AI for Science

Quadratic Unconstrained Boolean Optimization (QUBO) problems are widespread in both industrial applications and scientific studies. A QUBO problem corresponds to the optimization of a system of Ising spins defined on a generally sparse and heterogeneous graph. When the QUBO problem contains conflicting requests, the corresponding Ising system is frustrated, generating a complex energy landscape, which is hard to explore and optimize. Despite extensive algorithmic and hardware developments, finding low-energy configurations in these systems remains challenging (e.g., local-update heuristics typically become trapped in metastable states), especially when the (possibly frustrated) interactions generate extended correlated domains. We introduce CluMP (Cluster-based Message-Passing), an algorithm that performs collective updates on connected clusters of spins using information from Belief Propagation (BP). By controlling the amount of frustration within clusters, CluMP enables BP convergence on large subgraphs and proposes nonlocal rearrangements involving up to hundreds of spins in a single move. We benchmark CluMP against state-of-the-art local-update heuristics on spin-glass models defined on several graph topologies, including random regular graphs and lattice regular graphs in two and three dimensions. Cluster moves consistently bypass local trapping and reach lower energies with fewer effective operations than single-spin dynamics. These results demonstrate that frustration-tolerant cluster updates can be implemented efficiently on sparse graphs. The CluMP framework provides a scalable strategy for large-scale combinatorial optimization and inference problems, where exploiting medium- and long-range correlations is key to navigating complex energy landscapes.

Hyperspectral standoff detection systems such as Focal Plane Array (FPA) Fourier Transform Infrared (FTIR) spectrometers provide high spatial resolution in detecting airborne chemical contaminants that are invisible to the human eye but potentially hazardous. When two such systems are operated simultaneously with a suitable opening angle, they enable tomographic reconstruction of contaminant plumes with improved spatial and temporal accuracy. This work presents a mathematical model of these measurement capabilities and an algorithm to identify, localize, and quantify contaminant release sources. The objective is to develop a a tool that reconstructs release locations and predict the future plume evolution from standoff measurement data, thereby supporting early warning and situational awareness in hazardous material release scenarios. The transport of contaminants is modeled by an advection-diffusion equation, and the corresponding inverse problem for source identification is formulated accordingly. Owing to the severe ill-posedness and underdetermination of the problem, a sparsity-promoting regularization approach is employed together with a high-performance optimization algorithm. To incorporate the tomographic measurement data into the discrete formulation, a level-set description of a threshold concentration is used, allowing the measurements to be represented independently of the computational mesh and avoiding costly remeshing procedures.

We explore an extension to Newtonian gravity through a generalised Lagrangian function with the introduction of a second dynamical scalar field. Building on previous research into gravity with variable gravitational coupling, the work derives the complete field equations and applies a weak-field approximation. This leads to an effective post-Newtonian gravitational potential that includes key aspects of relativistic theories. The resulting N-body equations of motion highlight differences among inertial and gravitational masses, which can constrain the theory's free parameter through data from the Nordtvedt effect. By employing the method of osculating orbits for a two-body system, the study calculates the secular variation of the orbital pericenter and aligns this with the latest data on Mercury's perihelion shift, for another observational constraint on the model. Furthermore, a few examples of theories are discussed.

Coherent electron shuttling is a key requirement for realizing scalable silicon quantum computing architectures. However, in silicon qubits, the existence of nearly degenerate conduction-band valleys poses a significant challenge because non-adiabatic transitions to excited valley states cause spin dephasing via spin-valley mixing. In this paper, we propose a smooth velocity shuttling protocol to suppress these valley excitations. By mapping the time-domain design of the shuttling velocity profile onto the design problem of window functions in signal processing, we establish an analytical and intuitive design guideline that does not require computationally expensive numerical optimization. We demonstrate that the high-frequency sidelobes of the shuttling velocity spectrum can be effectively suppressed by applying a frequency-modulated gate voltage based on the Tukey window. Through statistical numerical simulations incorporating realistic spatial randomness of the valley landscape, we show that the proposed smooth velocity control significantly reduces the average spin infidelity in the moderate-to-low disorder regime ($|Δ_0|/σ_Δ\simeq \mathcal{O}(1)$). Our results underscore that this simple, control-level velocity shaping provides a robust pathway toward high-fidelity spin transport in large-scale silicon quantum processors.

