Papers

The origin of supermassive black holes (SMBHs) is still debated, with hypotheses ranging from massive seeds to super-Eddington growth to chaotic accretion. Recent discoveries of very high redshift SMBHs constrain their possible origin and growth scenarios. We investigate whether the chaotic accretion scenario can explain the most extreme high-redshift SMBH masses. We predicted the exponential growth timescale of black holes accreting chaotically in a slowly-spinning dark matter halo, in which uncorrelated gas inflows maintain a low spin and hence a low radiative efficiency. We compared this prediction against the growth timescales and seed masses implied by a compilation of all known SMBH mass estimates at $z>7$ and against the feeding and feedback conditions expected in such haloes. We found that in the period $7.6 < z < 10.6$, the most massive SMBHs likely grew very efficiently, with an exponential growth timescale $\tau \sim 32$~Myr, which is consistent with chaotic accretion predictions provided that growth was almost continuous. All high-redshift SMBHs with virial mass estimates could have grown from seeds with $M_{\rm seed} \lesssim 100 \, \msun$. We describe the evolution of the mass reservoir feeding such a black hole, suggesting that wind-driven feedback eventually removes most of the gas and changes the growth mode of black holes at $z \sim 7.5$. We make several predictions regarding the spins, host-galaxy correlations, and observable counterparts of high-redshift SMBHs that distinguish our model from its alternatives. These predictions will be testable in the near future. Chaotic accretion from stellar-mass seeds remains a viable origin for the earliest known SMBHs, without invoking massive seeds or super-Eddington growth.

Visual disruptions can arise while a robot is executing a task, leaving a vision-language-action (VLA) policy to respond without knowing the disruption type or timing. We introduce Self-supervised Adaptation from Leftover Trajectories (SALT), which uses the leftover trajectory, the unexecuted part of the previous action chunk, as self-supervision for test-time adaptation. Because consecutive chunks overlap in time, the leftover provides a temporally aligned target for the current prediction over the same future control interval. At the onset of a visual shift, the leftover can retain a plan formed before the corruption, so updating the policy toward it anchors the adaptation across the shift (Transition Anchoring). SALT keeps the adapted policy and regenerates the current chunk, whose leftover becomes the target at the next replan, carrying the correction forward along the execution trajectory (Sequential Correction Propagation). Supervision comes entirely from the policy's own predictions, requiring no disruption annotations, expert actions, or target-domain demonstrations, and a lightweight adaptation gate calibrated only on nominal trajectories decides when updates begin. On LIBERO-10, SALT increases average success across five persistent visual corruptions from 43.9% to 53.2% with SmolVLA and from 58.7% to 66.0% with GR00T N1.7, while largely preserving nominal performance. On a real robot, it raises task progress averaged over digital and physical disruptions from 0.49 to 0.61.

We introduce a unified framework of $Q\text{--}P$ spaces, designed to encompass a broad class of generalized topological structures arising from the interaction of two topological properties. We establish general permanence and transfer results, including heredity, productivity, additivity, and separation properties. Specializing to the Hausdorff setting, we investigate the intermediate classes \(C\)-Hausdorff, \(PS\)-Hausdorff, and \(CC\)-Hausdorff, and obtain the hierarchy \[ T_2 \Longrightarrow Ps\text{-}Hausdorff \Longrightarrow CC\text{-}Hausdorff \Longrightarrow C\text{-}Hausdorff \Longrightarrow T_1. \] For \(C\)-Hausdorff spaces, we derive intrinsic characterizations and establish the equivalence with the simultaneous properties of \(k\)-Hausdorffness and \(KC\). We further examine their connections with \(k\)-spaces, Fréchet spaces, local compactness, realcompactness, and \(\aleph_0\)-boundedness. Finally, carefully constructed examples and counterexamples demonstrate the strictness of the hierarchy and the necessity of the hypothesis. Thus, the \(Q\!-\!P\) framework provides a unified perspective on generalized separation axioms and reveals a rich structure between classical Hausdorffness and \(T_1\)-separation.

