Papers

Even if it were possible, attempting to model the motion of a fluid in a pipe by tracking each individual molecule would be an extraordinary waste of computation. The emergent laws of fluid mechanics accomplish the same goal on a laptop instead of a datacenter. But a sufficiently complex fluid such as in Earth's atmosphere again requires a datacenter, this time to resolve the many contortions in the fields making up our weather. Might there be some analogous set of emergent laws that would represent the effect of these perturbations without needing their explicit representation on a numerical grid? We argue that the existence of a symmetry principle implies such laws are likely to exist. The atmosphere is scale invariant, implying that the observed structure at one scale is nothing but a stretched version of that at any other scale, at least within the invariant regime. Though some emergent laws built on scale invariance exist, they are built on a challenging ontological foundation, making them hard to develop further. We propose a path forward, first by introducing a new fundamental building block of the atmosphere that we term a “turbulon”, inspired by a lesser-known generalization of turbulence theory that accounts for buoyancy. We then construct the Superposition of Turbulons and Eddies Atmospheric Model (STEAM), and show that simulated atmospheric volumes are broadly plausible when compared to state-of-the-art hydrodynamic simulation output. Some aspects of STEAM clearly need improvement, though other statistics reproduce observations better than hydrodynamic models, and visualizations of simulated clouds are strikingly realistic. We estimate STEAM's computational cost to be up to a million times less than hydrodynamic models, suggesting that the Navier-Stokes equations may not, in fact, be required for simulation and prediction of Earth's weather and climate.

As researchers tackle interdisciplinary problems, they face the need to deepen expertise in primary areas while rapidly acquiring knowledge in secondary domains. Generative AI (GenAI) is increasingly positioned to meet this need, from general-purpose chat assistants to Deep Research tools marketed as autonomous research agents. Prior work has examined how researchers use GenAI to support single-discipline or general research tasks. However, we know little about the goals and GenAI practices in interdisciplinary research. We conducted a longitudinal study and semi-structured interviews with 15 interdisciplinary researchers to examine how interdisciplinary researchers actually orchestrate GenAI. Findings show that researchers leaned on GenAI to fill knowledge gaps while maintaining epistemic agency for novelty discovery. We also uncovered an expertise paradox: GenAI outputs were hardest to verify when most needed. Our empirical insights motivate GenAI designs that calibrate verification to researchers' expertise, nudge toward cross-domain synthesis, and adapt prompting and outputs to disciplinary conventions.

Let $C(d,2)$ be the optimal dimension-free ratio between the polytorus and Boolean-cube norms of complex multiaffine polynomials of degree at most $d$. For every $d$, $C(d,2)$ equals the unrestricted complexification constant of the real cube for complex polynomials of degree at most $d$. Moreover, $\lim_{d\to\infty}C(d,2)^{1/d}=1+\sqrt2$, and for every $d\ge3$, $\beta(1-6/(5d))(1+\sqrt2)^d\le C(d,2)\le(1+\sqrt2)^d$, where $\beta=15\pi/[2(5\sqrt{26}+\log(5+\sqrt{26}))]=0.8473222863\ldots$.

We prove that fiberwise-connected finite locally free commutative group schemes over a ring $R$ are equivalent to an explicit full subcategory of the derived category of modules over the Cartier ring of $R$.

We establish linear time precutoff for oriented kinetically constrained models (KCMs) on a $d$-dimensional box of side length $n$, whenever the finite-volume spectral gap is bounded away from zero uniformly in $n$. A typical example is the North-East model, a $0$-$1$ spin system on the two-dimensional integer lattice that evolves according to the following rule: whenever a site's southerly and westerly nearest neighbours have spin $0$, with rate one it resets its own spin by tossing a $(1-q)$-coin; at all other times its spin remains frozen. This settles, in the oriented case, a conjecture which states that there is linear time precutoff for all KCMs in the (open) ergodic parameter regime ($q>q_c$). The result has been shown recently, for general update families, in a perturbative regime ($q$ close to $1$). The previous general upper bound on the mixing time for oriented update families, in the full ergodic regime, was $O(n \log n)$. The proof combines a spectral estimate for a certain killed process with a graphical coupling to a stationary process and a backward witness path argument, exploiting the orientation of the constraints, to control the probability of failing to couple.

