Papers

We present an $\infty$-categorical refinement of classical $KK$-theory and $E$-theory for graded, $G$-equivariant for a locally compact second countable group $G$, Real $C^*$-algebras. As a consequence, we obtain a couple of fiber sequences connecting the different $K$-theory spectra, and explicitly analyse the induced endomorphisms on $K_0$-groups. We also answer precisely some folklore questions in the area. Among other things, we show that graded $KK$- and $E$-categories are not a Dwyer-Kan localization of graded $C^*$-algebras.

Clinical decisions depend on anticipating how a patient's condition will evolve as treatment is adjusted and new evidence becomes available. Existing medical world models capture patient-state evolution, but often define the evolving state around a particular modality or task, limiting reuse as inputs or prediction targets change. These models also retain history through observation sequences or recurrent or latent summaries rather than exposing preceding patient states and clinical actions as an ordered transition context. This distinction matters because similar current measurements may follow different recent changes, treatment exposures, and observed responses. To tackle these limitations, we propose PatientWM, a clinical world-model framework built around a unified patient state and sequential latent history. Availability-aware encoders integrate observations available at each decision time into the patient state, while typed decoders produce task-specific clinical readouts without defining the state around any single target. PatientWM further conditions each next-state prediction on the current state and supplied action together with a bounded, ordered history of preceding latent patient states and clinical actions, preserving recent treatment and response context. Experiments across MIMIC-III, MIMIC-IV, and three external clinical datasets show superior performance in clinical prediction, multi-step forecasting, and action-planning evaluation. Transfer experiments further demonstrate the generalizability of PatientWM's patient state. Together, these results establish its ability to forecast clinical evolution and support planning.

Scaling fault-tolerant quantum computing (FTQC) requires orchestrating millions of operations continuously across a tightly coupled classical-quantum control stack. A critical challenge in this orchestration is determining the optimal hardware-software boundary, i.e., the Quantum Instruction Set Architecture (QISA). We focus on cryogenic, fixed-connectivity qubit platforms, such as superconducting and topological Majorana qubits. For these, existing QISA choices either operate at too low a level, violating cryogenic thermal budgets at scale, or at too high a level, imposing unrealizable complexity on the control engine. To reason about these tradeoffs, we introduce the ARIES instruction set hierarchy, which categorizes QISAs into distinct abstraction levels. To systematically evaluate these levels, we develop an end-to-end evaluation toolflow featuring a cycle-accurate classical driver and hardware engine simulator. Using eight quantum applications across superconducting and topological qubit platforms, we measure instruction bandwidth, transceiver power dissipation, and measurement processing overheads. We demonstrate that while physical-level abstractions exceed the thermal limits of dilution refrigerators by orders of magnitude, canonical-level abstractions significantly underestimate hardware routing congestion and execution runtimes. Our Level 4 QISA (local logical operations) offers an optimal architectural balance. As quantum hardware platforms scale toward FTQC in the coming years, our findings provide concrete architectural guidance for the design of their classical control stacks and instruction interfaces.

Vision-based document retrieval has improved rapidly, but current evaluation settings still emphasize short documents and single relevant pages. This leaves an important gap between leaderboard performance and realistic long-document retrieval, where the evidence needed to answer a query may span multiple pages or even multiple documents. We present VL-DocIR, a page-level benchmark for evidence-complete visual long-document retrieval. VL-DocIR is built from 29,641 HTML-born documents rendered into 388,548 page images from Wikipedia, arXiv, PubMed, and SEC proxy statements. It contains 271,760 queries across 23 domains and six evidence structures, covering single-page, within-document multi-page, and cross-document evidence configurations. Evidence annotations are grounded to rendered pages and HTML element identifiers, and queries are filtered with a cleaning pipeline targeting clarity, correctness, closedness, and the absence of explicit layout references. We evaluate leading single-vector and multi-vector visual retrievers together with diagnostic retrieval settings such as document-context scoring, query splitting, reranking, and round-robin ensembles. The strongest evaluated configuration reaches 69.3 nDCGAll@10 and 80.6 RecallAll@10 under structure-balanced Macro evaluation, but source- and evidence-structure breakdowns show that cross-document retrieval remains substantially harder than single-page retrieval, and that document-context scoring is especially helpful for within-document multi-page queries. VL-DocIR exposes these failure modes and provides a more demanding target for visual long-document retrieval. The new dataset and code will be made publicly available at https://yufanchen96.github.io/VL-DocIR.

