Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Tuesday 15 September
Answerability on Visual Question Answering is a novel and attractive task to predict answerable scores between images and questions in multi-modal data. Existing works often utilize a binary mapping from visual question answering systems into Answerability. It does not reflect the essence of this problem. Together with our consideration of Answerability in a regression task, we propose VT-Transformer, which exploits visual and textual features through Transformer architecture. Experimental results on VizWiz 2020 dataset show the effectiveness and robustness of VT-Transformer for Answerability on Visual Question Answering when comparing with competitive baselines.
Artificial Intelligence
Query-Aware Source-Risk Triage for Retrieval-Augmented Generation
Retrieval-augmented generation (RAG) pipelines may omit a source's material relationship to the query. We study a pre-generation triage layer that treats this relationship as query dependent. The method routes canonical query families for enhanced review and assigns retrieved pages to pass, contextualize, exclude, or review. It combines a four-dimension page score, rank-discounted family aggregation, intent-preserving query mutations, and a family-held-out router. A single-coded pilot of 200 real URLs supplies provisional calibration anchors; a 20,000-row scenario with synthetic domain identifiers supports controlled workload analysis. An oracle page gate defines a risk-coverage target for a future learned classifier. The evaluation shows why page-level frequency cannot substitute for family-level exposure and quantifies how calibration changes scenario activation. Annotation reliability remains unmeasured, and synthetic rankings omit real retrieval dynamics. The result is an auditable triage method and validation plan, not an estimate of deployed review workload, live-Web prevalence, or downstream answer-quality gains.
Read the full abstract
Cryptography and Security
The MAL Simulator: Cyber Operations Simulation based on Attack & Defense Graphs
We have developed the MAL Simulator, a cyber operation simulator based on the Meta Attack Language (MAL). The MAL Simulator is intended for decision-driven cyber attack and defense simulations, for system analysis and the development of automated agents. By building the simulator around an attack modeling language, it can be adapted to different target domains without modifying the source code. We used the simulator for two case studies where we trained two types of agents for automated cyber operations: a defensive agent and an offensive agent. To ground the experiments, we base the models in data collected from an emulated network implemented in the cyber range CRATE. We found that the trained attacker policy could reach the designated targets more efficiently than the compared search methods, and that the trained defender agent induced lower costs than a naive heuristic agent under noisy alert conditions. When testing the RL attacker against the RL defender, we found that the performance of the defenders dropped significantly. This emphasizes the importance of cyber attack simulators to facilitate training both offensive and defensive agents. The MAL Simulator and associated tooling is publicly available and provides common interfaces for compatibility with existing machine learning frameworks.
Read the full abstract
Statistical Methodology
Statistical Inference for Bivariate Functional Causal Discovery
Causal discovery methods aim to determine the causal direction between variables using observational data. Functional causal discovery methods rely on structural and distributional assumptions to determine directionality but typically lack statistical inference. This paper reviews the statistical guarantees of existing functional causal discovery methods in theory, software, and applied use, highlighting a key gap: the absence of a unified inferential framework that applies broadly across model classes. As a first step toward addressing this gap, we formalize a test-based approach for bivariate causal discovery by repurposing goodness-of-fit and independence tests within a hypothesis-testing framework. Because directionality is determined through two disjoint hypothesis tests, corresponding to four causal discovery outcomes, the approach provides explicit uncertainty quantification and diagnostic insight into assumption violations. Resampling is further used to estimate the rates of causal discovery outcomes, offering an additional layer of inference. We demonstrate the use and behavior of our inferential framework through simulations that vary the degree of assumption violation, as well as through real-data applications. We conclude with practical lessons and recommendations for advancing statistical guarantees in functional causal discovery.
Read the full abstract
Statistical Methodology
A Multiplicative Loss Function for Chain Ladder
Reserving models increasingly rely on loss-based estimation, where the loss function encodes the assumed error structure. Mack demonstrated this for the chain ladder, showing that the volume-weighted average estimator minimises a volume-weighted squared-error loss function that is additive in successive claim developments. This paper instead considers a multiplicative error structure and proposes the corresponding volume-weighted squared log-error loss function. Adapting Mack's distribution-free framework, we show that this loss function is minimised by the volume-weighted geometric average of the individual development ratios. This provides practitioners with an alternative estimator of development ratios for chain-ladder-based models, and a candidate loss function for machine-learning-based reserving models. We further show that the same estimator arises from two independent arguments: a stability requirement on successive ultimate loss projections, and maximum-likelihood estimation under a log-normal model. The estimator thus admits three complementary justifications: loss minimisation, reserve stability, and parametric likelihood. An out-of-sample study of 362 Schedule P company-line datasets supports the use of the proposed estimator in place of the volume-weighted average for the chain ladder, by showing that it improves predictive accuracy and reduces a slight over-prediction bias.