Experimental Fluid-Structure Interaction (FSI) often involves phenomena that occur simultaneously across vastly different spatial and temporal scales. For example, the separation bubble behind a cylinder and the vortex interactions in its wake, or the flow produced by an individual swimmer and the flow generated by the movement of the entire swarm. Imaging-based flow studies of such FSI phenomena are challenging and often lead to separate analyses of different flow regions. This work introduces simultaneous shadow and laser-based Particle Image Velocimetry (PIV) to obtain concurrent near-wall and far-field flow velocity measurements. The experimental setup developed here enables the first simultaneous comparison between shadow- and laser-based PIV, demonstrating that shadow velocimetry can achieve accuracy comparable to laser-based measurements at the macro scale. We then use this new combined flow measurement technique to fully capture the velocity field generated by a shrimp-inspired underwater robot, employing laser PIV to record the flow surrounding the robot and shadow PIV to focus on the actuators, imaging regions that the laser technique cannot resolve due to wall reflections and shadows behind the structures. The simultaneous application of shadow- and laser-based PIV offers a new experimental avenue for gaining insights into the physics of FSI phenomena.

Let $\Gamma$ be a finitely generated hyperbolic group, and let $Q$ denote the conformal dimension of its Gromov boundary $\partial\Gamma$. We prove that for every $p < \frac{Q}{Q-1}$, the group $\Gamma$ admits a proper uniformly Lipschitz affine action on $\ell^p$ with compression exponent $1/p$. The $\ell^p$-space hosting the action is designed from a Markov chain on a specialized hyperbolic filling of $\partial\Gamma$.

EEG electrode layouts differ across devices, from consumer headsets to 128-channel nets, and learned methods for completing missing electrodes assume a fixed layout and within-dataset training, so they cannot be applied to unseen devices. We propose the coordinate-conditioned detector-atom network (CC-DAN), a self-supervised convolutional dictionary model whose atoms are spherical-harmonic functions of electrode coordinates and whose detector pools observed electrodes with coordinate-dependent weights, making it independent of electrode number, order, and names. In a plug-and-play protocol, a single model pretrained on five public datasets (132 subjects) is applied zero-shot to nine held-out sets with 8 to 128 channels and compared with nearest-neighbor and spherical spline interpolation (SSI), a within-subject oracle, and learned baselines on signal error, band-wise error, and downstream tasks. With five or fewer observed electrodes, sparse real layouts, and hemispheric dropout, CC-DAN beats both interpolations in nearly all subjects and approaches the oracle; on the EGI net and 8- to 16-channel devices it keeps this advantage up to 15 electrodes, whereas SSI is superior with 15 or more electrodes on dense 10-05 montages. CC-DAN is also the most robust to electrode-position errors. Mean-squared-error gains do not carry over to downstream accuracy, and the amplitude shrinkage of MSE-optimal prediction degrades power-based measures, which a variance-matching loss and gain calibration mitigate. The resulting regime map tells practitioners where learned completion pays off and where classical interpolation remains the method of choice, and the plug-and-play protocol lets one pretrained model connect an unseen device to a fixed-montage pipeline without retraining.

Encrypted training relies on keeping server-side updates low-degree. This constraint traditionally excludes models whose weights inhabit a compact Lie group (notably variational quantum circuits, where every trainable weight is an $\mathrm{SU(2)}$ rotation). Expressed in Euler angles or discrete alphabets, these updates appear transcendental, historically demanding prohibitive costs: one client–server round per gate, or upwards of $25{,}000$ operations per weight. This penalty is strictly an artefact of coordinates. In the unit-quaternion (spin) chart, group composition is exactly bilinear (degree two, with coefficients in $\{-1,0,+1\}$). Consequently, encrypted rotation updates cost one multiplicative level and federated averaging costs zero in any levelled homomorphic scheme, completely eliminating bootstrapping. This implementation-independent algebraic property is confirmed across two cryptographic backends, introducing only $0.0$ and $-2.0\times10^{-12}$ rad of aggregation error. Leveraging this reduction yields a non-interactive protocol for encrypted federated training of hybrid quantum–classical networks. It includes correctness proofs for aggregation and sign handling, plus a compilation lemma proving parameterised entanglers add only constant-factor overhead without altering the depth class. Empirically, a paired five-seed study confirms zero measurable utility tax ($\Delta=+9\times10^{-6}$ MSE, $p=0.92$), and a noise-budget ablation falsifies the hypothesis that encryption noise regularises. These convergence trends replicate across datasets and scale to $20$ clients. Finally, hardware validation on a $156$-qubit processor achieves $0.9918$ fidelity against a $0.99957$ unencrypted control.

