Optics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Friday 11 September
Transverse light localization of Anderson type through disorders has made intriguing fundamental topics. Toward this, we found that in the disordered porous Si (PSi) a nearly "zero reflection" (from 450 to 900 nm) can be observed. The localization length versus disorder parameter is experimentally visualized and calculated based on a self consistent theory and solution of stochastic Helmholtz equation for electric fields traveling in a transverse disordered media with the results determined to be in a reasonable agreement with experiments. Based on these calculations for different pore diameters in a constant order parameter, diffusion length can vary from 260 nm to 2.4\mu m, accounting as controllability of such structure for further applications in photonic devices. Reflection from ordered PSi structure are evaluated that shows no similar responses to those seen in disordered PSi.
Representation Theory
On Donkin's Tilting Module Conjecture IV: New Frontiers
Let $G$ be a simple, simply connected algebraic group scheme defined over $\mathbb{F}_{p}$, and let $G_{r}$ be the $r$th Frobenius kernel. Donkin's famous Tilting Module Conjecture purports that a given indecomposable injective $G_{r}$-module can be realized as the restriction of a specific tilting module. The conjecture was first stated in 1990 and withstood proof for nearly 30 years. The authors discovered a counterexample in 2019. This paper provides some indication about the rich mathematics surrounding the Tilting Module Conjecture that also discusses an older conjecture by Humphreys and Verma. New versions of the Tilting Module Conjectures are formulated given multiple families of new counterexamples. Finally, through a new calculation, the authors present further evidence that the Tilting Module Conjecture should hold for reductive algebraic groups with an underlying root system of Type $\mathrm{A}_{n}$ for all fields of characteristic $p > 0$.
Read the full abstract
Geophysics
ADEPTS: An auto-differentiable framework for time-dependent nonlinear thermo-chemical mantle convection inversion
Time-dependent mantle-dynamics inversion must address the high dimensionality of the initial state, nonlinear rheology, and gradient propagation through long-term thermo-mechanical evolution. We develop ADEPTS, a two-dimensional staggered-grid finite-difference framework for mantle-dynamics inversion based on automatic differentiation. The forward model solves incompressible Stokes flow, temperature advection-diffusion, and compositional advection with temperature- and strain-rate-dependent viscosity and plastic yielding. For the nonlinear Stokes system, we compare two gradient strategies: unrolled differentiation through a fixed number of Picard iterations and implicit differentiation of the converged discrete residual equations. Numerical experiments show that unrolled differentiation remains stable even when the nonlinear solve is not fully converged, whereas implicit differentiation requires sufficiently accurate nonlinear solutions; otherwise gradient consistency and optimization convergence deteriorate. With sufficiently converged solves, implicit differentiation recovers accurate gradients and reconstruction quality comparable to unrolled differentiation. Joint thermo-chemical twin experiments show that ADEPTS can simultaneously recover a high-dimensional initial temperature field and low-dimensional physical parameters, including compositional density, reference viscosity, and stress exponent, while fitting final-time temperature, surface horizontal velocity, and surface normal stress. These results demonstrate the feasibility of differentiable time-dependent mantle-dynamics inversion and clarify the different convergence requirements of unrolled and implicit differentiation.
Read the full abstract
Computation and Language
A Hybrid Hierarchical 1D-CNN-BiLSTM Framework for Extractive Summarization of Biomedical and Clinical Text
Large language models have made abstractive summarization remarkably fluent, but fluency has come at a cost. Models trained to generate freely over long documents may hallucinate, producing summaries that read well but drift from what the source actually says. This risk is especially serious in biomedical and clinical text, where fabricated details can have serious consequences. We address this by removing free-form generation from the pipeline rather than trying to constrain it. Our Hybrid Hierarchical CNN-LSTM Summarizer reframes summarization as extractive sentence selection: a stacked, multi-kernel convolutional network progressively refines contextual representations, while a bidirectional LSTM models long-range dependencies across the document. A lightweight scoring head predicts sentence importance using binary cross-entropy supervision. For PubMed, a document-adaptive mean-plus-standard-deviation threshold with a Top-3 fallback determines extraction; clinical evaluations instead use a fixed top-30% sentence-selection policy. Because output sentences are copied verbatim, the framework prevents source-external generation, although omissions and contextual inaccuracies remain possible. Evaluated on PubMed and transferred without clinical fine-tuning to MIMIC-CXR and MIMIC-IV BHC, the model demonstrates competitive extractive summarization performance, with threshold and component ablations supporting adaptive extraction and the combined CNN–BiLSTM architecture. Clinical factuality analysis further highlights limitations in preserving important findings and negations. These results suggest that structural constraints offer a promising direction for source-grounded biomedical summarization, while emphasizing the importance of clinical completeness and contextual fidelity.