Read the full abstract
Information Retrieval
ReliGRec: Reliability-Oriented LLM-Based Generative Recommendation via User-Risk-Aware Prompt Routing
User behavior in real-world recommender systems is heterogeneous. While some users exhibit coherent preferences, others show abrupt interest shifts, bursty interactions, excessive repetition, or inconsistency with collaborative neighborhoods. Such deviations may arise from benign variation or manipulation, including shilling attacks, but do not alone establish malicious intent. Existing robust recommenders exploit user-risk signals through training-time reweighting or graph aggregation, whereas adapting generation to estimated user-level weak risk remains underexplored in LLM-based generative recommendation. We propose ReliGRec (Reliability-oriented Generative Recommendation), a weakly supervised framework whose name denotes its design goal rather than a supervised reliability variable. ReliGRec derives user-level weak-risk proxy labels from review-feedback signals for a subset of users and represents sequential behavior and collaborative context using a Behavior Token and temporal Graph Tokens, respectively. A Dual-View Weak-Risk Estimator fuses the representations to produce a user-level weak-risk score that selects a Simple or Cautious Prompt at inference. The Cautious Prompt is designed to encourage attention to stable, collaboratively supported evidence while reducing overreliance on isolated, short-term, or repeated interactions. The Behavior Token affects generation through weak-risk estimation and routing, whereas the aggregated Graph Token provides collaborative context for next-item Semantic ID generation. ReliGRec thus turns weak-risk estimation from an auxiliary prediction into a generation-time control signal. Experiments report competitive recommendation and weak-risk proxy-label prediction, while routing analyses characterize the recommendation-quality and inference-cost behavior of weak-risk-guided prompting.
Read the full abstract
Statistical Mechanics
Diffusivity in Dissipative Quantum Transport from an Exactly Solvable Krylov Chain
The computation of transport coefficients in interacting quantum many-body systems is rarely analytically accessible. Here, we develop a new Krylov-space mechanism that makes the leading density dependent correction to Green-Kubo diffusivity exactly calculable in the strong-dissipation regime of a one-dimensional noisy spin-$\frac{1}{2}$ XXZ chain. This is done by showing that the density-polarization-dressed bond coherence generates a Krylov subspace where repeated action of the dissipator closes exactly on an explicitly identifiable operator family. Within this subspace, the dissipative dynamics is then mapped onto a self-similar semi-infinite chain with a boundary defect. Surprisingly, all Lanczos coefficients of the Krylov chain and its boundary Green's function can be determined exactly. The latter then determines the exact leading density-dependent correction to the diffusivity. Finally, we calculate analytically the full boundary-to-bulk Green's function and find that it decays exponentially along the emergent Krylov chain.
Read the full abstract
Chemical Physics
Tailored coupled cluster method with sample-based quantum diagonalization: Application to titanium-based metallocene catalytic reactions for 1-hexene production
Sample-based quantum diagonalization (SQD) is a hybrid quantum-classical method for electronic-structure calculations. We applied SQD to a titanium-based metallocene catalyst system for 1-hexene production, where free-energy differences must be predicted within 1 kcal/mol to accurately determine product selectivity. However, currently accessible active spaces on quantum computers are limited in size, leaving significant dynamical correlation effects outside the active space. To address this challenge, we combined SQD with Tailored Coupled Cluster (TCC) theory. In this SQD-TCC framework, static correlation is treated by SQD, while dynamical correlation is incorporated through CCSD and its perturbative triples extension, TCC(T). We calculated the relative energies of two transition states governing 1-hexene selectivity. While SQD alone did not yield converged relative energies within practically accessible active spaces, SQD-TCC and SQD-TCC(T) provided reasonably converged results. Importantly, the relative energy obtained from SQD-TCC(T) differed by more than 1 kcal/mol from the corresponding CCSD(T) result, demonstrating the significance of static correlation in this system and its impact on predicted selectivity. Although TCC(T) calculations based on CASCI are feasible for small active spaces, the present system requires substantially larger active spaces that are beyond the reach of CASCI. By using SQD, we were able to access active spaces impractical for classical CASCI calculations. These results demonstrate that both static and dynamical electron correlation are essential for a reliable description of this chemistry and highlight the importance of incorporating dynamical correlation into quantum-computing approaches targeting chemically accurate simulations.