Local coordination cores are critically important in metal-organic frameworks (MOFs) and polyoxometalates (POMs), as they govern the fundamental chemical and structural properties of these materials. Existing methods, including element-resolved persistent homology (ERPH), have limited ability to translate local coordination cores into quantitative, scalable, and interpretable information. Here, we introduce persistent Mayer homology (PMH) to enable the detection of the geometric variability invisible by classical topology. By focusing on coordination cores formed by periodic metal-centered heavy-atom, carbon, and donor channels, PMH distinguishes different arrangements of the local coordination environment, linker framework, and potential coordinating atoms. We examine explainable PMH of orders $N=3$ and $N=5$ and separately combine each with ERPH, which characterizes the multiscale organization of chemically defined element groups across the extended framework, to predict eight nitrogen- and oxygen-related adsorption and self-diffusion properties. The resulting coordination-core PMH representations consistently improve predictive performance over ERPH alone, with the third-order Mayer representation providing the strongest overall performance. These results demonstrate that the inner coordination configurations captured by PMH contain essential structural information that is not fully represented by conventional element-resolved topological descriptors. This finding provides a quantitative interpretable topological framework for understanding local coordination environments and may facilitate the rational design of MOFs and POM-based materials.

Let $q$ be a prime power. An $(r, i)$-regular fat $\mathbb F_q$-linear set has exactly $r$ points of weight $i\ge2$, called fat points. All of its remaining points have weight one, and at least one such point occurs. We solve two open problems posed by three of the authors of this paper. First, we derive divisibility constraints on the point weights using the scalar stabilizer of the defining subspace. Consequently, a regular fat linear set of rank $2i$ in $\operatorname {PG}(1,q^n)$ has at most two or exactly $q + 1$ fat points. In the latter case, the fat points form an $\mathbb F_q$-subline. Second, we construct higher-dimensional regular fat linear sets whose fat points do not form an $\mathbb F_q$-subgeometry. In particular, for every $k\geq 3$ we obtain examples of rank $2k$ whose fat locus consists of an $\mathbb F_q$-subgeometry $\operatorname {PG}(k-2,q)$ in $\operatorname {PG}(k-1,q^4)$ together with one external point. This construction is then generalized further. Finally, we apply our projective-line results to $q$-matroid representability and present an infinite family of non-representable rank-two $q$-matroids arising from partial spreads.

Solar-blind ultraviolet communication (UVC) exploits near-zero solar background radiation and strong atmospheric scattering in the 200-280 nm band, enabling photon-level daytime communication and propagation modes beyond conventional line-of-sight (LOS) links. However, existing studies mainly focus on individual channel types, and a unified channel classification is still lacking. This paper presents a taxonomy of all four fundamental solar-blind UVC channels together with representative experimental validations and a comparative performance analysis. The proposed framework provides practical guidelines for selecting appropriate channel architectures for future solar-blind UVC systems.

Shapley-based measures are widely used for measuring variable importance in statistics and machine learning. However, these measures have a serious issue–a nonactive variable can get a positive importance measure when it is correlated with an active variable. Herin et al. recently proposed a measure called proportional marginal effects (PME) that mitigates this spurious effect. However, PME is developed for deterministic models and has high computational cost. In this article, we propose a new measure called proportional inclusive marginal effects (PRIME) that overcomes these two limitations of PME. The key idea in PRIME is to use a novel model-free variable selection technique to identify the active variables and include only them in the computation of the importance measure. Analysis of test functions and real datasets show that PRIME can accurately and reliably quantify variable importance from noisy input-output data.