Ciliary arrays pump fluid at low Reynolds number through non-reciprocal beating and phase coordination between neighbouring cilia. Previous studies have demonstrated that antiplectic metachronal waves are more effective than symplectic waves in enhancing transport, and have proposed several physically intuitive explanations for this preference. What remains incomplete is a predictive analytical understanding of how hydrodynamic coupling and beat geometry determine the flow-maximising phase difference. Here, we address this problem in two steps: we first use reinforcement learning to identify flow-maximising coordination in a bead--spring cilia model, and then introduce an analytically tractable reduced model, termed a tilted-slider model, to analyse the weak-coupling limit. Reinforcement learning identifies antiplectic coordination as the flow-maximising state in linear arrays, and shows that the phase difference between neighbouring cilia accounts for most of the flow enhancement. We then use the tilted-slider model to show that a shift of the time-averaged position opposite to the effective-stroke direction enhances fluid transport through its coupling with the elastic restoring force. The reduced model further reveals that antiplectic coordination can be optimal, consistent with previous studies, whereas symplectic coordination can instead become optimal depending on beat geometry. These results identify a simple elastohydrodynamic mechanism underlying flow-maximising metachronal coordination.

Many applications in quantum simulation, quantum chemistry, and quantum machine learning require not a single quantum state but an ensemble of states characterizing the heterogeneity of a target system. Preparing such ensembles state-by-state is prohibitive in both variational and fault-tolerant settings, thereby motivating a generative modeling approach. We introduce latent-conditioned parameterized quantum circuits (LPQCs), a hybrid quantum-classical framework in which classical neural networks map a latent variable sampled from a prior distribution to the parameters of a parameterized quantum circuit. We prove that LPQCs are universal approximators for probability measures over density operators in the 1-Wasserstein distance, extending classical universal approximation theorems to the quantum-distribution setting. We additionally introduce a multimodal latent prior and a mixture-of-experts circuit architecture, and show empirically that the latent-conditioned parameterization alleviates the barren plateau problem during optimization, a behavior for which we provide rigorous partial guarantees. Numerical experiments validate the framework on a synthetic multi-cluster ensemble of mixed quantum states and on a QM9-derived ensemble of 3-D molecular structures. In these tasks, LPQC outperforms recent quantum generative baselines and matches the generation quality of a classical neural-network baseline, while requiring an output dimension that grows only linearly with the number of qubits rather than exponentially. By leveraging classical expressivity in the latent space, LPQCs offer a tractable route to quantum generative modeling.

We propose KO-PDE-IDENT, a data-driven framework for identifying parsimonious partial differential equations (PDEs) with false discovery rate (FDR) control. PDE discovery from noisy observations is often hindered by extreme multicollinearity among candidate terms, which causes typical sparse-regression methods to select spurious terms. To address this problem, KO-PDE-IDENT initially mines a support set of potential candidate terms via model-X knockoff filters with finite-sample FDR control, then refines and ranks the surviving PDE alternatives. The framework integrates three components. First, knockoff feature statistics are constructed by coupling $\ell_{0}$-constrained adaptive best-subset selection with SHapley Additive exPlanations (SHAP), yielding an effective and computationally efficient difference statistic. Second, a recursive feature elimination (RFE) procedure removes terms whose marginal contributions are dispensable and assesses statistical necessity through knockoff-perturbed hypothesis testing. Third, the final model selection is formulated as a multi-criteria decision-making (MCDM) problem, where the optimal governing equation is the alternative that best balances a wide range of criteria such as predictive accuracy, model complexity and coefficient uncertainty. We evaluate KO-PDE-IDENT on five canonical PDEs under severe noise corruption. Empirical results show that our framework can exactly recover the true PDE structure, eliminating false discoveries while retaining all true underlying terms, with low coefficient estimation error.