Read the full abstract
Computation and Language
PunGraph: Retrieval-Enhanced Phonetic-Semantic Graph Reasoning for Pun Understanding
Puns are a challenging form of figurative language that exploit phonetic similarity and semantic ambiguity to convey multiple meanings. Although large language models (LLMs) demonstrate strong language understanding capabilities, they still struggle with pun reasoning due to limited phonetic modeling and uncontrolled end-to-end generation. We propose PunGraph, a retrieval-enhanced knowledge graph framework for pun understanding. PunGraph constructs a phonetic-semantic lexical graph using the Unisyn phonetic dictionary, IPA and G2P representations, and WordNet definitions, and retrieves candidate words or senses to constrain LLM reasoning within a structured candidate space. We further introduce WebPun, a new large-scale dataset containing 5,730 annotated heterographic and homographic puns. Experiments on SemEval-2017 and WebPun show that PunGraph consistently improves the performance of small-scale LLMs and achieves competitive results against strong proprietary models. Further analysis shows that retrieval-guided phonetic and semantic constraints effectively reduce common reasoning errors in pun interpretation, highlighting the benefits of integrating structured knowledge with LLMs. We release our code and dataset at https://github.com/ysu132/PunGraph.
Read the full abstract
Earth and Planetary Astrophysics
Multi-wavelength Constraints on Dust Dynamics and Size Evolution in Protoplanetary Disk Rings. II. Observational Implications
Spatially resolved dust rings in protoplanetary disks are widely used to infer disk and dust physics from multi-wavelength continuum observations. Their interpretation, however, often neglects grain growth and the evolution of the size distribution, limiting the connection between observed ring profiles and dust-evolution parameters. Building on a physical dust-ring model that includes coagulation and fragmentation, we develop a Bayesian inference framework that jointly incorporates radiative transfer and finite angular resolution. When applied to two rings in HD 163296 and two in LkCa 15, our framework yields gas-dependent estimates of the key dust-evolution parameters such as turbulence strength $\alpha$ and the fragmentation velocity $v_{\rm frag}$ in a self-consistent way. Most rings admit both a low-$\alpha$, low-$v_{\rm frag}$ branch with small grains, and a higher-$\alpha$, higher-$v_{\rm frag}$ branch with larger grains. Typical low-$\alpha$ branches have $\alpha\sim10^{-5}$–$10^{-4}$ and fragmentation velocities of order cm s$^{-1}$ level, whereas the higher-$\alpha$ branches reach $\alpha\sim10^{-3}$–$10^{-2}$ and fragmentation velocities of a few to $20$ m s$^{-1}$. The observed broad and wavelength-dependent profiles near the ring peaks can be reproduced by intrinsically narrow dust rings. This new framework offers a more direct route from multi-wavelength continuum data to the microphysics of dust growth and trapping—a connection that can be robustly tested with future high-resolution observations at longer wavelengths.
Read the full abstract
Combinatorics
Density and separation for augmented Zarankiewicz numbers
We study the augmented Zarankiewicz problem, in which disjoint pairs of cells are added to a binary matrix with no all-one $2\times2$ submatrix. The pairs must satisfy compatibility conditions, and the objective counts each original occupied cell and each added pair once. We show that starting with a maximum $C_4$-free matrix can lower the final optimum, answering a question of Qi, Cui, and Xu. Let ${z_A}(m,n)$ be the optimum over all $C_4$-free initial matrices, and ${z_L}(m,n)$ the optimum when the initial matrix must have the maximum number of occupied cells. As $n\to\infty$ with $n\le m=o(n^2)$, we prove \[ {z_A}(m,n)-{z_L}(m,n)\ge\left(\frac1{30}-o(1)\right)mn \] and determine the sharp second-order term: \[ {z_A}(m,n)=\frac{mn}{3}+\left(\frac1{\sqrt6}+o(1)\right)n\sqrt m. \] An explicit construction gives a separation at $m=n=1893$. We also find a sharp density threshold: when $n\to\infty$ and $m/n^2\to c>0$, the limited density ${z_L}(m,n)/(mn)$ tends to $1/3$ if and only if $c\ge1/12$. The proofs combine density and stability estimates, combinatorial constructions, and an exact polynomial certificate.