Virtual sensing repeatedly reconstructs physical fields from changing observations, often on a fixed geometry. We investigate how shared spatial computation reduces the energy of these updates while retaining the selected checkpoint and its evaluated predictions. In a heat-exchanger service, standard compiler freezing and explicit trunk reuse give similar operating energy reductions relative to graph replay: approximately 1% at one request per second and 20% at forty requests per second. In 15 W mode with fixed clocks, reuse with graph replay completes the same request sequence with 22.0 to 22.5% less energy than eager execution, including preparation and waiting. DeepONet and Fourier neural operator (FNO) controls distinguish the effects of reusable arithmetic and launch overhead. Preparation, artifact construction, and worker replacement add costs outside repeated inference. These results connect operator structure to operating energy and show how update frequency and execution lifetime govern the benefit of computation reuse in physical-field virtual sensing.

We introduce a notion of topological complexity of a group with respect to a family of subgroups, with the case of the trivial and diagonal families recovering the classical category and topological complexity, respectively. These invariants come equipped with a battery of upper and lower bounds derived from functoriality and Bredon cohomology. We introduce a new family of subgroups, the permutational family, and show that the corresponding invariant is a lower bound for the distributional topological complexity of Dranishnikov-Jauhari and Knudsen-Weinberger. As a first application, we show that the permutational and diagonal families coincide, and thus that classical and distributional topological complexity are equal, for a large class of torsion-free groups, extending work of Dranishnikov. Second, we show that the lower bounds of Grant-Lupton-Oprea are in fact lower bounds on distributional topological complexity; in particular, it follows that Farber's conjecture holds distributionally.

Language models gain accuracy from more compute at test time, but majority voting over independent samples saturates: as samples grow, the vote converges to the model's most frequent answer. Communication can add what sampling cannot: an agent that solves a problem can pass the key step to the others. We present TalkMesh, a decentralized mesh of small language model agents that learns when and what to communicate. Each agent samples a proposal and scores it with a trained confidence head. The most confident agent broadcasts a hint; agents below a confidence threshold revise, keeping each revision that outscores its proposal. Gossip consensus approximates the vote weighted by confidence without a coordinator. A talk policy, trained with group relative policy optimization on the change in correctness after revision, writes hints and revisions. With three agents, which together generate at most six outputs, the mesh reaches the accuracy of majority voting over 32 samples with each of three models. Trained with at most 8 agents and evaluated with 32, it raises accuracy from 0.568 under self-consistency to 0.705 (Qwen3.5-0.8B, GSM8K) and from 0.492 to 0.722 (SmolLM3-3B, MATH-500). When 4 of 8 agents collude on a wrong answer with fabricated confidence and poisoned hints, majority vote accuracy falls to 0.000 (Qwen3.5-0.8B, GSM8K). A defended mesh, whose agents rescore solutions with their own confidence heads, retains 0.507. Across reasoning, embodied coordination, and traffic signal control, messages improve a decision when the acting agent cannot observe the information it requires and another agent can send it.

Topological mechanical metamaterials generally derive their robustness from one of two distinct mechanisms: kinematic topology, which localizes zero-frequency floppy modes through geometric compatibility, or band topology, which localizes finite-frequency waves through topological bandgaps. Because these mechanisms arise from fundamentally different physical principles, mechanical systems are typically engineered to exploit either static or dynamic topological functionality, but not both simultaneously. Here we demonstrate that a single mechanical metamaterial can host two independent topological channels governing static and dynamic response within the same architecture. Using a generalized rotor-chain lattice with two coupled rotational degrees of freedom per unit cell, we realize the coexistence of a topological polarization mode and a finite-frequency topological band mode. We show that the two channels are governed by distinct invariants and can be programmed independently through geometry, angular asymmetry, and stiffness dimerization. Remarkably, despite being governed by distinct topological invariants and controlling different physical responses, the two channels undergo topological transitions at a common symmetry-controlled critical geometry. Experimental measurements using scanning laser Doppler vibrometry confirm simultaneous localization of a boundary floppy mode and a finite-frequency domain-wall state in the same structure. These results establish dual topological channels as a general design principle for multifunctional mechanical metamaterials and demonstrate how static deformation and dynamic wave transport can be programmed independently within a single architecture.