Direct and parametric driving together generate nonlinear response structures beyond conventional Duffing hysteresis, including multiple switching regions and isolated branches. Understanding how these structures are connected requires identifying the bifurcations that reorganize the response as the parametric excitation is varied. At a phase difference of 45 degrees between the direct and parametric excitations, we experimentally and analytically determine how the steady-state response changes as the parametric excitation is increased. A single resonance peak first splits into two while the response remains single valued. Two successive cusp bifurcations then create two saddle-node pairs and establish a double-hysteretic response. At the parametric-instability threshold, one saddle node of each pair is displaced towards unbounded amplitude and detuning, and an isolated-branch structure emerges above threshold. A parametric representation of the steady state distinguishes amplitude extrema from genuine saddle-node bifurcations and provides analytical conditions for these changes. Measurements on a dielectrically actuated pre-stressed SiN nanostring reproduce the predicted response and associated switching behavior. These results establish the bifurcation pathway connecting the different response structures at this phase configuration, providing a systematic basis for controlling multistability and switching in parametrically driven nonlinear resonators.

Favored carriers of the Diffuse Interstellar Bands (DIBs) are carbonaceous molecules whose constituents have been presumably expelled by dying stars along with those of dust grains. The parallel evolution of DIB carriers and grains and their potential interplay are unknown. We assembled published R(V) measurements of 505 hot stars. We inverted a catalog of extinction data based on Gaia XP spectra and NIR photometry, and used the resulting new 3D extinction map to locate the opaque areas responsible for R(V) for each star. Using archived spectra, we measured their 5780, 5797 and 8620 A DIBs. In parallel, we compared local and line-of-sight (LOS) R(V)s from recent 3D maps with local and LOS Gaia DIB strengths from our recent map. Results: R(V)s from hot stars reveal variations on large spatial scales, confirming with independent data the published 3D pattern. Gaia DIB strength maps show large spatial scale variations similar to those of R(V) maps, with higher abundance of DIB carriers corresponding to higher R(V). LOS Gaia DIB strengths and R(V)s from 3D maps also correlate positively in the R(V) range 2.3-3.4, but there is no connection for higher R(V), likely due to the disappearance of DIB carriers in dense clouds. Despite averaging effects along the LOS to hot stars, we find differences in R(V)-DIB relationships among the three DIBs, with an evolution from negative to positive correlation from the zeta- to the sigma-type DIBs. This can be understood in the frame of the bimodal evolution of R(V) with the extinction, with the zeta (resp. sigma) DIB carrier disappearance in the decreasing (resp. increasing) phase of R(V). Various combinations of DIB strengths and DIB ratios may allow to define the range of R(V), but the relationship between DIBs and R(V) also appears to depend on the wavelength range used to determine R(V).

Molecular systems with optically addressable electron spins are emerging as versatile quantum sensing platforms, offering tunable qubit properties and molecular-scale spatial control. Owing to their longer lifetimes, coupled nuclear spins provide an attractive resource for enhancing sensing performance. For molecular spin systems, the ability to initialise, coherently control and readout nuclear spins at room temperature is crucial for realising such benefits. Here, we present the first optical detection of nuclear magnetic resonance and coherent nuclear spin control with a molecular system at room temperature. Specifically, we demonstrate optically detected Rabi oscillations of $^{14}$N nuclear spins in the photoexcited triplet state of the organic chromophore 6,13-diazapentacene. The strong hyperfine coupling between the electronic spin and $^{14}$N nuclei enhances the nuclear Rabi frequency, enabling fast coherent nuclear-spin rotations. This work lays the foundation for leveraging nuclear spins as memories for enhanced molecular-based quantum sensing.

A class of models which brings to the fore the inner logic of the calculus of probability instead of assigning probabilities to sentences is introduced and investigated.