Read the full abstract
Dynamical Systems
Weighted averages and applications to sets of multiple recurrence
We introduce new techniques for determining combinatorial properties of sets of multiple recurrence by considering weighted averages with quickly growing weights. Our main result is a far-reaching generalization of Szemerédi's Theorem which additionally confirms a conjecture of Bergelson-Moreira-Richter and contains as special cases both a variant of the Polynomial Szemerédi Theorem due to Bergelson-Leibman-Lesigne and the fact that if $f$ belongs to a broad class of smooth functions and satisfies $x^{d-1}\prec f(x)\prec x^d$ for some $d\in \mathbb{N}$ then for any $\ell\in \mathbb{N}$, any invertible measure preserving system $(X,\mathscr{B},\mu,T)$, and any $A\in \mathscr{B}$ with $\mu(A)>0$, the set $\{n\in \mathbb{N}: \mu(A\cap T^{-[f(n)]}A\cap T^{-2[f(n)]}A\cap \cdots\cap T^{-\ell[f(n)]}A )>0\}$ is thick, meaning that it contains arbitrarily long intervals of natural numbers. Additionally, we formulate and prove a generalization to weighted averages of Boshernitzan's criterion for uniform distribution which we use in the proof of our main result.
Read the full abstract
Atmospheric and Oceanic Physics
Concurrent daytime and nighttime heatwaves in the late 21st century over the CORDEX-East Asia phase 2 domain using multi-GCM and multi-RCM chains
The adverse impacts of extreme heat on human health are greater when concurrent daytime and nighttime heatwaves (CDNHWs) occur than when daytime or nighttime heatwaves occur individually, due to reduced recovery time from heat exposure. This study projects future changes in CDNHWs across East Asia under RCP2.6, RCP8.5, SSP1-2.6, and SSP5-8.5 scenarios. Daily maximum and minimum temperatures are derived from 3-hourly temperatures at 25-km resolution produced by 12 GCM-RCM chains participating in CORDEX-East Asia Phase 2. During the historical period (1981-2005), the East Asian mean CDNHW occurrence period and rate from April to September are 10.9 days and 0.9 percent, respectively. By 2071-2100, they increase to 3 weeks and 3.7 percent under RCP2.6, 2 months and 20.5 percent under RCP8.5, 2 months and 15.6 percent under SSP1-2.6, and 3 months and 45.7 percent under SSP5-8.5. CDNHW intensity and spatial extent also increase substantially. The proportion of CDNHWs lasting more than 10 days increases from 0.2 percent historically to 1.2 percent, 7.2 percent, 6.1 percent, and 17.3 percent under RCP2.6, RCP8.5, SSP1-2.6, and SSP5-8.5, respectively. Particularly large increases in CDNHW occurrence and intensity are projected over Indochina, East and West China, and India. Under continued high greenhouse gas emissions, East Asia is expected to experience unprecedented heat stress as CDNHWs become substantially more frequent and intense by the end of the 21st century.
Read the full abstract
Statistical Methodology
Supervising the Chain Ladder
The chain ladder's volume-weighted pattern minimises an explicit loss function, yet is rarely booked as such. Practitioners adjust the pattern and record the final adjusted ratios. This paper treats the chain ladder's pattern selection as a supervised-learning problem. Judgement on pattern adjustments becomes a framework of defined penalties and hyperparameters on the chain ladder's loss function, treated here as an objective function in machine learning. Data weights are generalised with a decay and a power parameter for recency and volume weighting. Benchmark shaping and smoothness enter through a reference penalty and Whittaker-Henderson smoothing. The assembled objective is strictly convex and minimised by a single linear system. Each hyperparameter becomes an interpretable adjustment in its own right, declarable by judgement and categorised as an experience or a prospective adjustment. Experience adjustments can be set more objectively by a proposed training loop and a reserve validation score on held-out calendar diagonals. Further hyperparameter-based adjustments are written as almost-everywhere differentiable penalties that re-time or reshape the pattern. A worked example carries one real Schedule P triangle through an incurred and then a paid training stage, demonstrating the workflow.