Read the full abstract
Optimization and Control
Fast Relax-and-Round Unit Commitment with Topological Constraints
Recent developments in the US knowledge economy have created a significant growth in datacenter loads, with two major consequences for power generation. First, to compensate for growth in load, datacenters are encouraged to bring their own generating units. Second, in search of the remaining pools of dispatchable generation, utilities are increasingly turning to subtransmission and distribution level generating assets. Coupled with increasing loads and the resulting tighter grid conditions, both trends are likely to create a need to commit a large number of localized generating units under grid constraints. We propose an extension of Relax-and- Round Unit Commitment (RRUC) that is capable of committing generating units for larger problems faster than conventional methods, while staying within intertemporal and spatial MVA constraints. We demonstrate the performance of RRUC using synthetic congestible test systems ranging from 100 to 20,000 buses. RRUC consistently finds low cost solutions, independent of the problem size, and its run time increases sub-quadratically in the number of buses. RRUC can solve the 100 bus system in less than a second and the 20,000 bus system 7 minutes. In contrast, a leading state of the art solver cannot find a feasible solution to the 100 bus system in 15 minutes.
Read the full abstract
Human-Computer Interaction
Exploring K-12 Teachers' Perceptions of Students' Relationships with AI Companions: Boundaries, Intervention Strategies, and Design Implications
K-12 students increasingly form relationships with AI companions. Schools face growing expectations to teach AI literacy, yet existing frameworks treat AI as a tool rather than a relationship, and little is known about how teachers understand and act on students' relational use of AI. We conducted scenario-based interviews with 33 US K-12 teachers. Teachers welcomed academic companions but worried that intimate companions remove the developmental friction through which students learn to sustain human relationships. Teachers drew the boundaries of their jurisdiction by setting and observable wellbeing: within it they taught, talked, and watched; beyond it they positioned themselves as the adults best placed to notice and connect students with support. They envisioned AI companion literacy as shared work across the jurisdictions of counselors, parents, platforms, and policymakers, spiraling across grade levels. We introduce AI companion literacy as an extension of AI literacy and discuss implications for K-12 AI education.
Read the full abstract
Cryptography and Security
BadEngram: Backdoor Attack on Gated Memory Components in LLMs
To expand open-weight models' capacity without proportionally increasing computation, recent architectures incorporate gated parametric memory that retrieves learned values and injects them into intermediate representations. One representative design is Engram, which combines deterministic n-gram lookup with context-dependent gating over large learned memory tables. Despite these efficiency benefits, such modules create a distinct attack surface: their parameters can be modified independently of the backbone while directly shaping its computation. We introduce BadEngram, a post-training attack that exploits this surface to implant persistent, trigger-dependent behavior while leaving conventional backbone weights and the execution graph unchanged. We first establish the attack's feasibility and causally characterize its mechanism in a controlled Engram model, where BadEngram achieves 96.6% ASR on triggered inputs while limiting false activation on matched trigger-free inputs to 0.1% and preserving 99.6% clean accuracy. Replacing the retrieved memory values with their clean counterparts or closing the memory gates reduces ASR to at most 0.32%, confirming that the backdoor is expressed through the gated-memory pathway. We then test whether this vulnerability extends to production scale in Qwen3.8-Flash-Next's native Per-Layer Embedding subsystem. Using independently trained checkpoints for the two benchmarks, BadEngram achieves 47.8% ASR on HarmBench and 64.0% on AdvBench, while dormant-condition ASR remains 0.9% and 0.0%, respectively. These results identify native gated-memory parameters as a security-critical part of the model whose integrity cannot be inferred from an unchanged backbone.
Read the full abstract
Machine Learning
On the Potential of Multi-Task Learning in Predictive Process Monitoring
Predictive Process Monitoring (PPM) forecasts how ongoing organizational processes unfold, enabling information systems to move beyond execution support toward proactive analysis and monitoring. Although deep learning has improved prediction accuracy in PPM, most approaches follow a single-task learning (STL) setup, training a separate model per task. This increases maintenance effort and overlooks potential synergies. Multi-task learning (MTL), which jointly learns multiple prediction targets in one model, offers a promising alternative, yet its effectiveness in PPM remains underexplored. It remains unclear whether and under which settings MTL improves upon STL, which prediction tasks benefit most from joint learning, which task combinations are particularly synergistic, and if and how tasks should be balanced. To fill this gap, we present the first comprehensive empirical study of MTL for PPM, evaluating a variety of task combinations, neural architectures, and optimization methods. Overall, our results position MTL as a strong paradigm for PPM: we see substantial improvements in next-activity prediction and inherent mitigation of class imbalance using MTL, while task balancing is especially critical under low-capacity models.