Large audio-language models (LALMs) have become highly capable on a wide range of tasks spanning speech, music, and environmental sound. However, many modern LALMs resample input audio to 16 kHz mono, making all frequencies above 8 kHz unobservable. We present SpectralBridge, a lightweight extension that gives such a model access to the 8-16 kHz band by joining low- and upper-band representations with time-aligned paired cross-frequency attention. Applied to Audio Flamingo Next, the method trains only a small adapter and rank-8 LoRA modules, while leaving the original audio encoder, audio projector, and base language model frozen. SpectralBridge improves both zero-shot transfer and downstream fine-tuning: it raises zero-shot balanced accuracy on the ESC-50 audio classification dataset from 75.5 to 90.5, and when fine-tuned on Insect-10, a 10-way insect species task we construct from InsectSet459 recordings, under the same budget as a LoRA fine-tuned baseline, it reaches 55.4 balanced accuracy against the baseline's 26.2. On a benchmark where the answer depends only on the 8-16 kHz band, it gains 30 mAP over the frozen model. Erasing the upper-band stream from the trained model removes these gains, showing that they come from the high-frequency signal itself rather than from extra tokens or the adapted prompt format.

Nonlinear effects are widely utilized in quantum metrology to enhance measurement precision by leveraging complex dynamical behaviors. Here, we propose a quantum scrambling-integrated scheme that optimizes the selection of initial states and parameter encoding by exploiting the nonlinear dynamical trajectories and their geometric properties in phase space. We consider a two-mode Bose-Einstein condensate as a typical example. Within a certain parameter range, using uncorrelated spin coherent states located at saddle points and separatrices enables the measurement precision to reach the Heisenberg scaling. An analytical expression for the quantum Fisher information under long-time evolution is provided, and our analysis further reveals that the enhancement mechanism is governed by the sensitivity of the energy spectrum to parameter variations. Compared to the criticality-enhanced protocol that requires parameters to approach the critical point, our scheme offers a distinct advantage in its applicability over a broader range of parameters. Moreover, this scheme can also be applied to models that include one- and two-axis countertwisting interactions and chaotic dynamics. This Letter offers a new perspective for advancing the application of nonlinear dynamics to enhance the parameter estimation precision in quantum systems.

For the counts $(X_1,\dots,X_n)\sim\mathrm{Mult}(N;1/n,\dots,1/n)$ of $N$ observations in $n$ ordered bins we study the variance of the discrete total variation $d=\sum_{i=1}^{n-1}|X_{i+1}-X_i|$. An exact decomposition splits $\mathrm{Var}(d)$ into per-edge, adjacent-pair and disjoint-pair terms and yields, for fixed $n$ as $N\to\infty$, $\mathrm{Var}(d)/N=\alpha(n)-\beta(n)/N+\gamma(n)/N^2+o(N^{-2})$. For every $n\ge 2$, $N\alpha(n)$ is the classical normal-theory variance (Hoel, Kamat) of the sum of absolute successive differences of $n$ independent $N(0,N/n)$ variables; the correction $\beta(n)$ is obtained in closed form for $n\ge 3$ (the case $n=2$ is treated separately). Two disjoint edge differences $D_1=X_2-X_1$ and $D_3=X_4-X_3$ are exactly uncorrelated and asymptotically independent, yet $\mathrm{Cov}(|D_1|,|D_3|)=-1/(\pi n)+O(1/N)$: a correlation of relative order $1/N$, invisible in the Gaussian limit and absent under Poissonized sampling, whose constant is proportional to the exact joint fourth cumulant $\kappa_{2,2}(D_1,D_3)=-4N/n^2$ and is proved here on the lattice. The adjacent pair contributes the lattice-Edgeworth constant $a_*=-2\sqrt{3}/(9\pi)$, of which one quarter is a smooth Edgeworth term and three quarters a lattice-corner term; the disjoint skeleton $h(m)$ is closed through a Bessel generating function; and $\gamma(n)$ ($n\ge 5$) is expressed through a single triangular-lattice constant whose closed form we conjecture. Together with the mean of Part 1, the variance supplies the two moments used by moment-matched approximations to the null law of $d$. This is Part 2 of a series; the cyclic statistic $d^{c}$ will be treated in Part 5.