Read the full abstract
Computer Vision
Which Pretext Task Transfers? Self-Supervised Pretraining Objectives for Lung Ultrasound
Self-supervised learning (SSL) can reduce the need for labelled medical images, but the choice of pretext objective remains unclear for lung ultrasound (LUS). Contrastive learning, masked reconstruction, and joint-embedding predictive architectures (JEPA) differ in the space in which their targets are defined, yet existing ultrasound studies compare them under different corpora, backbones, and evaluation protocols. We compare these three objective families using the same encoder backbone, pretraining corpus, optimisation schedule, and frozen-evaluation protocol. Encoders are pretrained on COVID-BLUeS LUS videos and evaluated with linear, $k$NN, and attentive probes at 5%, 10%, 50%, and 100% label budgets. Evaluation is performed on POCUS using patient-level five-fold cross-validation and on the independently acquired Mendeley-Uganda dataset, which is excluded from both pretraining and probe fitting. At the full label budget under linear probing, VideoMAE and V-JEPA achieve $66.5 \pm 13.1$ and $65.4 \pm 11.7$ balanced accuracy on POCUS, while MoCo achieves $42.1 \pm 1.2$. On Mendeley-Uganda, the ranking reverses: MoCo performs best at $62.7 \pm 1.0$, followed by VideoMAE at $53.8 \pm 2.8$, while V-JEPA falls near chance at $35.1 \pm 4.9$. These results show that POCUS probe accuracy alone does not identify the objective that transfers best across datasets. We also outline planned representation-level analyses to examine this reversal. Code is publicly available at https://github.com/moeinheidari7829/LUSVideoSSL.
Read the full abstract
Nuclear Theory
A general decay form and absolute nuclear stability
A nucleus is conventionally called stable when no decay has been observed. We replace this phenomenological definition with a criterion based on conservation laws. For this purpose, we formulate a general decay form (GDF) for finite charge-neutral nuclear systems that includes all possible final configurations. A nucleus is absolutely stable if no allowed configuration has lower mass. A global search over all nuclei with measured atomic masses yields 84 absolutely stable nuclei.Besides the criterion of absolute stability, we introduce a dimensionless decay-accessibility entropy SGDF as a coarse-grained measure of the relative accessibility of GDF-allowed decay channels. We rank all naturally occurring isotopes according to their maximum SGDF values, identify their favored decay modes, and estimate the corresponding partial half-lives. We further identify promising candidates for natural alpha and two-neutrino double-beta decays. The resulting framework also provides a nuclear-scale perspective on the far-future evolution of matter, in which nuclear decay drives matter toward combinations of these 84 absolutely stable nuclei before still slower processes of matter evolution. These findings extend the conventional phenomenological stable-radioactive classification to a conservation-law-based framework for nuclear stability and provide a systematic guide for future rare-decay searches.
Read the full abstract
Quantum Physics
Pretty good quantum state transfer via transcendental edge weights
We prove that if we take a rooted product of a circulant graph with universal perfect state transfer with a path of fixed length whose end edge is weighted with a transcendental number, then there is pretty good state transfer between any pair of endpoints of these paths. As a consequence, in a path with an even number of vertices with transcendental weights on the two edges incident to the endpoints, there is pretty good state transfer.
Read the full abstract
Artificial Intelligence
QueryFormer: Winning Solution for KDD Cup 2026 Tencent UniRec Challenge
Post-click conversion rate (pCVR) prediction requires jointly modeling feature interactions and sequential user behaviors. The KDD Cup 2026 Tencent UniRec Challenge calls for a unified architecture addressing both. We observe that existing unified architectures often generate query tokens—the central information hub—with projection-based multi-layer perceptrons (MLPs), without explicit token-to-query attention for refining the query side. We propose QueryFormer, centered on a stackable unified field–sequence block that bridges non-sequential multi-field features and behavioral sequences, and provide a latency-aware scaling study over view width $H$, model width, depth, data, and compute. The block generates queries through cross-attention and packs sequence queries into shared-parameter attention. QueryFormer secured 1st place in the Industrial Track, achieving an official test area under the ROC curve (AUC) of 0.83254; a modest post-competition scale-up reached 0.832713. Within our grid, $H$-scaling improves validation AUC from 0.84540 to 0.84615 and beats HyFormer at comparable budgets. Ablation identifies query generation as the largest contributor. Packed shared-parameter cross-attention keeps H=8 inference latency to only 1.89x that of H=1, positioning the bridge as an efficient stackable unified block.
Read the full abstract
Cosmology and Nongalactic Astrophysics
Probing multi-state dark matter via optical absorption lines in DESI spectra
We search for dark matter through absorption lines imprinted on the spectra of background sources whose sightlines pass through foreground dark matter halos. Such a line is a generic feature of multi-state dark matter models that couple to photons through higher-order electromagnetic moments. We focus on the minimal case of a two-state system with a magnetic dipole transition, in which the ground state absorbs a photon and makes a transition to the excited state. For transition energy $\Delta E \sim O \,(\text{eV})$, the line falls in the optical band, within reach of spectroscopic surveys such as DESI. Stacking 145,087 spectra of DESI Early Data Release objects paired with foreground GAMA galaxy groups, we find a stacked optical depth consistent with zero and place $2\sigma$ upper limits on the magnetic dipole transition cross-section. The limits are strongest at $\Delta E \sim$ 3-5 eV for a few MeV dark matter mass in the parameter space that is inaccessible to both cosmological (BBN+CMB) and direct detection experiments. We forecast that the full DESI and the next-generation spectroscopic surveys will tighten these limits by more than an order of magnitude, establishing spectroscopic stacking as a competitive probe of electromagnetically interacting multi-state dark matter.