Read the full abstract
Combinatorics
Edge-addition monotonicity of positive p-energy fails for every p >= 1
At a 2021 AIM workshop, Guo conjectured that the positive square energy s+ = E+_2 should inherit the familiar edge-addition monotonicity of the spectral radius, rho(G + uv) >= rho(G). That conjecture was subsequently shown to fail at p = 2. Tang, Liu, and Wang then introduced positive p-energy, proved nonmonotonicity for every 1 <= p < 3, and in version 3 of their preprint (26 March 2025) explicitly conjectured that monotonicity should hold for p >= 3. We disprove this conjectured high-exponent extension completely: for every real p > 2 there are infinitely many connected graphs G and nonedges uv such that E+_p(G + uv) < E+_p(G). Together with the Tang-Liu-Wang counterexamples below 3, this shows that no exponent p >= 1 restores the spectral-radius-style monotonicity: positive p-energy can decrease under the addition of an edge for every real p >= 1. The construction is a chain of clique blocks joined by regular bipartite graphs. An equitable quotient converges to Q = I + cA(P_k). For noninteger p, a binomial-series sign argument for a fractional power of I - cA(P_k) gives the required negative endpoint entry. At integer exponents, choosing c across the first spectral threshold leaves exactly one negative eigenvalue, and path locality forces the positive spectral contribution to have negative sign. We also give a fully rational 38-vertex certificate at p = 4 and determine the complete failure interval of a fixed 17-vertex counterexample at p = 3.
Read the full abstract
Artificial Intelligence
Grounded Adjudication of Variations across Extracted TimeLines (GAVEL): Comparing Clinical Timelines Against Their Case Reports
Existing pipelines for clinical timeline extraction from case reports are evaluated using an expert reference and are limited by imperfect reference annotations and imprecise event alignment. We developed GAVEL, an LLM judge protocol that compares two timelines with the case report and returns a discrepancy type, verdict, and report passage for each difference. We evaluated the event matcher, reviewed 2,738 findings from GPT5.6sol and DeepSeek V3.2, ranked six LLM extractors and two human annotators, and tested GAVEL guided merging. True match rates were 60% immediately below and 48% immediately above the 0.10 cutoff. Manual review confirmed 89.4% and 88.6% of findings. Across 126 reports, merged timelines were preferred in 77.0% of comparisons (95% CI, 69.8 to 84.1%) and reduced discrepancies attributed to the evaluated timeline from 7.63 to 0.85 per report. GAVEL supports report-based comparison and revision without treating either timeline as ground truth.
Read the full abstract
Other Condensed Matter
Magnon Theory of Domain Wall Wavefronts and the Ballistic Diffusive Crossover in the Classical Anisotropic Landau Lifshitz Spin Chain
We study the far-from-equilibrium dynamics of domain-wall initial states in the integrable lattice Landau–Lifshitz spin chain. Building on the linear spin-wave description of domain-wall spreading, we derive the long-wavelength magnon dispersion directly from the microscopic integrable Hamiltonian and use it to characterize the propagation and broadening of the resulting wavefronts. The dispersion interpolates between a quadratic form at the isotropic point and a linear form in the easy-plane regime. In the isotropic limit, the quadratic dispersion leads to diffusive broadening, whereas in the easy-plane regime the domain-wall front propagates ballistically. Keeping the leading nonlinear correction to the linear dispersion gives a cubic dispersive term, which produces a characteristic $t^{1/3}$ broadening of the ballistic front. We further show that the sign of this cubic correction changes at a critical anisotropy, leading to a reversal of the dispersive wake from trailing to advancing. The resulting critical point occurs at $\Delta_c=1/7$, corresponding to $\gamma_c\simeq 1.047$. Numerical simulations of the full nonlinear spin dynamics confirm the analytical scaling and demonstrate the coexistence of the amplitude-independent spin-wave front with amplitude-dependent nonlinear soliton propagation.