Ideal Poisson–Voronoi tessellations are random tessellations that arise as limits of Voronoi diagrams for Poisson point processes as their intensity tends to zero. For the Riemannian metric on a symmetric space of higher rank, Frączyk, Mellick, and Wilkens (arXiv:2307.01194) proved that the corresponding ideal Poisson–Voronoi tessellation is almost surely completely connected, with an unbounded side between any two cells. This contrasts sharply with the geometry of ideal Poisson–Voronoi tessellations in hyperbolic spaces studied by D'Achille, Curien, Enriquez, Lyons, and Ünel (arXiv:2303.16831). In this paper we introduce and study ideal Poisson–Voronoi tessellations for Finsler metrics on higher rank symmetric spaces. We show that ideal Poisson–Voronoi tessellations with respect to polyhedral Finsler metrics can exhibit either behavior seen in these previous works: generically they resemble the hyperbolic case, where almost surely all cells have bounded sides. We further introduce the ideal Selberg–Poisson–Voronoi tessellation of the space of positive definite symmetric matrices, which has particularly nice properties.

The homotopy types of Oka manifolds are exactly those of countable finite-dimensional CW complexes. We construct realizations as open subsets of complex Euclidean spaces with explicit dimension bounds. Other realizations have good open covers by sets biholomorphic to $\C^N$, with prescribed finite nerves, or are unions of Fatou–Bieberbach domains whose nonempty finite intersections are again Fatou–Bieberbach domains. Every nonempty Stein manifold admits a proper holomorphic embedding into an Oka open subset of some $\C^N$, and this embedding is a homotopy equivalence. We obtain holomorphic representatives of ordinary and generalized cohomology classes, and realize every countable group as the fundamental group of an Oka domain in $\C^5$. The fundamental groups of algebraically elliptic manifolds are precisely the finite groups, and every virtually abelian Kähler group is the fundamental group of a projective Oka manifold. Finally, we classify holomorphic cofibrations in Lárusson's intermediate model structures and establish factorizations over aspherical targets into proper holomorphic embeddings and holomorphic fibre bundles with Oka fibre.

Recent multimodal deep search agents have advanced tool use and visual content retrieval, yet they still struggle to convert intermediate visual evidence into actionable guidance for subsequent search, limiting their effectiveness in long-horizon settings. To address this gap, we present Neos, a multimodal deep search framework that learns both how to acquire evidence and how to use it to guide subsequent decisions. To make this coupling learnable, Neos constructs training tasks with intermediate visual dependencies while filtering out textual shortcuts, and introduces a lightweight, trainable Multimodal Evidence Interface (MEI) that transforms retrieved content into goal-conditioned observations. At its core, an agent-environment co-evolution procedure treats MEI as a co-adaptive component of this loop. Specifically, reference-guided credit assignment uses verified successful trajectories to localize critical policy errors while preserving credit for useful decisions in failed searches, whereas rollout feedback selects informative local evidence-processing interactions for MEI supervised fine-tuning (SFT). By alternating policy reinforcement learning (RL) with MEI SFT, Neos allows the policy's evolving evidence requests to shape MEI training and the adapted MEI to reshape subsequent observations, closing the loop between evidence acquisition and evidence-conditioned action. Across ten benchmarks, Neos-9B and Neos-35B demonstrate leading performance among specialized multimodal deep-search agents.

Audio description (AD) aims to enhance the accessibility of visual media for blind and low-vision (BLV) persons. However, in higher education, where slide-based lectures are a default mode of teaching, lecture AD remains largely underexplored. The main reason is that existing production guidelines, evaluation metrics, and benchmark datasets are oriented toward entertainment content and do not transfer to lectures, where ADs must complement the lecturer's speech, resolve deictic references, and fit within narrow pauses. In this work, we propose AD4Edu, consisting of an AD benchmark, AD4Edu-Bench, and a rule-diagnosable evaluation metric, AD4Edu-Eval, for generating and evaluating ADs in higher education. Specifically, we consolidated established accessibility guidance, including WCAG 2.2, the DCMP Description Key, and NCAM's STEM description practices. We further collected silver reference ADs, judge-ordered preference pairs, and pairwise preferences from four BLV raters, and post-trained models with direct preference optimization (DPO) for detailed analysis. Our findings indicate that while the metric reliably detects violations of the standard, ADs that comply with it do not necessarily fulfill the subjective preferences of BLV persons, most notably on description length, underscoring the necessity of active participation of BLV persons in co-designing inclusive AI solutions.