Read the full abstract
Cryptography and Security
GPUThor: Amplifying Rowhammer Attacks via Non-Uniform Patterns to Exploit ECC-Protected GPUs
GDDR memory in GPUs is vulnerable to Rowhammer attacks, where rapid memory accesses induce bit flips in adjacent cells, enabling data tampering and privilege escalation. However, prior GPU Rowhammer attacks trigger only tens to hundreds of bit flips, orders of magnitude fewer than CPU attacks, severely limiting their practical impact. This gap stems from the reliance of existing GPU Rowhammer attacks on uniform hammering patterns that activate aggressor and decoy rows equally, which results in low hammering intensity for aggressor rows. We present GPUThor, a high-intensity Rowhammer attack on NVIDIA GPUs leveraging non-uniform hammering. GPUThor reverse engineers GPU memory-access coalescing behavior to enable non-uniform hammering patterns on GPUs, that activate aggressor rows more intensely than decoy rows. Additionally, by identifying refresh instances when in-DRAM mitigations are applied, it constructs longer attack patterns that escape mitigation across refresh intervals, further increasing hammering intensity. Together, these techniques yield 500X to 23,500X more bit flips than prior GPU Rowhammer attacks, across several NVIDIA GPUs (A4000, A4500, A5000, A6000), reaching bit flip rates close to state-of-the-art CPU Rowhammer attacks. GPUThor also enables the first Rowhammer exploits on ECC-protected GPUs, inducing uncorrectable double and triple bit flips, making denial-of-service and privilege-escalation attacks practical even on GPUs with ECC enabled.
Read the full abstract
Materials Science
Exotic centrosymmetric phase of acentric urea under high pressure
Urea is a simple prototype supramolecular crystal that exhibits rich polymorphism at low pressure due to broken and restored N-H-O hydrogen bonds. The high pressure polymorph (phase V') of acentric urea crystallizes in a centrosymmetric structure, which presents an appealing target because of its potential exotic structure, analogous to the symmetric ice phase X. The pressure-induced polymorphism of urea was studied using powder X-ray diffraction, infrared and Raman spectroscopy, second harmonic generation (SHG) measurements up to 20 GPa and ab initio crystal structure prediction (CSP) based on the constrained evolutionary approach. A strong decrease of the SHG signal at the transition pressure 10 GPa reveals that the high-pressure polymorph is indeed centrosymmetric, further confirmed by the selection rules observed in the lattice vibration modes, in contrast to chemical intuition for acentric urea. The structural evolution sequence obtained from X-ray diffraction, SHG and CSP calculations is as follows: phase I (P421m; Z=2) from 0 to 0.5 GPa, Phase III (P212121; Z=4) from 0.5 to 5.2 GPa, and phase V' (P21/m; Z=6) beyond 10.0 GPa which is energetically competitive with the theoretically predicted phase V (Pnma; Z=4). A phase X with distinct spectral and diffraction features forms between 5.2 and 10.0 GPa, which could be explained by a quantum disorder intermediate state between phase III and V', that is ascribed to the difficulty to disrupt the H-bonding network under extremely compressed environment. The softening of N-H vibrations and the change in intensity of the vibrations associated with the hydrogen bonding provide evidence for proton tunneling and charge-transfer interaction in phase X.
Read the full abstract
Quantum Physics
Partial projected ensembles reveal slow tilt-constrained information spreading
Projected ensembles reveal information about quantum many-body dynamics beyond the reduced density matrix. Here we investigate how spatial constraints affect the emergence of this information in a kicked Ising chain with a spatially varying longitudinal field. We consider partial projected ensembles, obtained by measuring a remote region while leaving an intervening buffer unobserved. Our numerical results reveal a pronounced contrast between nearly ballistic correlation onset in the ergodic regime and strongly delayed onset in the tilt-constrained regime, corroborated by quantum mutual information. Despite these different onset scales, the conditional-state fluctuations decrease approximately exponentially with buffer length in both regimes. Full projected ensembles additionally exhibit slow relaxation and persistent measurement-basis dependence in their higher moments. We establish an information-theoretic characterization of the connected second moment and use a solvable dephasing model with effective-basis measurements to illustrate the separation of phase accumulation and coherence loss. Within this model, a factorially suppressed coupling envelope yields sublogarithmic spreading consistent with the finite-distance onset trends. Our results identify partial projected ensembles as probes of the distinction between the time required for correlations to develop and the information that remains accessible under partial observation.