How a mobile impurity exchanges momentum and energy with a many-body environment is a central question in nonequilibrium quantum physics. This exchange becomes particularly complex when the impurity moves rapidly and is strongly coupled to its bath. Here, we realize an interaction-tunable cold-atom collider with fermionic $^{40}$K impurities immersed in a $^{23}$Na Bose–Einstein condensate. A species-selective Raman pulse simultaneously launches the condensate and quenches the interspecies scattering length, producing an initial relative speed 36 times the condensate speed of sound. We track the ensuing relative motion as the interspecies interaction is tuned from weak coupling to resonance. At weak and intermediate coupling, the impurity dynamics are well described by a finite-energy two-body collision model. Near resonance, however, the early-time impurity acceleration is markedly suppressed relative to the two-body prediction, corresponding to a strong enhancement of the apparent dynamical inertia. The enhancement is observed near two distinct Feshbach resonances. These observations provide evidence for a polaronic response in the strongly coupled, supersonic regime of a degenerate Bose–Fermi mixture.

For a finite forest $F$ let $i_k(F)$ be the number of independent sets of $F$ with $k$ vertices. Zhang and Li proved that the sequence $i_0(F),i_1(F),\dots,i_{\alpha(F)}(F)$ is unimodal for every finite forest $F$, which answers Erdős Problem 993. We give a second proof. It starts from their decomposition relative to a fixed independent set and from the bounds of Zhang and Li and of Fang, Lu, Nevo, Yao and Zheng that confine a valley of the sequence to an explicit window of ranks. For a forest with at least $25$ vertices, one moment argument excludes a valley at every rank of the window: at the activity where the hard-core mean equals the rank, the size of a random independent set is a mixture of binomial laws over an independent set of maximum weight, a valley is a moment inequality for this mixture, and it is excluded by duality given three bounds that hold for every forest, on the variance of the number of free vertices and on its Laplace transforms, and on the variance ratio. The variance bound is proved by hand up to finitely many interval checks and the other two bounds are verified by computer on finite interval-arithmetic coverings; on the resulting parameter domain a valley is excluded by exact tests on finitely many rational boxes while the mean number of free vertices is below an explicit starting mean between $19$ and $50$, and above it by one inequality, with explicit constants, for the fibers of a weighted valley kernel, proved by hand up to a finite list of explicit checks and averaged over the mixture. Forests with at most $24$ vertices are treated by exact counting, by hand except for exact rational evaluations of two explicit formulas at $43$ parameter triples. No forest is enumerated. A formal proof of the theorem in Lean 4 accompanies the paper.

Accurate modeling of p-type dopants in 4H-SiC is essential for understanding the mechanisms governing doping efficiency and carrier transport. In this work, we revisit the electronic structure of boron- and aluminum-related acceptors using hybrid density functional methods. Besides defect formation energies and thermodynamic transition levels, we present a quantitative look into the carrier capture kinetics within the multi-phonon emission framework. Our results reveal striking differences between the two most relevant p-type dopants. While $\textrm{B}_{\textrm{Si}}$ and $\textrm{B}_{\textrm{C}}$ exhibit nearly identical formation energies, consistent with the occurrence of both defects, the formation energy of $\textrm{Al}_{\textrm{C}}$ under intrinsic conditions is approximately 6.5 eV higher than that of $\textrm{Al}_{\textrm{Si}}$, confirming previous findings that the former is unlikely to occur. We further find that $\textrm{B}_{\textrm{Si}}$ possesses large electron and hole capture cross sections, identifying it as a plausible source of minority-carrier lifetime degradation of n-type material contaminated with boron. In addition, we predict a previously unexplored donor transition for $\textrm{B}_{\textrm{C}}$. If experimentally confirmed, this defect would represent additional problems to both p-type doped and boron contaminated 4H-SiC, not only because of its ineffectiveness as an electric dopant, but also due to trapping of up to two free holes, reducing the free-hole concentration and increasing scattering effects.

Physically-based differentiable rendering enables gradient-based optimization of scene parameters by matching rendered images to measurements, but has so far mainly focused on light transport. We extend this paradigm to medical ultrasound, where image formation resembles transient rendering: echoes are binned by time-of-flight rather than projected onto an image plane. We present UltraDiff, a modular framework for differentiable ultrasound ray tracing. UltraDiff formulates ultrasound image formation as a path-space integral, gated by travel time between the transducer and tissue interfaces, and derives a Monte Carlo estimator of both the forward model and its gradients with respect to scene parameters. We demonstrate this on an inverse geometry estimation: starting from a sphere, an SDF is optimized until simulated echoes match measured ones, recovering vertebral surfaces from simulated B-mode sweeps and from a real robotic acquisition of a spine phantom. Unlike state-of-the-art ultrasound shape reconstruction methods, which rely on pre-segmented images, our approach operates unsupervised on B-mode images through analysis-by-synthesis, while achieving competitive geometric accuracy. Implemented on top of Mitsuba 3, UltraDiff brings differentiable path tracing to a new sensing modality and provides a foundation for inverse problems in acoustic imaging.