The expectation value of the twist operator has been widely used as a polarization-based index for gapped and gapless phases in interacting quantum many-body systems. Although numerous studies support this usage in specific settings and rigorous results have established the validity of the criterion in important settings, the precise assumptions required for it to sharply distinguish gapped and gapless phases under more general conditions have not been fully clarified. In this work, we clarify the logical status of polarization-based indices by formulating symmetry-based statements under explicitly stated assumptions. We identify the role of ground-state degeneracy in the statements for gapped systems and clarify the distinct assumptions required to exclude gapless scenarios that could otherwise mimic gapped behavior in the thermodynamic limit. Building on this controlled framework, we construct a meaningful extension beyond one dimension, emphasizing that such an extension is nontrivial and cannot be obtained by a straightforward generalization of the one-dimensional twist operator. Our results delineate the regime in which polarization-based quantities are justified as sharply defined many-body indices.

Explicit dynamic finite element (FE) simulations are widely used for large deformation engineering analysis, but repeated simulations remain costly during design space exploration and optimisation. In explicit FE analysis, nodal kinematics and element level deformation measures evolve through coupled node element updates. This motivates graph learned simulators that approximate one step FE state transitions and roll them out autoregressively. However, many mesh based graph surrogates are node centred, which makes element level variables and native nodal elemental exchange less direct to represent. This work proposes CAttBiGNN, a cross attention based bipartite graph neural network for coupled nodal elemental learning. The graph represents FE mesh nodes and elements as distinct entities linked by directed node element edges, enabling nodal displacement increments and element level deformation states to be predicted on their native discretisation domains. An edge aware cross attention processor uses geometric edge embeddings to modulate directional node element message passing. For larger graphs, CAttBiUGNN combines the bipartite processor with graph downsampling and upsampling to improve long-range information propagation. The method is evaluated on dome shaped cold forming and corner shaped hot forming benchmarks. Comparisons with node centred baselines and bipartite and attention ablations show improved accuracy and balance in nodal displacement and elemental thinning prediction during autoregressive rollout. The results indicate that the proposed finite element inspired learned simulator can support manufacturability oriented field prediction and efficient design space exploration in large deformation sheet material forming.

Classical First-Passage-Time Distributions (FPTDs) have been extensively studied both theoretically and experimentally. Their quantum counterparts-Quantum First-Passage-Time Distributions (QFPTDs)-remain largely unexplored and have deep implications for both fundamental physics and the development of emerging quantum technologies. We measure the first QFPTDs using a motional mode of a single trapped ion. We develop a novel composite-phase laser pulse sequence to perform tunable stroboscopic single-shot projective measurements of the motional state of a trapped ion. We measure QFPTDs of the ion energy when coupled to electric-field noise. The measurement protocol developed here is broadly applicable to other quantum systems and provides a powerful method for exploring a broad range of QFPTD phenomena. With these results we open a new field of experimental investigations of QFPT processes with potential future relevance to quantum search algorithms, unraveling connections between classical and quantum dynamics, and study of the quantum measurement problem.

We show that a Hamiltonian in terms of the local real-space currents obeying an $\mathfrak{su}_1(2)$ affine Lie algebra eliminates the non-locality in the Hatsugai-Kohmoto model for a doped Mott insulator. We establish this local correspondence through the Bjorken-Johnson-Low prescription for anomalous commutators. With this result, we show that the charge susceptibility computed from the current Hamiltonian is identical to that with the elemental Fermionic fields. Consequently, the HK model is local in real space, though not in terms of the Fermionic fields, thereby eliminating the key criticism of this model and reinforcing the utility of current algebras for strong interactions.