Read the full abstract
Nuclear Theory
The chiral filter
Thermal neutron capture on a proton, $np\to d\gamma$, is measured to a fraction of a percent; the impulse approximation misses nine percent of it. This article follows that gap through five decades of an idea of Mannque Rho's, the chiral filter. Chiral symmetry fixes the pion-exchange current that fills the gap, and power counting explains why nothing else competes. The missing nine percent becomes a prediction, correct within errors. On the unprotected side, one constant fixed by tritium beta decay predicted solar $pp$ fusion and the hep process; $g_A^*$ remains open.
Read the full abstract
Quantum Physics
Phantom Codes: Hardness, Rate Optimal qLDPC Constructions, and Distance Limits
An $[[n,k,d]]$ stabilizer code is phantom if every in-block logical CNOT gate can be realized by a permutation of its physical qubits. This eliminates the large and complicated physical overhead normally required for logical entangling gates. Yet this symmetry is highly restrictive: phantom codes are rare, the number of logical qubits is limited to $k=O(\log n)$, and no phantom qLDPC family with growing logical dimension was previously known. We make three contributions. 1. We prove that recognizing phantomness of a given stabilizer code is at least as hard as Graph Isomorphism, even for CSS codes encoding only $k=2$ logical qubits. 2. We propose the first rate optimal phantom qLDPC families: for every fixed $D$, our CSS families achieve the maximal logical scaling $k=\Theta(\log n)$ and distance $d\geq D$. 3. We prove a distance no-go theorem: every phantom code family with $k=\omega(\sqrt{\log n})$ and check weight $w=O(1)$ satisfies $d\leq w$ for sufficiently large $n$. Thus our fixed-distance qLDPC families are distance-scaling optimal at the maximal logical scaling $k=\Theta(\log n)$.
Read the full abstract
Cryptography and Security
A Cyber Range Evaluation of Inductive Autonomous Network Incident Response Agents
We have applied a novel combination of classic relational database modeling and modern reinforcement learning to automated intrusion response to enable inductive decision agents. Unlike previous work in the area, our approach allows agents to handle variable quantities of network assets without retraining. We test the operational performance of our approach in a virtualized computer network with variable topologies, red-team automation and simulated user agents. The task of the agents is to prevent hosts in the network from being accessed by the red-team agent, while minimizing the availability costs induced from defensive measures. We test a combination of heuristic agents and policies trained using relational reinforcement learning. The reinforcement learning agents were overall more efficient at defending the system than the heuristic policy, and the closest to the performance of human operators at the task.
Read the full abstract
Machine Learning
On the Importance of Gating: Memorization vs. In-Context Learning in State Space Models
State Space Models (SSMs) have emerged as a compelling alternative to Transformers, enabling sequence modeling with constant memory and linear compute. Although SSMs exhibit reasonable performance and favorable computational characteristics, they continue to lag behind Transformers on tasks that require in-context learning and precise retrieval, slowing their adoption for large-scale language modeling. In this work, we demonstrate that both the success and failure of SSMs in these domains can be explained by studying the role of the gating mechanism, a prevalent component in modern recurrent networks. Specifically, we show through theory and experiments that this gating mechanism causes SSMs to first learn an in-weights "memorization" solution, while delaying, or even preventing, convergence to a correct in-context learning solution. Importantly, this happens even in cases where there are no fundamental limitations due to the architecture or its memory capacity. On the other hand, we find that gating is often beneficial for improving generalization to long sequence lengths. Our results illuminate the crucial role of the gating mechanism in shaping both the training dynamics and generalization of SSMs, and provide a basis for understanding and improving linear-time models.