Looped language models apply the same stack of layers T times to each token, which deepens the model without adding parameters but multiplies its key-value (KV) cache by T. The larger cache limits how many sequences a GPU can decode at once and slows each decoding step, which reads the whole cache. We propose Hybrid Latent Attention (HLA), which keeps exact keys and values within a sliding window of W recent tokens and stores each older token as a compact latent that the query of each loop reads directly, without reconstructing keys and values. We uptrain HLA on Ouro looped models (T=4) with 1.4B and 2.6B parameters, keeping the pretrained weights frozen and training only the added parameters to reproduce the original attention. The cache shrinks by 10.7x per token, fitting 4.0-8.8x as many concurrent sequences per GPU, and decoding throughput improves by 2.5x at 1K-token contexts and by up to 7.4x at 16K. HLA retains over 97% of the original accuracy on math, knowledge and reasoning benchmarks, and 96-100% on long-context retrieval up to 16K tokens. After supervised fine-tuning, it performs on par with the fine-tuned original model on competition-level math.

Due to rapid advances in Generative AI, commercial video generation tools can be used to produce fabricated surveillance footage that can fool both human viewers and automated synthetic video detectors. Since these tools are so widely accessible, a malicious user can create a harmful video clip at minimal cost. The production and dissemination of such videos in high-stakes settings, such as crime reporting and elections, can misdirect emergency response efforts or distort political discourse. Existing deepfake video datasets, used by the research community to develop deepfake detection algorithms, exhibit two limitations: (1) they emphasize benign web content rather than footage of possibly malicious activity, and (2) they rely on older or open-source generators that do not represent recent advances in generative systems. We assemble CCtv DeepFakes (CCDF), a video deepfake dataset, to address both gaps. CCDF contains 1840 videos (460 real and 1380 generated) spanning 16 crime and accident categories, with generated content produced using three leading commercial systems: Grok Imagine, Google VEO 3.1, and OpenAI Sora 2. CCDF is a highly realistic, small-scale, manually annotated dataset targeting evaluation of detection models. We release three versions of the dataset: the raw generated data, a cleaned version in which video metadata are standardized between real and synthetic samples to prevent detectors from exploiting trivial cues, and an altered version simulating low-effort post-processing attacks. We evaluate CCDF with ten recent state-of-the-art detectors covering different detection approaches. Our results suggest that these approaches do not reliably distinguish CCDF's generated videos from real ones, despite their strong reported performance on existing datasets. These results further confirm the benefits of the CCDF dataset for advancing research in DeepFake detection.

Implicit-process priors specify distributions over functions through sample-forward mechanisms such as Bayesian neural networks and stochastic simulators, but their function-space densities are typically unavailable. We introduce Generalized Matheron Variational Implicit Processes (GMVIP), a pathwise variational family for posterior inference with such priors. For Gaussian-process priors, GMVIP recovers the standard inducing-variable variational GP construction; for general implicit priors, its empirical covariance construction preserves the prior mean and covariance in the population limit. GMVIP constructs posterior samples by drawing a function from the prior and applying a correction anchored at a set of inducing inputs. The effect of this correction away from the inducing inputs is determined directly from prior samples, allowing the posterior to retain the structure and variability of the original implicit process. The (surrogate) prior and variational posterior use the same pathwise construction and differ only in the distribution of whitened inducing coefficients, yielding a tractable coefficient-space Kullback-Leibler divergence. Experiments on regression, classification, and forecasting with simulator-defined and retrieval-conditioned empirical trajectory priors show that GMVIP is broadly competitive with existing methods.

Redpine Science gives models and agents a single access point to a wide range of peer-reviewed literature, queried directly through the Model Context Protocol (MCP) and an API. This report evaluates Redpine Science on two levels: the relevance of the retrieved chunks, and a model's answer when it has access to Redpine Science compared to web search. Both public and expert-validated benchmarks are used. Public benchmarks are a widely accepted way to test model development and are comparable across labs, but risk saturation and memorization. To address this, we complement them with an expert-validated question set. In total, this report presents four evaluations. On ScholarQABench SciFact, the public answer-quality benchmark reported here, an agent with Redpine Science answers 94.4% of claims correctly against 87.6% with no retrieval. On the expert-validated question set, an agent with Redpine Science states 80.1% of the required claims against 70.2% for an agent restricted to web search. On the 668 queries of a public retrieval benchmark whose gold paper Redpine holds, stripped of any model reasoning, Redpine Science places the correct source paper in its top ten results for 83.1% of queries (Recall@10), against 79.3% for the benchmark's creator. A blinded expert relevance panel places Redpine Science's Precision@5 at 75.2% against 39.8% for the PubMed search tool. We release the expert-validated question set and instructions to reproduce every headline result above, at https://github.com/redpine-ai/benchmarks.