Let $G$ be a semisimple Lie group that is not isogenous to the isometry group of real or complex hyperbolic $n$-space. We show that the rational rank is a profinite invariant for irreducible lattices in $G$, unless $G$ has type $D_{2n}$ which forms a genuine exception. In particular, profinitely isomorphic lattices in $G$ of type $D_{2n}$ can have differing virtual cohomological dimension. As another corollary, uniformity is a profinite property for irreducible lattices in $G$ outside type $D_4$. Assuming CSP, the existence of triality symmetries gives uniform and non-uniform profinitely isomorphic lattices in type $D_4$.

While language models provide powerful mechanisms for evaluating textual predictability, it remains unclear whether the relative extents of LM surprisal remain stable across models. This study presents empirical evidence of the invariance of ranking based on cross-entropy related measures (e.g. mean, variance of each data-point), across four distinct small scale transformer-based models, for a given dataset of text sources. We consider two stylistic environments: a corpus of contemporary poetry and a corpus of academic research papers. We demonstrate that mean and variance of cross-entropy maintain robust cross-model rank correlation ($\rho > 0.90$) across both regimes. To test whether this agreement merely reflects generic differences in textual complexity, we compare it against six parameter-free compression baselines. The extent of agreement among language models is found to be substantially higher than that among model-baselines, for most of the measures. Geometric analysis of the structurally constrained academic regime reveals that higher-order moments are sensitive to specific models, although strong agreement persists. The findings suggest that relative extents of textual unpredictability may be independent of the language model they are computed against.

Tool-using LLM agents are increasingly used to automate workflows governed by procedural manuals, such as company rules for booking travel or procuring hardware. These manuals are written in natural language for humans and describe general requirements over such complex workflows. Ensuring compliance of LLM agents with these manuals is difficult, and state-of-the-art techniques use either LLM judges or manually written rules to forbid problematic tool-calls locally at runtime. This is insufficient because (i) only blocking tool calls keeps the agent clueless about how to make correct progress towards fulfilling its task, and (ii) can lead to overly conservative behavior of the agent, as tool calls might only be problematic in certain contexts. To overcome these challenges, we introduce MENTOR, which (i) additionally provides guidance for the agent to make procedural-compliant progress on the task, and (ii) ensures that this compliance is context-aware. We achieve these goals by building a novel history-dependent runtime shield that is obtained from the procedural manual via a multi-step autoformalization pipeline. At its core, MENTOR utilizes a game-based technique to ensure robustness against runtime uncertainty without over-resticting agents' behavior. This yields a formally assisted framework for procedural shielding of LLM agents that improves compliance well beyond existing methods. In particular, MENTOR lifts the compliance of shielded light-weight open-source LLMs above that of non-shielded proprietary models. Since a shield is synthesized once per manual and then serves every task, a MENTOR-shielded agent becomes a viable choice for repeatedly executed workflows, potentially reducing deployment costs when the offline construction cost is amortized over repeated tasks.

Spatial sudden death of entanglement refers to the disappearance of quantum correlations between subsystems after a certain distance called the entanglement length. Specifically, given a density matrix $\rho_{ABC}$ on a tripartite system $ABC$ with $B$ spatially separating $A$ and $C$, the entanglement length is the distance $\ell(A,C)$ after which the reduced density matrix $\rho_{AC}$, given by tracing out the $B$ subsystem, is separable. We show in the case of gapped ground states in any dimension that the entanglement length is finite on finite subsystems. Restricting to one-dimensional systems, we prove that normal matrix product states have a finite entanglement length independent of $A$ and $C$, and that a finite pure state with such uniform entanglement length has a matrix product state representation with bond dimension independent of $A$ and $C$. Consequently, in one dimension, entanglement length distinguishes matrix product states from all other pure states.