Large-scale recommenders increasingly adopt the sequential generative recipe behind large language models, bringing the Transformer into recommendation along with design choices made for text, including Rotary Position Embedding (RoPE). In language models, RoPE encodes token indices for relative position reasoning, but in recommendation, an interaction index records only event order, saying nothing about elapsed time, behavioral cycles across scales, or calendar phase. We revisit this choice and propose T-RoPE, a time-aware RoPE for sequential generative recommendation that replaces index-only rotation with timestamp-based angles, learnable temporal coefficients, multiscale frequency banks, shifted query alignment, and non-stationary key rotation. We prove that standard RoPE, even on timestamps, remains time-translation invariant and cannot distinguish seasonal contexts, and that T-RoPE breaks this invariance while preserving the RoPE interface. Across five public benchmarks, T-RoPE achieves the best result on every metric on every dataset, improving over the strongest baseline by 78–130% in HR@10 on the sparse PixelRec data and 8–12% across metrics on Amazon Books. On an industrial-scale e-commerce dataset with more than 6B interactions, it improves every metric over the HSTU + Time RAB backbone by 13–82%, with ablations attributing the largest gains to multiscale frequencies ($+56\%$ NDCG@50) and non-stationary keys ($+4\%$). An online A/B test in the Shop app yields positive lifts in conversion rate ($+0.33\%$) and order count ($+0.63\%$). We also provide forward and backward algorithms whose added cost is linear in sequence length and head dimension, keeping time-aware RoPE practical for large generative recommenders.

Using relations established by Bringmann and Kane in a recent preprint based on the function $\omega_{k+1,D}$ studied by Mono, Rolen, and the second author, we introduce two further functions naturally associated with a construction by Bringmann and Mono. We exploit results by Löbrich–Schwagenscheidt, Mono, and the second author on representations of these functions as traces of cycles integrals in order to provide more identities of the same kind.

The infrared spectra of many carbon-rich post-asymptotic giant branch stars are dominated by emission features from aromatic and aliphatic hydrocarbons, but the chemical structure of the carriers remains unidentified. Class D sources show unusual emission profiles with strong aliphatic emission, providing a stringent test of carrier candidates. Here we investigate whether hydrogenated carbonaceous molecular particles can be the carriers of these features. Rather than assuming candidate geometries we generate the initial structures using ab initio molecular dynamics simulations and optimize them using density functional theory. We then compare the computed emission spectra to JWST NIRSpec and MIRI/MRS observations of Class D sources. The structures providing the best agreements contain sizeable domains with defect-bearing aromatics and hydrogenated-aromatics connected by aliphatic bridges, in contrast to the "arophatic" cluster model of isolated two- to three-ring aromatics connected by aliphatic and olefinic bridges. As a representative example, we discuss in detail the molecular particle C130H96 (radius of 5.5 A) and compare its computed spectrum to the JWST spectra of IRAS 05110-6616, a Class D2 source with apparent shifts of the aromatic C-H out-of-plane bands. The calculated spectra are in qualitative agreement with the observed 3 micron profile, the aliphatic C-H bending bands, the broad 8 micron feature, the 11-14 micron C-H out-of-plane region, and the bands at 16 and 21 microns. We hypothesize that UV photo-driven fragmentation of such molecular particles releases aromatics linked by aliphatics, followed by the formation of polycyclic aromatic hydrocarbons and fullerenes as these objects evolve into planetary nebulae.

Let $E/\mathbb{Q}$ be an elliptic curve. Zywina's refinement of Koblitz's conjecture predicts that there are infinitely many primes $p$ of good reduction for which $\#E_p(\mathbb{F}_p)$ is prime precisely when $E$ has no congruence obstruction. For non-CM elliptic curves over $\mathbb{Q}$, we classify all primitive congruence obstructions. In particular, we show that a primitive obstruction of composite level can occur only at $6$, $10$, $14$, $15$, or $30$ and determine the corresponding Galois images. As a consequence, for non-CM elliptic curves, the existence of any congruence obstruction is detected by the mod $210$ Galois image, so the positivity of the Koblitz–Zywina constant is determined at level $210$. We also determine which primitive obstruction levels can occur simultaneously. Finally, we prove, assuming GRH, that if $E$ has no congruence obstruction, then for every $0<\kappa<1/8$ there are infinitely many primes $p$ of good reduction for which the least prime divisor of $\#E_p(\mathbb{F}_p)$ is greater than $\kappa\log p$.