Read the full abstract
Fluid Dynamics
A conservative Godunov-type mesh-free hydrodynamics scheme for weakly compressible multiphase flows
We present a Godunov-type mesh-free hydrodynamics scheme for the numerical simulation of weakly compressible single- and multiphase flows. The method employs a consistent second-order spatial reconstruction of the primitive variables at the inter-particle interface, applies slope limiting, and solves the moving Riemann problem using a subsonic HLLC approximate Riemann solver. Particles are advected with a quasi-Lagrangian material velocity incorporating an inherent self-relaxation mechanism that suppresses the formation of anisotropic particle distributions. Near material interfaces, one-sided renormalized gradients are used to handle density discontinuities and slope limiting is applied for each phase independently. The Lagrangian character of material interfaces is preserved by constraining the relaxation mechanism to act tangentially to the interface. For same-phase particle interactions, the Riemann problem is centered at the arithmetic mean of the two material velocities. For cross-phase interactions, it is centered at the contact-wave velocity. This choice naturally eliminates spurious mass fluxes across phase boundaries. The methodology is validated on a set of challenging benchmarks, including the inviscid Taylor-Green vortex, material interface advection, surface gravity waves near hydrostatic equilibrium, the Kelvin-Helmholtz instability, and the Rayleigh-Taylor instability under gravitational stratification.
Read the full abstract
Soft Condensed Matter
Local Strain-Dependent Anisotropy in Fibrous Networks
Cells in connective tissues reside within the extracellular matrix (ECM), which consists of a fibrous mesh that exhibits non-linear strain-stiffening behavior, driven by a transition from bending-to-stretching-dominated deformation. While bulk rheology captures macroscopic mechanical properties, cells actively sense and respond to local microscale heterogeneities and stiffness anisotropy in their environment. Characterizing ECM micromechanics is therefore essential for understanding the mechanical cues experienced by cells. This study quantifies local stiffness anisotropy in stretched fibrous gels by combining experimental and numerical approaches. Experimentally, we utilized optical tweezers microrheology to measure local stiffness in fibrin gels subjected to uniaxial stretch. The gels demonstrated gradual local stiffening along both the tensile and perpendicular axes, with a more profound increase along the tensile axis, resulting in local anisotropy. To investigate the physical parameters driving this phenomenon, we developed a 3D finite element model of a discrete random fiber network, successfully replicating the experimental local stiffening and anisotropy. Numerical analysis further revealed that within the sub-isostatic region, both fiber thickness and network connectivity strongly influence local anisotropy: slender fibers and higher connectivity amplify the anisotropy by up to an order of magnitude. This contributes to the formation of a highly anisotropic local environment, thereby playing a significant role in directing mechanically driven biological processes, such as cell migration and durotaxis. Our simulations also indicate that local micromechanical responses may differ from the material's global stiffening behaviors, highlighting the need for characterization at the microscopic scale.
Read the full abstract
Statistics Theory
Generative bootstrap processes
We study generative bootstrap processes obtained by resampling from fitted generative distributions. We establish necessary and sufficient conditions for their conditional weak convergence to the $P$-Brownian bridge, formulated in terms of finite-dimensional conditional weak convergence and conditional asymptotic equicontinuity. We further provide general sufficient conditions and a Giné-Zinn-type characterization relating this convergence to the $P$-Donsker property. As applications, we verify these conditions for prominent classes of generative models, including triangular normalizing flows, flow matching, score-based diffusion models, and Wasserstein generative adversarial networks.
Read the full abstract
Artificial Intelligence
OrchSLM: Probing the Dynamics of Small Language Model Orchestration
Although large language models (LLMs) have demonstrated remarkable capabilities, their reliance on cloud-scale infrastructure poses fundamental challenges for deployment in agentic pipelines, including latency, privacy, connectivity, and substantial computational cost. Small language models (SLMs) offer a compelling alternative: recent studies suggest that many repetitive and narrowly scoped subtasks in agentic workloads may be better served by specialized SLMs than by monolithic LLMs. However, the limited capacity and context windows of SLMs can constrain long-horizon reasoning and interaction-heavy orchestration strategies such as iterative verification and debate. This motivates a complementary, non-interactive paradigm in which heterogeneous SLMs independently generate candidate solutions and a router orchestrates their cached samples without further model interaction. To further understand the mechanisms of such orchestration, we introduce OrchSLM, a routing framework that unifies existing non-interactive orchestration methods and exposes their underlying design choices as controllable parameters. Using OrchSLM as a systematic probe, we reveal how orchestration behavior emerges from diverse knobs, including the task structure, model-pool composition, and multi-agent consensus.
