot laser diodes with only modest bias. This breakthrough further empowers white-lighting quantum-dot light-emitting diodes to exceed the 2035 U.S. Department of Energy's targets for general lighting, which currently accounts for ~15% of global electricity consumption. Our work opens a door for understanding and optimizing carrier transport in nanocrystalline semiconductors shared by various types of solution-processed optoelectronic devices.

Machine learning (ML) has been widely used to accelerate the discovery of organic light-emitting diode (OLED) materials, but its application to improving device-level performance has been limited. Here, we develop an ML workflow that explicitly incorporates exciplex-specific design criteria, including exciplex-considered high triplet energy criteria, deep lowest-unoccupied molecular orbital (LUMO) alignment, high bond dissociation energy (BDE), and proper reorganization energy, as factors that directly link the molecular structure with device stability and exciton dynamics. Based on this physics-informed approach, we screen n-type hosts for phosphorescence-sensitized fluorescent (PSF) OLEDs, and identify two silane-functionalized n-type hosts, DPSiTrz and DBiPSiTrz, that successfully form an exciplex with p-type BPP-BCZ. Bulky silane groups are introduced to prevent aggregation-induced quenching while maintaining donor-acceptor electronic coupling to form an exciplex. As a result, these exciplex hosts yield high triplet energies (>2.50 eV), reduced nonradiative decay, and minimal back energy transfer (BET) from the phosphorescent sensitizer Ir(ppy)2(acac). The fabricated green PSF OLEDs based on these exciplex hosts show external quantum efficiencies (EQEs) of up to 39.4%, with limited efficiency roll-off (L90 > 100,000 cd m-2) and long operational stability (LT95 = 134.4 h at 5000 cd m-2), validating that the exciplex-informed ML design rules translate into experimentally robust devices. These results demonstrate that an ML-enabled molecular design strategy that can represent device-level exciton behavior and long-term stability is an effective method to discover high-efficiency, durable OLEDs.

Mechanical strain presents an effective control over symmetry-breaking phase transitions. In quantum paraelectric SrTiO3, strain can induce ferroelectric order via modification of the local Ti potential energy landscape. However, brittle bulk materials can only withstand limited strain range (~0.1%). Taking advantage of nanoscopically-thin freestanding membranes, we demonstrate an in-situ strain-induced reversible ferroelectric transition in freestanding SrTiO3 membranes. We measure the ferroelectric order by detecting the local anisotropy of the Ti 3d orbital signature using X-ray linear dichroism at the Ti-K pre-edge, while the strain is determined by X-ray diffraction. With reduced thickness, the SrTiO3 membranes remain elastic with >1% tensile strain cycles. A robust displacive ferroelectricity appears beyond a temperature-dependent critical strain. Interestingly, we discover a crossover from a classical ferroelectric transition to a quantum regime at low temperatures, which enhances strain-induced ferroelectricity. Our results offer new opportunities to strain engineer functional properties in low dimensional quantum materials and provide new insights into the role of ferroelectric fluctuations in quantum paraelectric SrTiO3.