Large language models are often post-trained on expert demonstrations using cross-entropy (CE), even when the downstream objective is not to imitate the demonstrated solution but to produce any output accepted by a verifier. This mismatch is seen in verifiable domains with multiple correct solutions, such as mathematical reasoning and code generation, where training data may contain only one expert solution per problem. We show that minimizing cross-entropy can be misaligned with minimizing verifier risk; two policies can assign identical likelihood to the observed demonstrations while placing different probability mass on incorrect outputs. This is formalized through a learning-theoretic counterexample in which CE minimization selects a suboptimal policy. We identify that controlling the support of the learned policy can solve this problem by preventing probability mass from spreading to unsupported outputs. Since support size is non-differentiable and computationally intractable, we propose entropy-regularized cross-entropy (ER-CE), using token-level Shannon entropy as a tractable proxy. Finally, across mathematical reasoning and code-generation benchmarks, we find that entropy-regularized training consistently improves verifier accuracy over standard cross-entropy. Our results identify a simple failure mode of imitation-based post-training in verifiable tasks and provide a practical objective that is better aligned with producing correct outputs.

Language models are often evaluated on curated benchmarks that underrepresent the complexity of enterprise deployments. We introduce HARDEN, a constrained evolutionary search method to adapt the input of existing evaluation cases into more challenging variants while keeping their expected outputs fixed. HARDEN searches along generated domain-specific complexity axes while enforcing feasibility constraints such as preserving task semantics, realism, and execution validity. Across FinQA, PubMedQA, and ContractNLI and three Qwen3.5 model scales (35B-A3B, 122B-A10B, and 397B-A17B), HARDEN reduces task-model accuracy by 22.7% on average and by up to 49.9% relative to single-pass baselines using the same feasibility checks. These results show that evolutionary search can produce substantially harder valid evaluation cases.

In this work, we present MOCHA, the first, to our knowledge, reinforcement learning based approach to computing a family of Pareto-optimal policies across the design space of a robot using a single network. Specifically, MOCHA leverages the hypernetwork architecture to learn a network that produces specialized network parameters optimized for a given objective and parameterized robot design; we term this a multi-objective design hypernetwork (MDH). We demonstrate the capabilities of MDHs to represent a complex family of design-dependent strategies on two distinct robot morphologies, each with six design dimensions and across 2-3 objectives. Moreover, we propose an approach for efficiently producing a Design Pareto set using evolutionary search of the learned policy network, generating the optimal design-policy combination for each objective prioritization. Lastly, we provide an efficient method for computing generalist robot designs which achieve the best cumulative performance across the entire set of objectives.

CryoEM map interpretation requires features that are spatially localized, consistent across samples, and informative across spatial scales. Most deep learning methods for map annotation extract features from fixed voxel grids. However, implicit neural representations (INRs) are able to model volumetric data as scale-agnostic, coordinate-conditioned functions. INRs are therefore attractive for cryoEM, but fitting a separate INR for each map is too expensive for large-scale feature extraction and produces representations that are not aligned across samples. We introduce Atelier, a self-supervised framework that amortizes INR fitting for reconstructed cryoEM maps. Pretrained on 5,439 Electron Microscopy Data Bank maps, Atelier is a transformer-based hypernetwork that generates high-fidelity reconstructions across a wide range of protein structures, including large multi-subunit assemblies. Beyond reconstruction, the INR generated by the pretrained transformer exposes a continuous, local feature field through its intermediate activations at any spatial query point, a property that voxel grid and patch-tokenizer architectures do not naturally provide. Used as auxiliary channels to a 3D nested U-Net annotation head trained from scratch, these coordinate-conditioned features improve performance on eight voxel-level property prediction tasks over a volume-only baseline. Our results demonstrate that amortized implicit neural representations are an effective primitive for geometry-aware analysis of cryoEM data.