Read the full abstract
Optimization and Control
Inverse Learning of the Altruism and Cost Level in Mixed-Individual Mean Field Games
Understanding how humans respond to incentives, both at the individual and collective levels, is crucial to the design of effective policies. Within the continuous-time stochastic framework for large interacting populations, mean field games (MFGs) model populations of non-cooperative agents, whereas mean field control (MFC) describes the fully cooperative benchmark, interpreted in our setting as fully altruistic behavior. Mixed-individual MFGs interpolate between these two extremes through a parameter governing the degree of altruism. A central challenge for regulators and policymakers, however, is that intrinsic altruism levels and other private structural parameters, such as individual labor costs, are typically unobservable. To address this challenge, we develop an inverse learning framework for mixed-individual MFGs. Our approach enables the recovery of (latent) altruism and labor cost levels from noisy observations, with experiments demonstrating the feasibility and accuracy of our method. These findings underscore the promise of inverse MFG methodologies for uncovering latent preference structures in large populations, with important implications for incentive design, empirical behavioral modeling, and data-driven policy analysis.
Read the full abstract
Materials Science
Symmetry- and Property-Aware Crystal Generation with Reinforcement Learning for Inverse Materials Design
The inverse design of crystalline materials ultimately seeks structures with desired physical properties. However, for many functional responses, a favorable numerical value is meaningful only when supported by the symmetry of the underlying crystal. Without the appropriate crystallographic constraints, an apparent response may be ill defined, accidental, or not symmetry protected. Here we introduce SPARC, a symmetry- and property-aware reinforcement learning framework that optimizes physical objectives while preserving the structural conditions required for their realization. We demonstrate SPARC on two complementary tasks. The first targets strong uniaxial dielectric anisotropy, a tensorial response that is well defined only within appropriate crystal classes. The second maximizes the spectroscopic limited maximum efficiency, a scalar device-level objective without a prescribed symmetry class, allowing the framework to identify favorable crystallographic motifs. These results show that symmetry is not merely an additional design constraint, but a physical foundation for generating candidates with meaningful, robust, and realizable functional properties.
Read the full abstract
Materials Science
Particle Inertia-Driven Pore Formation over Material Property Effects in Laser Powder-blown Directed Energy Deposition
Laser Powder-blown Directed Energy Deposition (LP-DED) offers flexibility for process and materials and high productivity (~5 kg/h), but process-induced pores often compromise mechanical properties. This study utilizes in-situ X-ray synchrotron imaging to compare pore formation mechanisms in Ti-6Al-4V (Ti64) and stainless steel 316L (SS316L), focusing on the interplay between particle dynamics and thermophysical properties. Four distinct pore formation mechanisms were identified, with most large pores originating from the closure of cavities formed behind incident particles impinging on the melt pool. High Weber number (We >> 1) governs this behavior, indicating that particle inertia, rather than thermophysical property differences, is the primary driver of large pore formation. The study demonstrates that increased energy density leads to larger melt pool volumes, facilitating deeper particle penetration. This greater penetration depth directly correlates with increased pore diameters. While thermophysical properties secondarily influence pore-formation frequency and cavity symmetry, particle inertia remains the dominant factor. These findings provide a physically grounded basis for understanding and controlling porosity in powder-blown DED processes.
Read the full abstract
Artificial Intelligence
Governing at Machine Speed: An Adaptive Intelligence Architecture for Real-Time AI Policy Enforcement
Enterprise AI adoption has reached 78% of organizations globally, yet the infrastructure to govern that adoption has not kept pace. This paper identifies and characterizes the attestation deficit, a structural condition in which organizations maintain governance policies but cannot produce auditable, tamper-evident evidence of enforcement within regulatory timelines. Drawing on empirical data from the Stanford 2026 AI Index Report (362 documented incidents), the IBM/Ponemon 2026 Cost of a Data Breach study (USD 4.99M average cost, 92% lacking access controls), and the EY/AIUC-1 Consortium survey (38% end-to-end monitoring, 17% agent-to-agent coverage), this paper demonstrates that the governance failure is organizational and architectural rather than technical. To address this deficit, we propose AGIL (Adaptive Governance Intelligence Layer), a conceptual five-layer architecture designed to use machine learning for real-time AI governance enforcement. The proposed layers include: (1) Autonomous Discovery for shadow AI detection via behavioral fingerprinting, (2) Behavioral Risk Classification unifying security, hallucination, privacy, and accountability scoring, (3) a Policy Enforcement Gateway for inline permit/deny/modify decisions at sub-100ms latency, (4) a Continuous Attestation Engine generating tamper-evident audit trails as a byproduct of enforcement, and (5) Adaptive Policy Intelligence for ML-driven policy evolution across jurisdictions. AGIL is presented as a theoretical framework and architectural proposal; empirical validation through controlled deployment remains a direction for future work.