Systematicity, the probabilistic mapping of form to meaning, permeates language at all levels, and sublexical cues have been shown to govern human pseudoword processing. Yet whether LLMs exhibit comparable sensitivity to these cues remains unclear. We tested five LLMs on two Italian two-alternative forced-choice pseudoword experiments and compared their responses with a human behavioural baseline. LLMs aligned more reliably with humans when real-word options provided a lexical familiarity cue than in the pseudoword-only condition, where they fell substantially below fastText, a character-n-gram model. In addition, the sublexical cosine-similarity cue that reliably drove human–fastText agreement did not consistently transfer to human–LLM alignment, and reasoning-token expenditure bore no consistent relation to human processing difficulty. These findings suggest that LLMs do not necessarily share the sublexical cues that govern human pseudoword processing; we discuss tokenization and training-data coverage as candidate explanations.

LLM agents are increasingly capable of executing complex tasks and of recursively improving themselves on easy-to-verify objectives such as software engineering and mathematics. Since alignment is much harder to verify, this creates a growing risk of capabilities increasing without appropriate safety alignment, especially as capabilities expand to auto-research and cybersecurity. Existing approaches focus on capability self-improvement using verifiable feedback or on alignment training with supervision from stronger models or curated data, creating an external supervision bottleneck for alignment. We ask whether current models can improve their own safety alignment, and propose SIGMA, a data generation and training pipeline enabling alignment self-improvement that generalizes to out-of-distribution settings. Given only a "Model Spec" stating the model's desired behavior, SIGMA leverages a model's reasoning capabilities to strengthen its own safety reasoning. SIGMA first performs spec-guided task synthesis, using the candidate model as a task designer agent to generate diverse alignment dilemma scenarios and convert them into training tasks that stress-test its understanding of the Model Spec. Next, SIGMA conducts self-judged alignment training through supervised fine-tuning and rubric-based reinforcement learning with the model itself as the reward model. Despite training only on single-turn chat data, SIGMA improves safety alignment in multi-turn agentic environments (AgentHarm harmfulness decreases from 22.6 to 14.8; Agentic Misalignment decreases from 79.1 to 3.8), outperforms Deliberative Alignment and Constitutional AI baselines, and retains general capability. Analyses show that a Model Spec balancing harmlessness and helpfulness, test-time reasoning for safety deliberation, and high-quality rubrics from SIGMA's task designer agent are crucial for effective self-improvement.

Black holes are usually described either as classical geometries on which quantum fields propagate or as objects whose complete quantum description lies beyond semiclassical gravity. Hawking radiation already shows, however, that a black hole is not an exactly isolated system: the horizon carries a thermal scale and, after coarse graining, may be associated with fluctuations, dissipation and environmental correlations. Whether this horizon-scale openness can be encoded directly in the gravitational background, before any stochastic treatment of the metric, remains an open question. Here we answer it by introducing, for the first time, a conserved Hawking-scale anisotropic source representing the coarse-grained near-horizon atmosphere and solving the Einstein equations for a static black hole at fixed ADM mass. The dressed geometry has a renormalized mass function, a self-consistently shifted horizon and a modified surface gravity. Because the same lapse governs the horizon and timelike geodesics, the thermal response is accompanied by a correlated orbital restructuring: the effective potential is renormalized, the ISCO moves inward and the near-horizon residence time increases. The construction adds no Langevin force to a fixed Schwarzschild background; rather, it defines a deterministic dressed mean geometry, providing a controlled background on which stochastic metric fluctuations may subsequently be treated in the Einstein-Langevin framework. This establishes the deterministic mean-field sector as the necessary first step for future stochastic gravity programs addressing horizon-scale openness.