A deployed B2B music-discovery platform serves both query-driven search (text prompts, vibe tags) and seed-driven recommendation (seed-track and artist stations) over one licensed catalog, one LAION-CLAP joint audio-text embedding space, one candidate-generation filter, and one ranking head – and neither path consumes end-listener behavioral signal. In this content-only regime, curator judgment is the principal feedback signal available, and offline cosine similarity predicts it poorly: 38% of cosine-nearest neighbors are rejected by curators. The rejections reveal a clean partition: a majority (55%) are sound failures the encoder could address (style, tempo, mood mismatch), and a substantial minority (37%) are context failures orthogonal to the waveform (wrong language, holiday content, devotional content, rights and lyric flags). We deploy this sound-vs-context decomposition as feedback infrastructure, routing each failure mode to the layer that can absorb it: context failures to a constraint filter at candidate generation, sound failures to an embedding reweighting head at the representation layer – both below the search/recommendation split, so a single curator loop maintains both experiences. On 1,200 curator judgments collected over two production rounds one month apart, the combined intervention reduces rejection rate from 38.17% to 28.83% (-24.5% relative, McNemar chi-squared = 22.4, p = 2.2e-6). An accounting decomposition attributes 4.08 pp of the drop to filter-eligible categories and 5.25 pp to the rest; the deployment was unblinded and compound, so this is a production accounting bound, not a causal estimate. We present this as an industrial case study rather than a validated general method, and close with lessons for content-only discovery: the failure partition is orthogonal to the paradigm partition, and corrections land in layers shared by both paradigms.

Large-scale signal processing systems, including aeronautical and aerospace platforms, increasingly rely on tensor operations, whose distributed execution is constrained by communication, memory, and latency bottlenecks. In this paper, we propose a tensor-aware coded computation framework for distributed mode-1 tensor–matrix multiplication that operates directly on tensor subtensors, avoiding explicit unfolding and preserving multi-dimensional structure. For a fixed partitioning configuration, the proposed tensor-PolyDot scheme achieves the same recovery threshold, communication cost, worker-side computation, and memory requirements as conventional matrix-based PolyDot schemes. However, by leveraging multivariate encoding, it replaces high-degree univariate interpolation at the fusion node with structured low-dimensional decoding. This leads to a substantial reduction in decoding complexity and latency, without affecting any other system metric. Additionally, the tensor-based approach improves memory locality and enables parallel decoding across tensor modes, making it well-suited to modern hardware architectures. Numerical results show decoding speedups proportional to the tensor partitioning factor, reaching up to an order-of-magnitude improvement in practical settings. These advantages make the proposed framework particularly relevant to latency-critical aeronautical and aerospace distributed signal processing systems.

We present a new inference-time sampler for diffusion models that gives a pretrained model a capability it was never trained for: constructing the atlas of the population it synthesizes. The sampler converges from every random seed to the population's central anatomy, which we call the intrinsic atlas. The advantage is threefold. (1) It requires no retraining. A diffusion model that has already learned a coherent population, including the released ones, yields its atlas in a single inference pass without involving deformable registration. (2) It applies to multiple domains, such as brain MRI, chest X-ray, faces, and 3D shapes. (3) It extends to subpopulations. One age-conditioned model gives an atlas at any age in its training range, and the resulting family reproduces the CSF expansion of healthy aging. Evaluated as a registration target, the intrinsic atlas is best or second-best on every dataset against classical and learned templates, and the most central template on held-out brain MRI cohorts. Atlas construction can be reframed as a byproduct of generative modeling: a diffusion model is a learned representation of population structure, and the atlas is what it already contains.

Many real-world high-dimensional time series exhibit long-memory, but Gaussian graphical model testing in this regime remains understudied. We develop a direct, data-adaptive test statistic for assessing conditional independence in the graph structure of stationary Gaussian time series. We establish a finite-sample, Berry–Esseen type Gaussian approximation bound for the statistic, which applies to both short-memory and long-memory time series. The testing procedure uses a block bootstrap method, on which we provide a finite-sample validity result including in the ultra-high-dimensional scenario, and can be extended to comparing graphical structures in two-sample tests. We also develop a consistency-empowered correction to the statistic and show that such tests attain asymptotic consistency in both size and power. Our proposed method is applied to real-world fMRI data to understand functional connectivities within brain in different periods.