Read the full abstract
Materials Science
Cracking the case: fluctuations enhance ductility in refractory alloys
Refractory body-centered cubic (BCC) alloys are attractive candidates for structural applications at extreme temperatures, yet combining room-temperature ductility with high-temperature strength remains the unsolved challenge. Ductility in crystals requires that dislocations emit from a sharp crack tip before brittle cleavage, but continuum theories that treat disordered alloys as chemically homogeneous incorrectly predict brittleness for many experimentally ductile alloys. Here we show that atomic-scale stress fluctuations in disordered alloys create an additional local stress intensity at the crack tip, enabling dislocation loop nucleation below the cleavage threshold. Accounting for these fluctuations yields a local ductility criterion where alloys deemed brittle by conventional fracture mechanics can be intrinsically ductile. Atomistic simulations with machine-learned interatomic potentials and an analytic fracture mechanics model correctly predict composition-driven brittle-to-ductile transitions in binary and ternary Mo-Nb-Ti alloys, in Nb-Ti alloys at 4 K, and in several commercial BCC alloys at room temperature. Guided by this criterion, we predict, fabricate, and test the Hf$_{15}$Mo$_{15}$Nb$_{32}$Ti$_{38}$ alloy, confirming its room-temperature ductility. This fluctuation-driven mechanism provides a quantitative basis for designing ductile multicomponent BCC alloys.
Read the full abstract
Combinatorics
First-Derivative Chromatic Symmetric Reconstruction For Proper Trees
Richard Stanley asked whether every tree $T$ is determined up to isomorphism by its chromatic symmetric function $X_T$. We approach this open problem by regarding $X_T$ as a polynomial in the power-sum symmetric functions $p_1, p_2, \dots$ and studying the invariant $\Phi_T = \left.\frac{\partial {X_T}}{\partial p_1}\right\rvert_{p_1 = 0}$. We prove that $\Phi_T$ distinguishes, among proper trees, every proper tree in which any non-leaf vertex with two or more non-leaf neighbors has a number of leaf neighbors occurring at no other non-leaf vertex. We give an equivalent formulation of Stanley's question in terms of $\Phi_T$. Finally, we show that counts of spanning forests with at most $t$ edges cannot distinguish all $k$-vertex trees when $t<\left\lfloor k/2\right\rfloor$, for every $k\ge4$.
Read the full abstract
Artificial Intelligence
Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent Failures
The increasing deployment of AI agents in long-horizon tasks yields massive execution logs. Diagnosing failures within these records is crucial for reliability, as it transforms outcome-level signals into actionable interventions. The sheer scale of the data renders human review impractical, driving the need for automated root-cause attribution (RCA). However, automated RCA methods using LLMs suffer from low diagnostic accuracy, especially as execution traces grow larger. They struggle because relevant information is often sparse, distributed across distant actions, and disconnected from the visible failure, reducing root-cause attribution to a massive search problem. Existing RCA methods typically rely on one-shot LLM judgments to diagnose failures from execution traces. While effective for shorter trajectories, these judges tend to settle on a plausible diagnosis early, leaving critical evidence in longer traces unexamined. We introduce Continual Search, an iterative framework that nudges the judge, over successive turns, to keep searching for unresolved diagnostic evidence. We evaluate Continual Search across four existing RCA benchmarks. Recognizing the lack of massive execution traces in current benchmarks, we introduce MegaRCA-Mix to evaluate RCA at scale. MegaRCA-Mix provides a challenging testbed of 50 human-annotated failure trials spanning long-horizon, execution-heavy tasks. Across multiple benchmark suites and model families, Continual Search consistently improves attribution performance. On MegaRCA-Mix, for example, it improves Opus-4.8's F1 score by 29%, from $0.471$ to $0.608$. More interestingly, within the same model family, lower-tier models can even surpass their higher-tier counterparts, demonstrating that effective search supersedes raw model scale.