Recently proposed by the author theory of the Q-balls mechanism of high-Tc superconductivity in cuprates is applied to explanation of known experimental data. The Q-balls (nontopological solitons) of coherently condensed spin/charge density wave fluctuations (SDW/CDW) with zero static mean and with the wave-vector that connects the 'nested' regions of the Fermi surface in doped cuprates cause pairing of the 'nested' fermions into local superconducting condensates. Hence, the Q-balls possess lower total energy in comparison with not condensed thermal SDW/CDW fluctuations in the same volume. Here it is demonstrated analytically that scattering of itinerant fermions on the Q-balls causes: linear temperature dependence of electrical resistivity in the interval of temperatures above T$_c$, reminiscent of the famous 'Plankian' behavior in the 'strange metal' phase; the famous hourglass dispersion close to forming Q-balls SDW fluctuations antiferromagnetic wave vectors and anomalous phonons dispersion softening close to CDW fluctuations wave vectors in the Brillouin zone. The diamagnetic response of Q-balls gas and contour plot of the Q-balls phase diagram, with lower temperatures dome touching the upper 'strange metal' one, are in qualitative accord with experimental data in high-T$_c$ cuprates. The superconducting condensates inside the Q-balls induce a spectral gap on the nested parts of the Fermi surface that might be responsible for a pseudogap phase in cuprates, where the Q-ball scenario was supported recently by micro X-ray diffraction data in HgBa$_2$CuO$_{4+y}$.

Ordinary, s-wave superconductors have been recognized as being topological phases of matter, in which the dynamical gauge field implies less understood global features. Using the tools of topological field theories and generalized symmetries, we provide an updated description of these systems. At very low energies, the Higgs model reduces to the BF theory, which exhibits topological order. Furthermore, the gauge field must be a spin$_c$ connection, to describe the spin of fermions forming Cooper pairs. Gauging implies that superconductors are inherently bosonic systems, yet they are endowed with a gravito-magnetic anomaly that is the remnant of their fermionic origin. We recognize that this anomaly is related to the Gaiotto-Kapustin-Thorngren bosonization, achieved via gauging fermion parity $(-1)^F$, now included in the gauge dynamics. This anomaly characterizes gauged electronic matter in great generality in three and four spacetime dimensions, forbidding trivial massive phases at low energy. It holds beyond the validity of the Higgs model, nd in other kinds of superconductors as well. It also appears in the nontrivial massless phase of three-dimensional electrodynamics, recently understood.

A low-energy nuclear transition in the isotope thorium-229 has been excited in thorium-doped crystals with laser light. This opens the perspective towards a highly stable and robust solid-state optical nuclear clock. The required laser radiation at 148 nm wavelength has so far been produced using pulsed laser systems where only a small fraction of the incident photons has been resonant with the narrow nuclear transition. Here we show that the nuclear resonance can be excited with a continuous-wave narrow-bandwidth laser source with a power of less than 1 nW, and that the resonance signal can be detected in absorption rather than in fluorescence. This eliminates the slow nuclear fluorescence decay from the detection process and offers a considerable advantage for clock operation through fast signal acquisition. The VUV laser source is based on three sequential frequency doublings, starting from a diode laser at 1187 nm that is well suited for linewidth narrowing and for frequency comparisons with optical atomic clocks. We use absorption spectroscopy for the quantitative characterization of two different Th-centers in calcium fluoride crystal and measure the isomeric shift between them. One of the centers shows a very small static electric crystal field gradient 0.1 V/$Å^2$, to be compared to gradients in the range of 100 V/$Å^2$ observed earlier. This indicates a center with high symmetry of the ions surrounding the Th nucleus, promising nuclear resonance lines that are nearly independent of the lattice spacing.

Surface growth driven by non-monomeric deposition has remained largely unexplored. We investigate a model based on the deposition of blobs with a power-law size distribution $P(s)\sim s^{-τ}$. We find that the critical exponents vary continuously with $τ$, recovering Kardar--Parisi--Zhang behavior only for $τ\ge 3$. For $τ<3$, roughness scaling exhibits strong corrections and scale invariance breaks down. We show that this behavior originates from the emergence of a second dynamical length scale $ζ$, corresponding to the linear size of the largest cluster, in addition to the usual correlation length $ξ$. The coexistence of these two relevant scales signals the breakdown of the usual Family--Vicsek scaling. These results point to a new phenomenology of surface growth beyond the standard scale-invariant paradigm.