The noncentrosymmetric superconductor NbReSi (Tc=6.5 K) is being actively explored for unconventional pairing that is allowed by its broken inversion symmetry. Despite broad experimental interest, what drives its superconductivity remains an open question. We show that the electronic states at the Fermi level are dominated by Nb and Re $d$-states. Vibrations of the same heavy-metal framework contribute more than 95% of the electron-phonon coupling, with the strongest coupling arising from the sublattice-selective breathing mode involving strongly bonded Re atoms. Harmonic Migdal-Eliashberg theory overestimates Tc by more than 60%. The zero-point anharmonic hardening of these modes reduces the coupling by 20%. The leading vertex correction becomes important as well despite low phonon energies due to the contribution of spatially localized Re-d orbitals to the electronic band structure near the Fermi level. Together, anharmonicity and vertex corrections bring the calculated Tc and superconducting gap into close agreement with experiment, establishing NbReSi as a moderately coupled, phonon-mediated superconductor whose quantitative description requires going beyond harmonic Migdal–Eliashberg theory.

Platform policies are increasingly tested on artificial users, making agent fidelity important. Yet convincing fake profiles could also manipulate perceived public opinion before elections. Validation has concentrated on agreement with human behaviour and has paid little attention to whether an agent behaves in line with the profile it was given. The present study profiled eight Serbian participants through a questionnaire, a deep interview, and a written self-presentation, recorded their reactions to sixty-eight social media posts, and asked four language models to predict those reactions under five prompt conditions varying profile content and instruction style. Attitudinal content improved prediction over demographic backstories by a wide margin. Agents matched their stated profiles more closely than participants matched their own survey answers, and consistency proved unrelated to fidelity once profile information was present. Instructing models to respond intuitively and immediately rather than analytically gave the highest fidelity of any condition and cut the compression of individual differences from seven times the human level to three. The advantage held on posts about topics the questionnaire never raised, where that condition reached the highest fidelity of any setup and beat a crowd baseline by a wide margin, which suggests that agents prompted this way could serve as general-purpose simulated users rather than specialists on the topics they were profiled for. Results may bear implications for the development of language models, because intuition-based setups appear better suited to some tasks than reasoning-based ones.

Accurate direction-of-arrival (DOA) estimation from a small multiple-input multiple-output (MIMO) radar array is limited by its aperture, while adding antennas increases hardware cost. Snapshot-based array emulation reconstructs large-array observations before applying multiple signal classification (MUSIC), but reconstruction errors can distort the spectral peaks used for DOA estimation. We instead generate an idealized large-array MUSIC spectrum directly from small-array measurements. Coarray compression combines redundant virtual channels, and the resulting covariance and MUSIC spectrum condition a diffusion model. Consensus over peaks in multiple generated spectra yields the DOA estimates. Simulations with fluctuating targets show that the method outperforms a scene-matched snapshot-reconstruction network and an otherwise matched deterministic spectrum predictor. Small- and large-array MUSIC and their Cramér-Rao bounds provide reference comparisons. The method improves angle estimation using only the small array at inference.

Let $S$ be a closed translation surface of genus $g\ge 2$ with $\mathop{area}(S)= 4\pi g$. We show that for any $\lambda \in (0, 1)$ there exist $\lfloor \lambda \cdot g\rfloor$ homologically independent simple closed curves of length at most $ C(\lambda) \cdot \log(g)$, where $C(\lambda)$ is a constant that depends only on $\lambda$. This result is obtained from a mixed graph construction based on a Voronoi graph of the surface with the cone points as seeds and its dual graph. We also give a complementary result using only the Voronoi graph that produces $2g$ short loops that form a homology basis. This construction, however, depends on the length of a shortest saddle connection of the surface $S$.

Beyond collision avoidance, socially competent robot navigation requires adherence to implicit social conventions that vary across contexts, cultures, and deployment requirements. Many conventional navigation policies learn a single normative behavior, either through reinforcement learning against a fixed reward function or imitation of human demonstrations, exposing no interface for adjusting that conduct at runtime. We present a diffusion-based navigation framework whose social behavior can be tuned at deployment: a desired style is specified, such as how closely the robot passes, which side it yields to, or how much it defers to groups, and the planner adapts accordingly. Continuous style axes can be followed independently or composed, spanning a behavioral space rather than discrete, primitive-based specifications. A feasibility projection layer separates learned social behavior from kinematic feasibility and collision avoidance. A single-axis sweep illustrates a tradeoff curve that strictly dominates the evaluated fixed-behavior baseline configurations, and stylistic differences are replicated in real-world demonstrations.