Read the full abstract
Optimization and Control
Matching Higher-Order Oracle Complexity for Smooth Monotone Variational Inequalities
We establish near-optimal higher-order oracle bounds for smooth monotone variational inequalities. For fixed $p\ge2$, let $F$ be monotone on a known compact convex set $X$ of diameter at most $D$, with $\operatorname{Lip}(D^{p-1}F)\le L_p$. Each feasible query returns the complete jet $(F,DF,\ldots,D^{p-1}F)$, and the goal is to find $x$ with tangent residual $\operatorname{dist}(0,F(x)+N_X(x))\le\varepsilon$. Writing $Q=L_pD^p/\varepsilon$, we improve the $\widetilde O_p(Q^{1/p})$ upper bound of Chen et al. to $\widetilde O_p(Q^{2/(3p-1)})$ via a dimension-independent deterministic algorithm that returns an explicit tangent-residual certificate. We prove a matching $\Omega_p(Q^{2/(3p-1)})$ lower bound for arbitrary adaptive deterministic algorithms and randomized algorithms with per-instance success probability at least $2/3$, without span or tensor-update restrictions. Hence the high-dimensional worst-case oracle complexity is $\widetilde\Theta_p((L_pD^p/\varepsilon)^{2/(3p-1)})$. The same method applies to smooth convex–concave minimax problems, improving the fixed-geometry accuracy exponent from $4/(3p+1)$ to $2/(3p-1)$ and matching the known lower-bound exponent.
Read the full abstract
Software Engineering
Rethinking Software Development as a Self-Adaptive Socio-Technical System
Agentic AI is substantially shortening the timescale at which development context (e.g., goals, evidence, uncertainty, and risk) evolves. However, human-agent organization across development tasks is typically treated as fixed, which is problematic because development configurations may become suboptimal as the development context changes. We propose viewing software development as a self-adaptive socio-technical system in which both the software project and the development configuration (participants, responsibilities, authority, information, and verification) adapt as the development context evolves. This yields two coupled forms of adaptation: evolving the software and reconfiguring how subsequent engineering is performed. We illustrate the idea with a proof of concept, outlining challenges and future plans.
Read the full abstract
Optics
Versatile Sleeve-and-Bulk Fabrication of Inversely Designed Nanophotonic Structures with High Pattern Transfer Fidelity
Inverse design is a powerful computational tool for generating nanophotonic devices that should outperform their traditionally designed counterparts. However, translating these theoretical gains into physical devices is challenging because the irregular and non-intuitive features that make inversely designed structures so effective also make them exceptionally difficult to fabricate, typically necessitating substantial process optimization unique to each device design. We describe a versatile two-step nanofabrication process that achieves pattern transfer fidelity in excess of 90% for each of three distinct inverse device designs with no device-specific process optimization.
Read the full abstract
Optimization and Control
Data-Driven Flutter Suppression via HODMD, DMDcand Constrained MPC
This work presents an interpretable data-to-control framework for actuator-constrained flutter suppression by combining higher-order dynamic mode decomposition (HODMD), dynamic mode decomposition with control (DMDc), and model predictive control (MPC). The primary case is a pitch-plunge typical section with a trailing-edge flap and rational unsteady aerodynamics. HODMD identifies the dominant frequency and growth rate in subcritical, near-critical, and post-critical regimes from four measured structural channels. With 2% RMS sensor noise, the near-critical growth-rate error is 1.8 x 10^-3 s^-1, substantially lower than with standard DMD without delay embedding. A five-state real reduced-order model is formed from the HODMD subspace, and its flap-input matrix is estimated from a small-amplitude PRBS record using DMDc. For an independent chirp input, the normalized error over all channels is 2.4 x 10^-3. The constrained MPC stabilizes the true post-critical plant for an initial pitch perturbation of 8 degrees with a flap limit of 1 degree, whereas saturated LQR designed from the same model reaches the prescribed validity limit. Three supporting studies assess transferability. SU2 simulations show HODMD frequency recovery consistent with the available spectral resolution; OpenFOAM dynamic-mesh calculations show nearly linear articulated-flap moment authority from 2 to 8 degrees with frequency-dependent gain; and a two-way SU2 Python-FSI case demonstrates bounded disturbance rejection using a proportional-derivative baseline. These studies support modal identification, actuator authority, and closed-loop feasibility, but are not presented as CFD-level MPC validation. The framework provides a reproducible route from early-time measurements to constrained aeroelastic control.
Read the full abstract
Robotics
STAGE: Diagnosing Semantic Transfer at Grounded Execution in Embodied Agents
Embodied language grounding requires more than identifying the referent of an instruction: recovered semantics must also control the action an agent exposes. We study this missing link as a semantic-action gap, where instruction semantics are recoverable but weakly expressed in native continuous actions. We introduce SAT-Bench, a fixed-observation counterfactual benchmark that holds the visual scene and agent state fixed while changing only instruction semantics. On LIBERO target-name and pixel-grounded relation swaps, target recovery reaches 100.0% and 95.8%, whereas OpenVLA action sensitivity remains only 6.8% and 7.7%. The gap persists across 1,000 additional compositional and temporal/procedural counterfactuals, with overall action sensitivity of 6.1%. Hidden-state, threshold-free, cross-policy, and rollout diagnostics further support this semantic-action transfer failure. We introduce VISA, a lightweight execution-time interface that converts recovered semantics into ALLOW, DEFER, target-consistency, and verified-selection decisions. VISA reduces invalid-instruction blind execution from 92.7% to 2.8% while preserving 94.0% of normal commands, and verified selection further improves target-consistent action exposure without updating the underlying policy. Overall, embodied language evaluation should measure semantic-action transfer, not semantic parsing alone.