Dynamic modulation of material properties in space and time enables powerful control over wave propagation, yet existing theories largely rely on idealized, nondispersive models. In realistic media, frequency dispersion can strongly reshape wave dynamics, especially near resonances in highly dispersive platforms such as epsilon-near-zero materials. Here, we develop a general frequency transition theory for electromagnetic scattering at moving interfaces between dispersive media. From phase continuity, we derive nonlinear frequency transition relations and show that dispersion fundamentally reshapes the space-time scattering landscape, enabling additional propagating solutions with no counterpart in nondispersive systems. Applied to Drude, Lorentz and double-Drude media, the theory reveals how resonant dispersion, material loss and negative-index branches reorganize the scattering channels. For the two-wave scattering class, we further introduce a mixed-domain formulation that combines time-domain interface kinematics with frequency-domain constitutive relations, yielding closed-form scattering coefficients. These results establish a unified framework for dispersive space-time scattering and open opportunities for dispersion-based transition engineering in realistic materials.

Lacking a universal law for granular flows across gravitational environments, fundamental processes such as hopper discharge remain vulnerable to failure in low gravity environments. A central challenge is clogging, the spontaneous arrest of flow through a constriction; yet previous studies report contradictory results on its dependence on gravitational acceleration. We identify the granular Bond number as the missing control parameter, defined as the ratio of intrinsic cohesive interactions among particles to gravity. Based on an in-bulk measurement of this quantity, we propose to rescale Earth-measured data for predicting granular behavior in low gravity. We present experiments of granular flow through an orifice under true reduced gravity (Moon and Mars), using an active drop tower, and extraterrestrial soil simulants as model cohesive materials. Our experiments reveal substantially increases in clogging probability, contrary to previously predicted, which depends on the properties of the material itself. When rescaled by the Bond number, seemingly conflicting results can be explained and collapse into a unified state diagram, predicting clogging across materials and gravitational accelerations. This establishes a general framework for the cohesion-to-gravity competition. Future space missions to the Moon, Mars, and asteroids will rely on such predictions of granular behavior in low gravity.

Given a graph \(G\) with adjacency matrix \(A\), consider the homogeneous Kuramoto energy $E_G(\boldsymbolθ):=\frac{1}{2}\sum_{1\leq i,j\leq n}A_{ij}\bigl(1-\cos(θ_i-θ_j)\bigr)$. We call \(G\) \emph{second-order globally synchronizing} if every second-order stationary point of \(E_G\) is fully synchronized. This property implies \emph{global synchronization}, namely that, up to a measure-zero set of initial conditions, trajectories of the Kuramoto model converge to a fully synchronized state. A fundamental graph-theoretic question is to identify which graph structures have this property. Existing guarantees for global synchronization typically require large minimum degree which forces the graph to be very dense, or good expansion properties. In this paper, we show that synchronization can also arise from a different, purely structural mechanism. More precisely, we prove that threshold graphs, a classical recursively defined graph class, are second-order globally synchronizing, and hence globally synchronizing. Thus, globally synchronizing graphs need not be very dense, have large minimum degree, or satisfy strong expansion-type conditions. The proof exploits the recursive construction of threshold graphs: local phasor constraints imposed by second-order stationarity are propagated along the construction sequence until full synchronization is forced.

Physics-informed neural networks (PINNs) solve time-dependent partial differential equations (PDEs) by learning a mesh-free, differentiable solution that can be evaluated anywhere in space and time. However, standard space-time PINNs take time as an input but reuse a single network with shared weights across all times, forcing the same features to represent markedly different dynamics. This coupling degrades error performance and can destabilize training when enforcing PDE, boundary, and initial constraints jointly. We propose Time-Induced Neural Networks (TINNs), a novel architecture that parameterizes the network weights as a learned function of time, allowing the effective spatial representation to evolve over time while maintaining shared structure. The resulting formulation naturally yields a nonlinear least-squares problem, which we optimize efficiently using a Levenberg-Marquardt method. Experiments on various time-dependent PDEs show up to 4 times improved relative error and 10 times faster convergence compared to PINNs and strong baselines.