Pyrazole derivatives are of growing interest due to their diverse pharmacological activities. However, their biological activity is often highly sensitive to subtle structural modifications. Existing quantitative structure-activity relationships (QSAR) approaches frequently fail to capture the conformational flexibility and nonlinear structure-activity relationships (SAR) of such heterocyclic scaffolds, creating a gap in the accurate prediction of their biological profiles. Therefore, there is a strong need for more robust and predictive computational frameworks. This study addresses this gap by integrating four-dimensional (4D)-QSAR descriptors with hybrid machine learning (ML) techniques to improve predictive accuracy and provide a more reliable tool for structure-based drug design. In this work, it was aimed to investigate the SAR of a series of pyrazole-based compounds using this advanced integrative computational strategy. The dataset consisted of 54 pyrazole derivatives, of which 50 compounds were used for model construction and 4 compounds were reserved as a test set for validation. Although the test set was limited in size, the selected compounds were structurally representative of the training set, sharing the same core scaffold while covering different substitution patterns and biological activity values. The 4D-QSAR approach included multiple conformations of each compound and utilized matrix-based representations of geometric and electronic properties to capture dynamic molecular behavior. A pharmacophore model was generated using EMRE software based on the spatial and electronic features of used compounds. EMRE is an in-house software developed by our research group. It has been employed in several previously published 4D-QSAR studies for electron-conformational matrix of contiguity construction, pharmacophore modeling, descriptor matrix generation, and activity prediction. EMRE operates on standard geometric and electronic descriptors derived from quantum-chemical calculations, ensuring methodological transparency and reproducibility despite its proprietary implementation. Comparable performance trends obtained with EMRE-based 4D-QSAR models have been reported in previous studies, supporting the validity of the software for pharmacophore-driven QSAR analysis (Şahin et al., 2011; Sahin and Saripinar, 2020; Sahin et al., 2021).Using this framework, a total of 204 molecular descriptors were computed using Spartan 07. To reduce redundancy and prevent overfitting, descriptor selection was optimized through a genetic algorithm (GA)-based procedure (Fernandez et al., 2011), and only statistically significant descriptors with low intercorrelation were retained for model construction. Subsequently, multiple ML algorithms, including artificial neural network, decision tree, and hybrid models, were evaluated to enhance prediction accuracy. Among all the tested models, the gradient boosting machine and random forest (GBM+RF) hybrid algorithm yielded the highest predictive performance, with an R2 value of 0.99978. To assess the robustness of the ML models, the training and validation procedures were repeated using different random seed initializations. The resulting performance metrics showed only minor variations across runs, indicating that the predictive performance of the GBM, RF, and GBM+RF hybrid models was not sensitive to random seed selection. The overall dataset comprised 54 pyrazole derivatives, with 50 molecules used for model construction and 4 reserved for validation. Although the high R2 value indicates strong internal consistency, it should be interpreted with caution due to the relatively small sample sizes for the construction of the model and the test subset. Overall, the integration of 4D-QSAR and ML approaches demonstrated strong predictive capability and effectively captured the key geometric and electronic features associated with biological activity. The electron conformational-GA computational strategy provides a robust framework for the rational design and virtual screening of pyrazole derivatives, leveraging multi-conformer modeling and a diverse set of molecular descriptors to identify potentially active compounds. The study was limited by the small size of the training and test set and the absence of experimental validation, which may constrain the generalizability of the findings. Future work could address these limitations by applying the model to larger and more diverse ligand dataset, performing virtual screening of compound libraries to discover novel hits, and validating promising candidates through in vitro assays. Overall, these findings support the potential of this scaffold for drug discovery, while further experimental studies are warranted to confirm the predicted activities and refine the predictive power of the model.

The design of organic chromophores with a high photoluminescent quantum yield (PLQY) is crucial for various optoelectronic applications. However, the vast chemical space of organic chromophores poses a significant challenge for experimental screening. Here, we report a molecular fingerprinting-based deep learning pipeline to discover organic chromophores with the desired PLQY. We convert 713 organic chromophores into 2048-bit fingerprints and screen them using machine learning (ML) techniques to predict their effect on PLQY. Support vector and gradient boosting regressor models achieve good predictive performance, with R 2 values ranging from 0.68 to 0.88. By breaking retrosynthetic analysis, we designed 5200 new organic chromophores with desirable PLQY. Furthermore, we visualize and screen 1840 chromophores using structure-activity landscape analysis. Our work demonstrates the power of molecular fingerprinting and ML in designing new chromophores with desired optical properties, providing a useful strategy for accelerating the discovery of high-performance organic materials.

The development of novel materials for radioactive iodate adsorption is critical for nuclear waste management. Layered double hydroxides (LDHs) are attractive iodate adsorbents because of their compositional flexibility and anion adsorption mechanisms. However, limited physicochemical understanding of LDHs synthesizability and adsorption mechanisms makes conventional trial-and-error approaches infeasible for exploring the numerous compositional spaces of multi-metal LDHs. In this study, a machine learning-assisted experimental approach is used to discover optimal multi-metal LDHs for iodate adsorption, leveraging its ability to discover hidden rules and predict unexplored compositional spaces. Active learning, based on positive-unlabeled and random forest models, was used to expand the exploration from an initial set of 24 binary and 96 ternary LDHs to 196 quaternary and 244 quinary candidates, requiring experimental trials for only 16 % of the total candidates. The discovered novel multi-metal LDH composition, Cu3(CrFeAl)1, exhibits an exceptional iodate adsorption capacity of 91.0 ± 0.2 %. The first application of Shapley Additive ExPlanations enhances model explainability, revealing that ionic size similarity is essential for synthesizability, whereas a higher electronegativity difference improves adsorption capacity. This study demonstrates, for the first time, the potential of machine learning-assisted discovery of multi-metal LDHs for radionuclide decontamination, paving the way for the accelerated development of new adsorbents to remediate hazardous materials in the environment.