Read the full abstract
Artificial Intelligence
TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language Models
Timeseries multimodal large language models (TS-MLLMs) have recently begun leveraging the reasoning capabilities of large language models (LLMs) for question-answering tasks. However, these models often fail to capture dynamic temporal patterns, providing only implicit reasoning that lacks the underlying explanations critical for high-stakes applications like healthcare. While reinforcement learning (RL)-based timeseries language models aim to address this, they often fall short because they are trained on narrow, in-distribution data and struggle with out-of-distribution compositional questions. To address these challenges, we present TimeThink, a synthetic framework for eliciting compositional timeseries reasoning. Core timeseries primitives (e.g., trend, seasonality) are domain-independent and can be deterministically generated. Guided by this premise, TimeThink first designs a synthetic data generator that produces atomic and composite question-answer pairs, providing objective ground truth with reasoning traces. Building on this framework, TimeThink employs a reinforcement learning with verifiable rewards (RLVR) training strategy that encourages explicit reasoning. Unlike template-reliant methods, this approach enables the model to learn the underlying logic of composition rather than simply imitating traces. Extensive experiments show that TimeThink, trained only on synthetic data, significantly outperforms strong baselines on both synthetic and real-world benchmarks.
Read the full abstract
Quantum Physics
Reusable selected-time channels for quantum simulation of non-Markovian dynamics
Non-Markovian quantum-state diffusion (QSD) provides a microscopic trajectory description of open-system dynamics, but trajectories associated with a chosen initial state are not directly reusable as quantum channels. We develop a constructive framework that converts the QSD propagator into an input-independent Kraus channel for a fixed bath preparation and a selected evolution time. The construction incorporates reservoir memory through a finite-mode approximation of the bath covariance and yields a deterministic channel representation that can be compressed for approximate implementation with a system-ancilla circuit. For the qubit systems considered here, the construction closes exactly in vacuum and extends systematically to finite temperatures. Two-qubit examples with structured reservoirs validate the vacuum channel construction, illustrate its finite-temperature extension for a chosen initial state, and show that the complexity of the reservoir description can differ substantially from the ancilla space required for quantum realization. The framework therefore establishes a direct bridge from microscopic reservoir memory to reusable non-Markovian quantum simulation and provides a foundation for optimized hardware implementations and extensions to larger systems and multitime dynamics.
Read the full abstract
Solar and Stellar Astrophysics
SDSS-IV MaStar: Determination of Stellar Parameters Using Bayesian Averaging
We present the stellar parameters for 59,266 high-quality spectra of 24,130 unique stars from the MaNGA Stellar Library (MaStar) in the Sloan Digital Sky Survey (SDSS) Data Release 17 (DR17). The median signal-to noise ratio per pixel of the spectra is 96. We derive four stellar parameters, effective temperature (Teff), surface gravity (log g), metallicity ([M/H]), and {$\alpha$}-enhancement ratio ($\rm [\alpha/M]$), by comparing the data with BOSZ (ATLAS-9 based) and MARCS theoretical atmospheric models. We adopt a Bayesian method and use color and absolute magnitude derived from Gaia to select a subset of theoretical models for each star. We then perform full-spectrum fitting to estimate the likelihood of each model in the subset and then compute their likelihood-weighted mean parameters as the final parameters. We set stellar-parameter quality flags to facilitate the use of the derived stellar parameters. The MaStar stellar parameters derived herein span an effective temperature range of $\rm 2,600 \leq T_{eff} \leq 29,861 K$, a surface gravity range of $0 \leq \log g \leq 5.5$, a metallicity range of $-4.9 \leq \rm[M/H] \leq 1.0$, and an {$\alpha$}-abundance range of $\rm -0.98 \leq [\alpha/M] \leq 1.0$. We compare these parameters with those from APOGEE and Gaia for stars in common, finding general consistency within the uncertainties. However, some artifacts and systematic differences are present, and we discuss their potential causes. These new stellar parameters are available through the MaStar SDSS-IV DR17 value-added catalog (https://www.sdss4.org/dr17/mastar/mastar-stellar-parameters/).