We show almost sure wellposedness of mild solution to the cubic nonlinear wave equation in a weighted Besov space over $\mathbb R^2$. To achieve this, we show that any weak limit of $ϕ^4$ measures on increasing tori is invariant under the equation. We review and slightly simplify the periodic theory and the construction of the weak limit measure, and then use finite speed of propagation to reduce the infinite-volume case to the previous setup. Our argument also gives a weaker invariance result on the nonlinear Schrödinger equation in the same setting.

We generalize the augmented Roothaan-Hall (ARH) Hessian formalism to the self-consistent field (SCF) optimization of spin-restricted open-shell (RO) wavefunctions, encompassing high-spin, low-spin, and two-determinant electronic states. A detailed ARH formulation is presented. We demonstrate that ARH is a highly efficient optimization algorithm for rapidly identifying accurate SCF minima, primarily owing to its systematic construction of an effective Hessian, particularly in the case of Euclidean quadratic energy functions. The ARH is built upon a universal energy formulation, including grid-based integration, for spin-restricted closed-shell, spin-unrestricted and RO density functional theory (DFT), thereby unifying and simplifying their numerical implementation. The performance of the present method is evaluated using two benchmarking studies. First, for a series of iron-sulfur clusters exhibiting different spin states, which represent notoriously challenging SCF problems, the ARH algorithm demonstrates superior convergence efficiency relative to L-BFGS and truncated Newton methods, requiring much fewer RO-SCF iterations to achieve convergence. Second, the ARH approach avoids convergence to higher-energy stationary points in two-determinant RO-SCF calculations for singlet excited states of selected photoactive compounds. Finally, an application of the ARH-based RO-SCF is illustrated by an investigation of the mechanistic origin of the spin-crossover phenomenon in Ni(II)-porphyrin complex utilized as a contrast agent.

Hydrogen embrittlement (HE) has persisted for more than a century as one of the most intractable problems in materials science. The prevailing view1 that diffusive H governs embrittlement has fostered the widespread assumption that H trapping at crystal defects mitigates HE. Here we overturn this conventional paradigm. Using plasma/ion irradiation of tungsten, we decouple -- for the first time -- H-induced crack nucleation from subsequent cavity propagation, and reveal nucleation as a two-stage mechanochemical fracture instability enabled by trapped H in the absence of diffusive H. In the first stage, H accumulation to a critical occupancy at dislocation cores acts as a chemical fuse, collapsing the local cohesive strength to a threshold at which infinitesimal external loads can trigger atomic decohesion. This bond rupture instantaneously enables the second stage: confined recombination of atomic hydrogen into molecular form. The abrupt release of chemical energy within an atomically restricted volume generates a transient inflation pressure that drives a dynamic, brittle jump to an internal macroscopic cavity. By separating mechanical decohesion triggering from energetic crack driving, our results provide a deterministic framework for the onset of H-induced crack nucleation under low-stress conditions. Furthermore, we place experimentally the classical H-enhanced decohesion model on an atomistic foundation and elevate it from phenomenology to prediction. Finally, by shifting the focus from experimentally elusive diffusive H to directly measurable trapped H, this work reframes HE as a deterministic, quantifiable instability, establishing a new paradigm for understanding and mitigating H-induced failure in high-strength metals.

The celebrated analogy between the pressure-temperature phase diagram of a liquid-gas system and the field-temperature phase diagram of a ferromagnet has long been a cornerstone for understanding universality of phase transitions and critical phenomena. Here we extend this analogy to a highly frustrated antiferromagnet, the spiral Ising compound Nd$_3$BWO$_9$ with kagome layers. In its phase diagram, we identify a metamagnetic transition line with a critical endpoint (CEP) located at $μ_0H_{\mathrm{c}} \simeq 1.04$ T and $T_{\mathrm{c}} \simeq 0.3$ K. Above the CEP, an Ising supercritical regime emerges with crossover lines that follow a universal scaling law, as evidenced by the specific heat, magnetic susceptibility, and magnetocaloric measurements. Remarkably, we observe highly sensitive magnetic cooling near the emergent CEP, characterized by a divergent magnetic Grüneisen ratio $Γ_H \propto 1/t^{β+γ-1}$, with $β+ γ\simeq 1.563$ the sum of critical exponents of the 3D Ising universality class and $t \equiv (T-T_{\rm c})/T_{\rm c}$ the reduced temperature. Adiabatic demagnetization from 2 K and 4 T reaches a minimum temperature of 195 mK, via a self-cascading process that combines supercritical and topological cooling. Our findings open a new avenue for studying supercritical phenomena and magnetic refrigeration with the frustrated rare-earth compounds RE$_3$BWO$_9$ and, more broadly, in Ising-anisotropic antiferromagnets such as spin ices.