Read the full abstract
Human-Computer Interaction
Ptolemy: A Semantic Map of Exploratory Data Analysis
A central challenge in exploratory data analysis (EDA) is keeping track of what has already been examined in order to decide what to analyze next. In practice, analysts often run dozens of analyses while building an understanding of a dataset. However, most tools provide little support for maintaining an overview of this evolving process, instead exposing only a linear history of analysis steps. These tools show sequence, what came before, but not position, how a current analysis relates to the broader space of possible analyses. As a result, analysts must mentally reconstruct which parts of the space they have explored and where gaps remain, increasing the risk of redundant work or overlooked patterns. We present Ptolemy, a navigational interface that externalizes analysis history as a semantic map. Each analytic step is represented as a point positioned by embeddings derived from a structured description of its effective data view (e.g., columns, filters, transformations), allowing spatial distance to reflect analytic similarity. In a mixed-methods study comparing map, canvas, and tree representations, we find that maps improve global orientation and local comparison, while ordered layouts reduce decision cost. These findings surface a trade-off between orientation and actionability, and highlight design principles for supporting strategic exploration in EDA.
Read the full abstract
Optics
Learning Transferable Self-Supervised Priors for Super-Resolution Reconstruction in Structured Illumination Microscopy
Structured illumination microscopy (SIM) extends the optical passband, and reconstruction of detail beyond it depends on prior knowledge. Hand-designed regularizers depend on how well their structural assumptions match the specimen; learned priors can be sensitive to changes in imaging conditions and specimen structure. We introduce SIMAdapter, which pretrains a network that predicts the emitter and the point-spread function (PSF) by self-supervision on 23,237 raw SIM stacks from BioSR, BioTISR, and simulations spanning different PSFs and specimen structures, then adapts it to a single unlabeled target stack. Adaptation refines the network against a differentiable image-formation model, with the light pattern calibrated from that stack. Both stages take their supervision from the raw measurements and need no paired high-resolution reference. On two held-out synthetic domains, SIMAdapter reaches a mean emitter normalized root-mean-square error of 0.156, compared with 0.403 for Sparse-SIM. The same adaptation started from a network pretrained on BioSR alone is less accurate in both domains. In three experimental case studies, adaptation reduces flanking artifacts and yields more distinct profiles across filament pairs, mitochondrial boundaries, and calibration lines. A single pretrained network can thus be reused across SIM measurements, with each reconstruction refined against its own raw data.
Read the full abstract
Machine Learning
What Does Layer-Importance Reveal About Transformers and State-Space Models?
Transformers and state-space models (SSMs) are the two dominant families of sequence models, and a central open question is how far the analytical knowledge built for transformers transfers to SSMs. We address this through the lens of layer importance which underpins compression, selective fine-tuning, and interpretability across both families. We decompose layer importance into two distinct notions. Necessity captures how much the pretrained model depends on a layer's existing contribution, measured by the loss increase from bypassing it. Plasticity captures where the model absorbs new information during fine-tuning, measured by the magnitude of task-specific weight updates. Our analysis reveals that the two families behave fundamentally differently: in every evaluated residual transformer up to $14$B parameters, Necessity and Plasticity anti-align across depth, whereas in the evaluated Mamba-style SSMs they point to overlapping regions. The sign of this alignment also predicts downstream adaptation behavior. In the evaluated transformers, concentrating updates in the most plastic layers increases catastrophic forgetting, while this tier-dependent effect disappears in the evaluated Mamba-style SSMs.
Read the full abstract
Numerical Analysis
Structure-Driven Inversion: A New Paradigm for Solving Inverse Problems
Inverse problems are predominantly solved within the optimization-driven paradigm, which formulates the problem as objective-function minimization and approaches the solution by iterative search. Though universal, it suffers from limitations in efficiency, interpretability, and multi-parameter decoupling. This paper proposes a new paradigm—Structure-Driven Inversion (SDI). Instead of iterative search, SDI identifies and exploits the intrinsic structure of the problem to construct a solution method. It has two types of structure—mathematical structure and physical structure—and two corresponding driving inversion types: Mathematical Structure-Driven Inversion (MSDI) and Physical Structure-Driven Inversion (PSDI). The current representative method of MSDI is Mathematical Structure-Driven Pseudo-Inverse Inversion (MSDPII), which constructs a pseudo-inverse in the spectral domain via unitary diagonalization. The current representative methods of PSDI are Physical Structure-Driven Waveform Inversion Imaging (PSDWII), which performs inversion through virtual-source projection, and Physical Structure-Driven Back-Propagation (PSDBP), which implements three structural projections for deep neural networks via physical-system analogy. SDI is not a rejection but a complement and extension of Optimization-Driven Inversion (ODI). To the best of the author's knowledge, no existing study has systematically presented structure-driven inversion as an independent paradigm; this paper aims to fill that gap.