In many practical applications of Bayesian inference, it is difficult or undesirable to compute the gradient of the negative log-posterior, motivating the study of derivative-free samplers, also called zeroth-order samplers. The aim of this paper is to generalize and rigorously analyze a multiscale derivative-free sampler recently proposed in the literature by Pavliotis, Stuart and Vaes (2022). In the discrete-time setting, the multiscale sampler we study bears a strong resemblance to other recently proposed methods, such as the so-called zeroth-order perturbed stochastic gradient descent algorithm (Jin et al., 2018) and the zeroth-order Langevin Monte Carlo algorithm by Roy et al. (2022). We present error estimates with respect to all the parameters of the problem, both in the continuous-time and discrete-time settings. The error estimates are in three different metrics: strong convergence, convergence of ergodic averages, and uniform-in-time weak convergence.

Effective vocal coaching connects performance issues with corrective advice, but recordings paired with expert diagnoses and advice are scarce. Online singing communities provide a potential source of supervision, but observations and advice are scattered across unstructured comments. We introduce PitchIn, audio-language models for natural-language vocal coaching, and the PeerCoach Dataset: 14,445 instruction examples covering Focused QA, Overall Feedback, and Recognition QA. We compare target combinations on Qwen2.5-Omni and train PitchIn on Focused QA and Overall Feedback using Qwen2.5-Omni and MOSS-Music. On VocalCoachBench, the proportion of requested diagnosis-correction pairs matching an expert issue with valid advice rises from 21.3% to 61.0% for Qwen2.5-Omni and from 31.4% to 51.2% for MOSS-Music. Direct training on source comments yields only 19.8% for Qwen2.5-Omni. For Qwen, Focused QA alone reaches 65.6%, but adding Overall Feedback improves alignment with expert reviews and human preference. Both models improve expert-review alignment over pretrained baselines and are significantly preferred for usefulness (92.0% and 67.3%, respectively). Demo examples are available at https://hayeonbang.github.io/PitchIn.

Artificial intelligence faces two growing challenges: computational bottlenecks and energy inefficiency inherent to the von Neumann architecture. Neuromorphic computing offers a promising alternative, yet its immature ecosystem still relies on conventional systems for data supply, requiring efficient conversion between tensor and spike dataflow. We observe a clear mismatch, where data conversion is slow while downstream computation is fast. As data size grows quadratically with large language models, this mismatch becomes a major bottleneck. In this work, we examine the data-conversion components that appear on the critical path of training and inference in neuromorphic LLMs. The work presents FlashT2S (Flash Tensor-to-Spike): GPU-Accelerated Tensor-to-Spike Conversion Operators, which applies architecture-aware and data-centric optimizations to reduce conversion latency. Experimental results show that FlashT2S delivers an average speedup of 850.78x over the native operators across individual operators and 802.83x in realistic workloads. Additionally, compared with GPU kernels automatically generated by torch.compile, our kernels achieve 5-11x speedups. Notably, it achieves near-optimal absolute performance: large-scale conversions are reduced to tens of microseconds (10-99 us), while small-scale conversions reach single-digit microsecond levels (less than 10 us). Furthermore, its standardized output enables seamless integration across heterogeneous backends, including diverse neuromorphic hardware.

We study quasibound states of a massive charged Dirac field on the Einstein–Euler–Heisenberg (EEH) black hole, the extension of Reissner–Nordström (RN) by the leading weak-field vacuum-polarization term of QED. We derive the background to first order in the Euler–Heisenberg coupling $a$. We show that it admits a nondegenerate horizon at $Q=M$ and an extremal charge $Q_{\rm ext}=M[1+a/(40M^2)+\mathcal{O}(a^2)]>M$, with horizon data that depend on $\sqrt a$ near extremality. From the separated radial equations, we derive an identity that excludes real-frequency quasibound states on any background of this class, so damped branches cannot become unstable. Since the EEH corrections first enter at $\mathcal{O}(r^{-5})$ in the electrostatic potential, the trapping condition $M\mu^2-qQ\omega_R>0$ and the hydrogenic spectrum coincide with RN, and no fine-structure-order splitting arises. Two-sided shooting with a Muller root finder shows that the real frequencies shift only by near-zone amounts that are linear in $a$ and strongly suppressed with angular momentum. The damping rates respond in a branch-dependent way. At $Q/M=0.99$ the $\ell=0$ ground state becomes slightly longer-lived, while the $(n_r,j,\ell,\lambda)=(0,\tfrac32,1,-2)$ state decays up to $80\%$ faster for $a/M^2\le1$, and near the trapping threshold the damping rate increases by up to $60\%$. Switching on the metric and potential corrections separately, together with a WKB analysis of the inner barrier, identifies the mechanisms: Coulomb screening always increases leakage, while the metric correction controls the sign and size of the $\ell$-dependent effect. Characteristic time-domain evolutions with Richardson-extrapolated matrix-pencil extraction reproduce the frequency-domain shifts.