We prove that, for all $x > 0$, the product of modified Bessel functions satisfies \[ \frac{1}{2\sqrt{x^2+\nu^2+1/5}}<I_\nu(x)K_\nu(x)<\frac{1}{2\sqrt{x^2+\nu^2-1/4}} , \] where the lower bound holds for $\nu>-1$ and the upper bound for $\nu\ge 1/2$, and that the sum of the logarithmic derivatives satisfies \[ -\frac{x}{x^2+\nu^2-1}<\frac{I_\nu'(x)}{I_\nu(x)}+\frac{K_\nu'(x)}{K_\nu(x)}<-\frac{x}{x^2+\nu^2+7/20}, \] where the upper bound holds for $\nu>-1$ and the lower bound for $\nu\ge 1$. All four constants are best possible. The four proofs share the same elementary structure: in each case the relevant auxiliary function satisfies a second-order differential inequality, and the inequality follows from the strong maximum principle. Among other applications, we obtain sharp monotonicity properties of the product and improved bounds for the ratios $I_{\nu-1}(x)/I_\nu(x)$ and $K_{\nu+1}(x)/K_\nu(x)$.

Given a topological space $X$ and a family of subgroups $\mathcal{F}$ of $\pi_1(X)$, the $\mathcal{F}$-category of $X$ is given as one less than the cardinality of the smallest cover of $X$ by open subsets whose fundamental groups lie in $\mathcal{F}$. In this article we study $\mathcal{F}$-categories of aspherical spaces and obtain cohomological lower bounds, a maximality result and bounded cohomology classes whose vanishing properties are crucial for determining $\mathcal{F}$-categories. For this purpose, we study the bounded Adamson cohomology of a family of subgroups and discuss universality properties of bounded cohomology classes. We apply our techniques to derive some applications to the study of the monotonicity of $\mathcal{F}$-categories under degree-one maps between manifolds and to provide a counterexample to a question of Capovilla, Löh and Moraschini.

Rewriting a hot path from Python into Rust is a common way to speed up security analytics, and large speedups are routinely reported. A rewrite usually changes the language and the algorithm at once, so a single factor can credit the language with a gain that comes from a better algorithm. We study this on a sliding window threat scorer modelled on the traffic light calculator of the SentinelSphere platform. Five implementations, three in Python and two in Rust, produce bit identical scores, confirmed by a shared checksum, and we time them from 100 to one million events in a bounded and a burst regime. The language alone contributes between about 4 and 29 times. Replacing a per event rescan of the window by an incremental update contributes more than 11,000 times in the burst regime, so the end to end factor reaches about 195,000 times when the window keeps filling and levels off near 13,000 times when it does not. A power law fit to the timings reported for the original rewrite gives growth exponents of 1.85 for Python and 0.88 for Rust, the signature of an algorithmic difference. Performance claims for security tooling should therefore report the language and algorithm contributions separately.

Reliable operation of plasma processes requires a power supply that can regulate and limit the arc current despite rapid changes of the electrical load. Gas-flow fluctuations, arc motion, changing attachment conditions, and electrode interactions can cause abrupt arc-voltage variations and may lead to arc interruption. The power supply must therefore provide fast current control and a suitable actuation layer for supervisory stabilization. This paper presents an adaptive average current-mode controller for a buck converter supplying a nonlinear plasma arc load. Based on an averaged model of the converter inductor dynamics, the controller combines current-error feedback with feedforward compensation of the measured arc voltage. Online adaptation of the inductance and its effective series resistance compensates for converter-parameter uncertainty without requiring an online nonlinear arc model. A Lyapunov analysis establishes bounded estimation errors and asymptotic current tracking for the averaged closed-loop system under bounded-reference assumptions. Experiments with an argon plasma arc demonstrate contact-opening ignition and current-reference tracking under varying arc- voltage conditions, supporting the controller as a fast current-control layer for plasma applications.

Let $h_n$ denote the complete homogeneous symmetric function of degree $n$, $s_\nu$ the Schur function indexed by a partition $\nu$, and $\langle\cdot,\cdot\rangle$ the Hall inner product. The plethysm coefficient $\langle h_n[h_m], s_\nu\rangle$ is the multiplicity of the Weyl module $W_\nu(\mathbb{C}^N)$ in $\mathrm{Sym}^n(\mathrm{Sym}^m(\mathbb{C}^N))$ whenever $\nu$ has at most $N$ parts. For a partition $\lambda$, let $\lambda[mn]=(mn-|\lambda|,\lambda_1,\lambda_2,\dots)$. Bowman and Paget proved that $\langle h_n[h_m], s_{\lambda[mn]}\rangle$ is independent of $m$ and $n$ once $m,n\ge|\lambda|$. We prove that this coefficient is already constant for $m\ge\lambda_1$ and $n\ge\lfloor(|\lambda|+\lambda_1)/2\rfloor$, and that its stable value equals $\langle s_\lambda, H[H-1-h_1]\rangle$, where $H=\sum_{i\ge0}h_i$; the proof uses the skewing operators adjoint to multiplication by complete homogeneous symmetric functions. We give a short new proof of the Bowman-Paget theorem based on stability of vector-partition functions, which count the ways of writing a vector of non-negative integers as an unordered sum of non-zero such vectors (its parts). Using a Vandermonde-type determinantal identity, we deduce that the stable value is a signed sum of numbers of vector partitions all of whose parts have entries summing to at least 2. The same method gives a short proof of Brion's stability theorem. We show that the stable value equals the multiplicity of $W_\lambda(\mathbb{C}^k)$, where $k=\lfloor|\lambda|/2\rfloor$, in an explicit representation of $GL_k(\mathbb{C})$ independent of $m$ and $n$, obtained by modifying the polynomial induction of Narayanan, Paul, Prasad and Srivastava. Finally, we prove an explicit positive formula for the stable value, which counts the number of lattice points in a polytope, when $\lambda_1\le 4$.