In twisted bilayer graphene, a unified understanding of the mechanisms governing temperature-dependent electronic spectra and thermodynamic properties remains controversial despite extensive theoretical efforts. Here, we present a comprehensive theoretical framework that quantitatively accounts for scanning tunneling spectroscopy, quantum twisting microscopy, and thermodynamic properties of magic angle twisted bilayer graphene. We demonstrate that the observed behavior arises from the interplay between electron correlations and external symmetry-breaking induced by strain and lattice relaxation. These effects act cooperatively to shape the emergent electronic behavior, leaving characteristic signatures across spectroscopy, compressibility and entropy.

Universal machine learned interatomic potentials (uMLIPs) embody a growing area of interest due to their transferability across the periodic table, displaying an error of about 0.6 kcal/mol against the Matbench Discovery test set. However, we show that achieving more accurate predictions on out-of-domain tasks requires fine-tuning. Additionally, we investigate the existence and influence of model biases in molecular dynamics (MD) by examining two approaches for data generation: from multiple MD trajectories in parallel, which we call naive fine-tuning, and from a single MD trajectory with fine-tuning after set intervals, which we call periodic fine-tuning. Our results find that naive fine-tuning generates constrained datasets that fail to represent MD simulations, and thus downstream fine-tuned models fail during extrapolation. In contrast, periodic fine-tuning yields models which are more generalizable and accurate, producing low-error dynamics. These findings indicate the role of uMLIP bias in fine-tuning, and highlights the need for multiple fine-tuning steps. Lastly, we relate unphysical behavior to principal component space, and quantify extrapolations through Q-residual analysis, which are useful as a proxy for epistemic uncertainty for larger simulations.

Understanding and controlling the energy level alignment at interfaces between lead halide perovskites and electron transport layers is crucial for optimizing perovskite-based devices such as solar cells. In this work, we investigated the energy level alignment of C60 on in-situ cleaved MAPbI3 single crystals in multiple repeat experiments using photoelectron spectroscopy aiming to resolve inconsistencies reported in earlier studies. Our results show that both materials remain chemically stable upon interface formation, in contrast to the strong reactions typically seen when metals are evaporated on perovskite surfaces. By analyzing the Pb 4f and C 1s binding energies in detail, we determined that C60 consistently exhibits a downward energy shift toward MAPbI3, which works against efficient charge extraction. The magnitude of this shift, however, varies between different sample positions, highlighting that small variations in sample surfaces can lead to significant differences in energetic alignment. At higher C60 coverages of more than 5 monolayers, a constant HOMO-valence band offset of 0.52 eV was obtained. These findings underscore the decisive role of surface chemistry on interfacial energetics, explain performance variability in perovskite devices, and demonstrate the need to control and accurately measure surface properties. Furthermore, the observed energetic alignment can explain why further interface modification by charge blocking layers or surface passivation is needed for optimized device efficiencies.