Grid searching a large and high-dimensional chemical space with density functional theory (DFT) to discover new materials with desired properties is prohibitive due to the high computational cost. We propose an approach utilizing Bayesian optimization (BO) with an artificial neural network kernel to enable an efficient and low-cost guided search on the chemical space, avoiding costly brute-force grid search. This method leverages the BO algorithm, where the kernel neural network trained on a limited number of DFT results determines the most promising regions of the chemical space to explore in subsequent iterations. This approach aims to discover new materials with target properties while minimizing the number of DFT calculations required. To demonstrate the effectiveness of this method, we investigated 63 doped graphene quantum dots (GQDs) with sizes ranging from 1 to 2 nm to find the structure with the highest light absorption. Using time-dependent DFT (TDDFT) only 12 times, we achieved a significant reduction in computational cost, approximately 20% of what would be required for a full grid search. Considering that TDDFT calculations for a single GQD require about half a day of wall time on high-performance computing nodes, this reduction is substantial. Our approach can be generalized to the discovery of new drugs, chemicals, crystals, and alloys in high-dimensional and large chemical spaces, offering a scalable solution enabled by the neural network kernel.

Sulfanilamide (SN), a synthetic broad-spectrum antimicrobial that inhibits folic acid synthesis and suppresses bacterial growth. However, long-term use has caused allergic reactions, skin problems, crystalluria, nephrotoxicity, and other side effects. SN has developed resistance, and its associated side effects underscore the urgent need to discover safer alternatives with greater efficacy and reduced toxicity. In this study, we attempted to design new SN derivatives by incorporating various functional groups into their basic structure. Derivative structures were geometrically optimized utilizing density functional theory (DFT) and B3/LYP 6-31G+(d, p) basis set to calculate their physicochemical and spectrochemical properties. Molecular docking and molecular dynamics (MD) simulations were conducted against the dihydropteroate synthase (DHPS) protein (PDB ID: 1AJ2) to predict the binding affinities of analogs and stability at the active site. ADMET and PASS analyses evaluated toxicological and pharmacological profiles. Most of the derivatives showed lower energy gaps (5.14 eV to 5.30 eV) than SN (5.34 eV). All derivatives showed stronger binding affinities (-5.5 to -6.7 kcal mol-1) compared to SN (-5.4 kcal mol-1). ADMET results showed good pharmacokinetics, with some derivatives exhibiting higher GI absorption and most falling under toxicity class III. Overall, SN7 (-6.5 kcal/mol), SN17 (-6.6 kcal/mol), and SN18 (-6.7 kcal/mol) have exhibited better performance. Thus, our research reveals that the studied analogs can serve as novel alternatives to SN with superior quality. However, further experimental and biological studies are necessary to validate these theoretical findings and confirm their potential antibacterial efficacy. The online version contains supplementary material available at 10.1007/s40203-025-00526-y.

The current research aims to discover applications of QML approaches in realizing liabilities within smart contracts. These contracts are essential commodities of the blockchain interface and are also decisive in developing decentralized products. But liabilities in smart contracts could result in unfamiliar system failures. Presently, static detection tools are utilized to discover accountabilities. However, they could result in instances of false narratives due to their dependency on predefined rules. In addition, these policies can often be superseded, failing to generalize on new contracts. The detection of liabilities with ML approaches, correspondingly, has certain limitations with contract size due to storage and performance issues. Nevertheless, employing QML approaches could be beneficial as they do not necessitate any preconceived rules. They often learn from data attributes during the training process and are employed as alternatives to ML approaches in terms of storage and performance. The present study employs four QML approaches, namely, QNN, QSVM, VQC, and QRF, for discovering susceptibilities. Experimentation revealed that the QNN model surpasses other approaches in detecting liabilities, with a performance accuracy of 82.43%. To further validate its feasibility and performance, the model was assessed on a several-partition test dataset, i.e., SolidiFI data, and the outcomes remained consistent. Additionally, the performance of the model was statistically validated using McNemar's test.