Dynamic multimodal learning aims to learn robust representations by adaptively modeling information discrepancies across modalities. However, existing methods still suffer from two limitations: (i) static cross-modal alignment strategies usually impose uniform constraints on all samples while overlooking sample-wise variations, potentially leading to unreasonable over-alignment; and (ii) confidence- or uncertainty-aware fusion methods often fail to adequately account for feature magnitude and confidence differences across modalities. For modality pairs with significant feature magnitude differences or small confidence gaps, it might be unreliable to strictly align fusion weights according to confidence. To address these issues, we propose an Alignment- and Calibration-driven Multimodal Learning framework (ACML). Specifically, ACML incorporates a dynamic cross-modal triplet alignment module, which enforces strong semantic consistency for high-confidence positive pairs while encouraging diverse representation learning between high- and low-confidence positive pairs according to their confidence gaps. Additionally, ACML introduces a difference-aware attention calibration strategy that adaptively adjusts attention regularization based on feature magnitude and confidence differences across modalities, thereby mitigating biases caused by unreasonable fusion constraints. Extensive experiments on multiple multimodal benchmark datasets demonstrate that ACML consistently achieves superior performance and robustness over recent state-of-the-art methods.

Two graphs $G$ and $H$ are homomorphism indistinguishable over a graph class $\mathcal{F}$ if, for every graph $F \in \mathcal{F}$, the number of homomorphisms from $F$ to $G$ is equal to the number of homomorphisms from $F$ to $H$. Lovász (Acta Mathematica Academiae Scientiarum Hungarica, 1967) showed that two graphs are isomorphic if, and only if, they are homomorphism indistinguishable over all graphs. Subsequently, homomorphism indistinguishability relations of a long list of natural graph classes have been equated with natural graph isomorphism relaxations. Given the wealth of such results, Atserias, Kolaitis, & Wu (LICS 2021) asked for an axiomatic characterisation of homomorphism indistinguishability relations. By exhibiting topological and geometric structure associated with homomorphism indistinguishability, we derive such an axiomatic characterisation. Here, a central ingredient is a novel characterisation of graph parameters of the form $\hom(F, \star)$ for some graph $F$ alternative to a previous result of Lovász & Schrijver (JCTA 2010). Moreover, we investigate the topology of homomorphism indistinguishability and discuss repercussions for the Ulam–Kelly Reconstruction Conjecture.

During epidemics, limited resources dictate critical decisions about whom to isolate or monitor. Conventional public health strategies intuitively prioritize known infected individuals and their immediate contacts. Here, we reveal that this contact-centric approach is fundamentally suboptimal under realistic resource constraints. By formulating this challenge as target-aware network dismantling, we uncover a counter-intuitive, budget-dependent intervention transition. Generating-function analysis demonstrates that when intervention budgets are scarce, isolating structurally central individuals, rather than the immediate neighbours of the infected, protects a vastly larger population. The conventional neighbour-targeting strategy only becomes optimal as available resources increase. Because this critical transition point shifts dynamically with the intervention budget, network topology, and spatial distribution of infections, static strategies inevitably fail. To overcome this, we develop Risk-Aware Isolation Learning (RAIL), a deep reinforcement learning framework that adaptively selects optimal interventions based on the evolving residual network. Across diverse synthetic and real-world networks, RAIL consistently maximizes population protection and suppresses outbreak spread. Ultimately, our findings challenge prevailing epidemiological intuitions, establishing a generalizable framework for adaptive interventions against localized risks in public health and other complex systems.