Land Surface Temperature (LST) is a key variable for various applications, such as urban climate and ecology studies. Yet, existing satellite-derived LST products provide either high spatial or high temporal resolution, resulting in a fundamental trade-off between the two. To address this trade-off, we combine observations from a geostationary and a polar orbiting satellite and provide LST fields at high spatial and high temporal resolution (1 km at 15-min intervals). We demonstrate their application for intraday forecasting of LSTs. To estimate LST fields at high spatiotemporal resolution, a U-Net model is trained to map LST fields from SEVIRI/MSG (3 km and 15 min resolution) to LST fields from Terra/Aqua MODIS (1 km, 4 overpasses per day) that are collocated in space and time. The presented model has been trained on LSTs across large European cities with a population exceeding 1 million inhabitants, and achieves an RMSE = $1.92$°C and near-zero bias MBE = $0.01$°C on the hold-out test set. As a second step, we present an LST nowcasting model based on ConvLSTM architecture, trained across downscaled LST fields with forecast lead times of 15 to 75 minutes. The nowcasting model outperforms a persistence and a Climatological Rolling Median benchmarks, with RMSEs of $0.57$ to $1.15$°C for the considered lead times and biases ranging from $-0.1$ to $0.14$°C. An additional validation conducted against independent MODIS overpasses confirms robust performance. Our LST forecast model at high spatiotemporal resolution is directly applicable to operational satellite-based LST monitoring.

Large language models can write scientific code, but direct paper-to-program translation remains fragile when correctness depends on tacit conventions rather than explicit equations. We frame this as a \textbf{knowledge-externalization} problem: index choices, gauges, fermionic signs, contraction order, validation gates, and scaling constraints must be made explicit before code generation. We evaluate a multi-stage, human-in-the-loop workflow on two quantum many-body tasks. DMRG from Schollwock's pedagogical review serves as calibration: specification-guided implementations pass in all 16 model pairings, compared with 6/13 direct attempts, and a prose-specification ablation shows that externalized content, not \LaTeX{} form, is the active ingredient. Pfaffian conversion of HFB states to MPS from the five-page Letter by Jin et al. serves as the stress test: no public implementation is available, and success depends on tacit sign, gauge, ordering, and scalability conventions. Here the workflow yields 11/26 audited passes, while direct prompting yields none. Cross-specification transfer is asymmetric: non-GPT specifications implemented by GPT~5.5 pass 4/4, whereas GPT~5.5 specifications implemented by weaker models fail 4/4. The contrast supports a two-bottleneck picture. Externalization resolves the first bottleneck -- paper-to-code ambiguity -- well enough to make DMRG reproducible and Pfaffian-MPS auditable. The remaining failures expose a second bottleneck in implementation-model capability. Iterative meta-specification moves this boundary but does not eliminate it. The resulting \emph{Paper-to-Program Many-Body} skill is both a reusable implementation protocol and a diagnostic instrument for AI-assisted many-body programming.

We prove that the Hirzebruch quadratic form of a complex hyperplane arrangement is non-positive on the set of stable weights, and we identify the zero locus within this set with flat logarithmic connections of a distinguished type. The proof uses Kempf--Ness and the frame-potential inequality.

Foundation models trained on DNA sequences have achieved strong performance across biological tasks including variant effect prediction and genome design. These models rely on massive public genomic datasets comprising trillions of nucleotide tokens. Unlike natural language, DNA sequences lack semantic transparency, making corrupted or adversarially crafted entries difficult to detect during data curation. We present the first systematic study of training data poisoning in genomic language models, targeting both pre-training and fine-tuning stages. At pre-training, using Evo 2 and GENERator architectures, we show that less than 1% adversarially crafted sequences in the training corpus can selectively degrade generative performance on targeted genomic contexts while leaving unrelated sequences unaffected. We evaluate three scenarios: corruption of TATA-box promoter motifs, disruption of CTCF binding sites, and insertion of synthetic sequences absent from all training genomes. At fine-tuning, we demonstrate two additional attacks. First, poisoning a subset of CTCF sites in a ClinVar-derived corpus installs a conditional backdoor in a LoRA-adapted model that activates almost exclusively when the trigger sequence is present. Second, using frozen Evo 2 7B embeddings, targeted label corruption of downstream training data selectively compromises a clinically relevant variant classification task, demonstrated on BRCA1 variant effect prediction. These results show genomic foundation models are susceptible to targeted data poisoning with minimal footprint. We urge the field to adopt data provenance tracking, integrity verification, and adversarial robustness evaluation as standard components of the genomic model development pipeline.