While terrestrial colliders and fixed-target cosmic-ray interactions are bounded at center-of-mass (CM) energies $\sqrt{s} \lesssim \mathcal{O}(10^5)\,\text{GeV}$, we demonstrate that bound sub-parsec binary blazars form broad-band, beam-beam hadronic “Cosmic Colliders”, overcoming such limitation. Driven by modest jet misalignments ($\theta_{\rm mis} \sim 1^\circ - 6^\circ$), co-streaming proton beams collide at CM energies reaching $\sqrt{s} \sim 10^{10}\,\text{GeV}$ - six orders of magnitude above the Large Hadron Collider. Operating in a strictly optically thin regime ($\tau_{pp} \ll 1$), the system performs a continuous CM energy scan, probing asymptotic QCD scaling, gluon saturation, and Froissart unitarity, while opening a possible high-sensitivity window into physics beyond the Standard Model. Secondary PeV-EeV neutrinos and sub-TeV $\gamma$-rays fall directly within reach of current and planned observatories such as LHAASO, CTAO, IceCube, and GRAND, establishing binary blazars as laboratories for physics at the ultimate energy limit.

We study whether an unnormalized graph Laplace operator can determine both the geometry of a manifold and its sampling density. On a fixed compact smooth manifold without boundary, we use a known positive, continuous, strictly decreasing radial kernel at a fixed positive bandwidth. For positive sampling densities, the operator built from geodesic distances determines the Riemannian metric and identifies the density up to equality almost everywhere. When the metric is induced by a Euclidean embedding, the operator built from ambient distances also determines the embedding up to rotations, reflections and translations, and the density. On compact parameter classes with positive continuous densities, minimizing the distance between empirical and popula- tion operators gives estimates that converge to the true parameters under independent, identically distributed sampling. This recovery result assumes that the empirical operator can be evaluated at arbitrary points of the manifold.

We present and analyze a spatially and temporally adaptive high-order method for the one-dimensional Richards equation describing vertical water transport in unsaturated soil, with an emphasis on resolving sharp wetting fronts while satisfying a fully discrete water mass balance. The spatial discretization uses an arbitrary-order local discontinuous Galerkin spectral-element formulation with collocated Legendre-Gauss-Lobatto quadrature, and the time integration is based on a high-order Rosenbrock-Wanner method. We focus primarily on the mixed formulation of the Richards equation, which advances water content as the conserved variable while evaluating the nonlinear diffusion operator in terms of pressure head; the resulting fully discrete scheme satisfies the water mass balance in exact arithmetic, independent of spatial resolution and time step size, including under spatial and temporal adaptivity. In contrast, the pressure-head formulation advances pressure head through a nonlinear capacity function and satisfies the corresponding conservation identity only at the semi-discrete level. Numerical experiments examine convergence through spatial and temporal refinement studies, with mass conservation assessed under spatial and temporal adaptivity using the solver flux output method (SFOM). The SFOM results confirm machine-precision mass conservation for the mixed formulation, whereas the pressure-head formulation exhibits mass balance errors above roundoff. Moreover, in the temporal refinement studies, the mixed formulation yields more accurate solutions than the pressure-head formulation for a prescribed time step size. The combination of high-order accuracy, spatial and temporal adaptivity, and fully discrete mass conservation makes the proposed mixed-form discontinuous spectral-element method a promising approach for simulating unsaturated flow in next-generation land models.