Partially Relevant Video Retrieval (PRVR) aims to retrieve untrimmed videos containing moments relevant to a given text query. Despite recent progress, existing PRVR methods suffer from two key limitations: a fixed video decomposition scheme that causes semantic dilution, and source-domain overfitting induced by task-specific training. In this paper, we propose TF-PRVR, the first training-free framework for PRVR. TF-PRVR leverages frozen vision-language features to construct video-specific hierarchical representations. It derives temporal semantic signals from frame-level features and applies frequency-based multi-scale analysis to identify adaptive temporal boundaries, producing hierarchical segments with coherent event-level semantics. Built on these segments, TF-PRVR constructs a unified multi-scale graph and propagates query relevance across temporally and semantically related nodes. A moment-aware scoring strategy then aggregates temporally aligned relevance across scales, emphasizing consistently supported moments while suppressing isolated false responses. Without task-specific training, TF-PRVR preserves the general-purpose alignment capability of pre-trained vision-language models and avoids dataset-specific overfitting. Extensive experiments demonstrate consistent performance across datasets with diverse visual and temporal characteristics, suggesting a practical direction for training-free PRVR.

A graph is beautiful if each of its induced subgraphs is the intersection graph of all maximal nonempty 1-rectangles of a binary matrix, with adjacency defined by a common cell. Beautiful graphs were introduced as a hereditary class of Berge graphs, but a complete forbidden-induced-subgraph characterization was not obtained. We prove that a finite graph is beautiful if and only if it has no induced $C_4$, gem, net, watch, or odd hole. More precisely, these graphs are exactly the $C_4$-free comparability graphs admitting a partial order in which every interval is a chain. The proof constructs such an order from inclusion-maximal closed neighbourhoods: their representatives induce a bipartite graph whose domination regions admit compatible orientations. One order matrix then represents the graph and, through its principal submatrices, every induced subgraph. The representation step is formulated using classical double-bound graphs and the established correspondence between maximal bicliques and interval-intersection-closed posets. Consequences include polynomial-time recognition, the exact minimal obstruction families, and a corrected characterization in the $K_4$-free case.

This article demonstrates the classical equivalence between standard and conformal cosmological models regarding the apparent luminosity of type Ia supernovae. We show that previous claims of observational discrepancies stem from differing interpretations of physical quantities when transitioning between Einstein and Jordan frames. Specifically, we establish that in the conformal picture, SNe Ia are not "standard candles" in the conventional sense because their absolute luminosity depends on redshift due to variations in local measurement units.

World-action models (WAMs) couple future prediction with robot control, yet existing systems often vary the video backbone, interaction structure, supervision, and inference procedure simultaneously, making their design choices difficult to compare. We introduce OPENWAM, an open world-action modeling framework built around a common causal robot-video foundation and configurable video-action interaction. Starting from Wan2.2-5B, we perform causal robot-video pretraining on over 10,000 hours of video, then integrate an action expert through a shared Mixture-of-Transformers architecture that supports joint, video-then-action, action-then-video, and decoupled generation. OPENWAM achieves high success rates on four LIBERO suites and real-world bimanual tasks; robot-video training with causal adaptation improves VTA success on LIBERO-Long from 68.4% to 97.8%. The same configurable architecture naturally extends to inverse and forward dynamics, allowing us to study how counterfactual transitions improve independently trained dynamics models beyond demonstrations alone. When only the video predictor is adapted to a new task, a frozen local-context inverse dynamics model trained on counterfactual data and demonstrations achieves 84.0% mean success across four held-out LIBERO-90 tasks, compared with 47.0% for a full-context inverse model and 21.5% for a local-context model trained only on demonstrations. For forward dynamics, counterfactual supervision reduces RGB prediction error by 34.5% and raises outcome identification from 21.1% to 71.3% among 16 same-state outcomes. OPENWAM provides a common testbed for comparing WAM interaction designs and for studying dynamics learning from video data beyond successful demonstrations.

We study how a parity–time-symmetric photonic environment changes the charging of Dicke and Tavis–Cummings quantum batteries. For atoms prepared in the collective ground state the excitation number grows as the square of the charging time, with a coefficient set by the initial mean photon number; the counter-rotating terms of the Dicke model add one to four times that number. Conservation of the total excitation number turns this into an exact ceiling on the stored Tavis–Cummings energy, the transition energy times the smaller of the photon and atom numbers, met numerically to within two per cent, which the Dicke terms overshoot by a few per cent at ultrastrong coupling. We then couple a battery to a lossy–gain pair, keeping the parity–time symmetry intact by putting an identical battery in each cavity; the closed moment system puts the exceptional point at the loss rate. Below it the gain cavity fills and the gain-side battery charges while the lossy-side one stays empty; above it photons are shuttled away and dissipated. Solving the full Lindblad equation with batched Monte–Carlo wave functions, we find that the gain-side battery gains thirty-nine, twenty-six and twenty-eight per cent in stored energy, power and ergotropy in the broken phase and loses sixty-five, thirty-five and sixty-five per cent deep in the unbroken phase, where the lossy-side battery overtakes it. Restoring the counter-rotating terms leaves all of this unchanged within the statistical error; what they add is a reactive oscillation at twice the transition frequency that does no net work but nearly doubles the peak of the instantaneous power. The extra energy is of lower quality, the recoverable fraction dropping from about ninety-eight to ninety per cent. The optimum sits on the broken side, close to but not at the exceptional point.

{A GEANT4-based simulation study of a silicon–tungsten (Si–W) sampling calorimeter is presented to investigate electromagnetic shower development and detector response in the energy range of 1–120 GeV. Two detector configurations are considered: a reduced sequential-sampling prototype (Configuration-I), where a single silicon detector layer is placed downstream of progressively increasing tungsten absorber depths to sample the longitudinal shower evolution, and a fully instrumented sampling calorimeter (Configuration-II) consisting of twenty silicon layers interleaved with tungsten absorbers. The detector response is studied in terms of longitudinal shower development, shower maximum position, energy deposition, detector linearity, and calorimetric performance. Configuration-II provides simultaneous longitudinal sampling and allows event-by-event reconstruction of the deposited energy, yielding the expected linear response and energy resolution. Configuration-I reproduces the average longitudinal shower profiles and exhibits a strong linear correlation with the integrated shower response of Configuration-II. Although the absence of event-by-event multi-layer sampling prevents a direct determination of calorimetric energy resolution in Configuration-I, the observed correlation between two configurations is used to relate the response of the reduced prototype to that of the fully instrumented calorimeter and to estimate the corresponding calorimetric performance. The results demonstrate that the reduced prototype successfully reproduces the main characteristics of electromagnetic shower development observed in the fully instrumented detector, establishing its usefulness for detector characterization and test-beam studies when the construction of a complete prototype is limited by available resources.

This paper extends complex Gaussian process regression-based transfer function (TF) estimation from the single-input single-output setting to closed-loop multiple-input single-output (MISO) systems. Accurate estimation of individual TFs in closed-loop MISO systems is challenging as feedback and correlation between measured signals can obscure the contributions of each input. We formulate a Gaussian process-based estimator for MISO systems, provide analytical conditions under which each system module can be separately identified, and show that the estimator converges to the true TFs in the zero noise limit. Moreover, we demonstrate how conventional closed-loop bias handling techniques relying on external excitation can be employed to construct appropriate regressors. Monte Carlo simulations support the analytical identifiability results and highlight the trade-off between closed-loop bias and error introduced by constructing exogenous regressors partly from endogenous signals. The presented work offers an expressive, nonparametric alternative to existing estimation approaches for MISO systems, while providing a probabilistic characterization of the estimation uncertainty.

Metal-organic frameworks (MOFs) are widely explored for gas separation, catalysis or energy storage, yet their integration with other electronically functional materials remains largely unexplored. Here, we integrate the layered Hofmann-type spin-crossover (SCO) MOF {FeII(pyS2Et)2[PtII(CN)4]} into electrical nanodevices to control the properties of two-dimensional (2D) quantum materials. Our molecular approach exploits the stimuli-responsive nature of SCO layered materials as switchable building blocks. We demonstrate selective modulation of the electronic transport in van der Waals heterostructures interfacing SCO with 2D quantum materials (few-layer graphene, magnetic CrSBr, superconducting NbSe2) through the strain induced via thermal and light-induced spin transitions. In graphene, the conductivity is selectively switched by light. In spin-valves based on CrSBr bilayers, the MOF triggers magnetic hysteresis (absent in pristine CrSBr) and enables tunable non-volatile zero-field memory. In NbSe2, light modulates the superconducting critical current and transition temperature. These results establish layered stimuli-responsive SCO MOFs as active molecular control elements for 2D quantum materials, providing a route to strain-mediated control of electronic, magnetic, and superconducting functionalities and extending MOF-based architectures beyond traditional porous-matter applications towards multifunctional electronics and spintronics.