Cryptography and Security
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Wednesday 2 September
Trusted Execution Environments (TEEs) provide hardware-supported isolation through enclaves that protect code and data independently of software abstractions. However, TEEs alone cannot enforce information-flow security. This problem is further aggravated in LLVM-like low-level languages that allow unrestricted pointer manipulation and unstructured control flow. Moreover, using TEEs effectively typically requires manually partitioning applications into enclave and non-enclave components, a process that is labor-intensive, error-prone, and lacks fine-grained control. We address these challenges with a three-step approach. First, we formalize SIR, an enclave-oblivious calculus based on LLVM IR, equipped with a novel permissive type system that enforces security against low-level attackers. To obtain meaningful guarantees, SIR combines information-flow control with security-aware coarse-grained memory safety. Second, we extend SIR to SIREN, an enclave-aware calculus that enforces noninterference against stronger attackers capable of observing arbitrary non-enclave memory. Third, we develop a type-driven, type-preserving compilation from SIR to SIREN that automatically produces secure enclave-aware programs, eliminating manual partitioning while providing fine-grained control over host-enclave boundaries. We implement and evaluate SPLITR on thirteen microbenchmarks and real-world workloads, including applications from SGXGauge, on Intel SGX hardware. SPLITR scales to OpenSSL (425,953 LLVM IR instructions) and supports multiple objectives that expose trade-offs among enclave TCB size, host-enclave transitions, and boundary data movement. For OpenSSL, optimizing for transitions reduces them from 393 to 187. Runtime overhead is dominated by fixed enclave costs for short-running workloads, whereas long-running applications better amortize these costs and approach native performance.
Adaptation and Self-Organizing Systems
Phase-delays shape multistability and basin sizes in Kuramoto networks: analytical estimates from network structure
We study how network connectivity and heterogeneous phase-delays shape the spatiotemporal dynamics of finite oscillator networks. Phase-delays can destabilize global synchronization and promote phase-locked patterns, including states with uniform phase gradients and more complex combinations of these modes. Yet, how connectivity and phase-delays jointly determine which states the network selects remains unclear. Here, we show that the spectrum of a composite matrix, which combines connectivity and phase-delays, governs not only the linear stability of the network's collective states but also their basin sizes. This, in turn, enables analytical estimates of basin size of phase-locked states for individual networks from connectivity and phase-delays alone. Applying this framework to nonlocal and global networks, including cases with random phase-delays, we uncover multistability and strong asymmetries in basin sizes, revealing chiral dynamics that conventional stability analysis cannot detect.
Read the full abstract
Robotics
Modeling What Changes: Sparse, Residual World Models for Object-Centric Manipulation
Monolithic world models predict the entire next state at every step, spending capacity re-predicting the static majority of a scene and injecting error into it. We ask whether explicitly modeling change (a per-object change gate plus a residual delta head that perturbs only the objects the gate flags) is a more effective and interpretable bias for physical prediction and control. On a MuJoCo tabletop pushing benchmark scaling from 3 to 8 objects, the sparse/residual model predicts next-state poses 2.5 to 4.6 times more accurately than a dense multilayer perceptron at 8.6 to 11.1 times fewer parameters, sustains change-detection F1 of 0.80 to 0.87 where the dense baseline is degenerate, transfers across object counts with zero retraining (99.4 percent F1 retention), and reaches about 90 percent of its full-data accuracy with a quarter of the data. In autoregressive rollout it compounds far less error, hugging the no-motion floor while the dense model drifts. Finally, inside a sampling-based planner, prediction-only models fail (though a true-simulator oracle solves the task with the identical planner, confirming the planner is sound), but once featurized and trained for the states a planner visits, the sparse model begins to plan (0.23 plus or minus 0.06 success over three seeds) while the dense monolith stays at zero at every seed. Modeling what changes, rather than re-predicting the whole world, is a simple, effective bias for object-centric physical AI; code, data generators, and all checkpoints will be released upon publication.
Read the full abstract
Quantum Physics
Analytic Maximal Violation of Extended MABK Inequalities for Generalized GHZ States
We analytically characterize the maximal quantum violation of the extended Mermin-Ardehali-Belinskii-Klyshko (EMABK) family of inequalities by $n$-qubit generalized Greenberger-Horne-Zeilinger (GHZ) states. We develop a correlation-tensor approach in which the Bell value is expressed as the Frobenius inner product of the quantum correlation tensor and an effective coefficient tensor. A rank constraint on the latter, together with the von Neumann trace inequality, yields an upper bound governed by the two largest singular values of the reshaped correlation-tensor. For generalized GHZ states, we determine the complete singular spectrum and obtain a piecewise analytic upper bound. We then construct two complementary measurement strategies that saturate the bound for EMABK throughout the entire parameter range. The first is a purely MABK strategy in which all measurement directions lie in the equatorial plane of the Bloch sphere. The second is a hybrid strategy that recursively combines lower-order MABK anti-diagonal operators with the fully-$\sigma_z$ tensor product. The resulting piecewise analytic expression recovers the standard MABK Tsirelson bound $2^{(n-1)/2}$ in the maximally entangled limit and approaches the classical local-hidden-variable bound in the product-state limits. It further shows that every entangled generalized GHZ state exhibits a strict quantum-classical separation under the EMABK inequality, thereby eliminating the nonviolation region of the standard MABK inequality in the partially entangled regime.
Read the full abstract
Optics
Rydberg-atom microwave angle-of-arrival detection via cylindrical vapor-cell-mediated field redistribution
Microwave angle-of-arrival(AoA) measurement is essential for radar, communication, and spectrum mon-itoring. Existing Rydberg-atom-based AoA schemes employ phase-difference measurements with local oscil-lators, standing-wave fluorescence imaging, or amplitude-ratio readout with internal metal reflectors. Here we demonstrate a new approach: a cylindrical glass vapor cell serving directly as an angle-encoding dielectric structure, eliminating the need for multiple apertures, local oscillators, or imaging optics. The cylindrical geometry produces angle-dependent reflection and field redistribution, mapping the incident AoA onto the effective microwave field sampled by the Rydberg ensemble. This effective field is read out optically via the Autler-Townes(A-T) splitting in the electromagnetically induced transparency(EIT) spectrum. Full-wave simulations and experiments at 11.64 GHz confirm a deterministic, geometry-mediated response over 0{\deg}-90{\deg}.Over the monotonic operating range(55{\deg}-125{\deg}), the angular resolution(minimum distinguishable increment)is 0.05{\deg}, and the angular accuracy(RMSE of repeated measurements) is 0.13{\deg} across the full range, improving to 0.05{\deg} in the optimal region(65{\deg}-115{\deg}). Occupying a sensing volume of~2.5 cm3, this method offers a compact, single-sensor pathway toward Rydberg AoA receivers.
Read the full abstract
Quantum Physics
A generalized harmonic oscillator problem for a spin-1/2 fermion
The exact bound-state wavefunctions and the corresponding energy equation are calculated for a new generalized harmonic oscillator problem describing a spin-1/2 fermion in 3+1-dimensions, involving scalar, vector, and tensor couplings acting simultaneously within a particular plane of motion. For the scalar and vector coupling, singular harmonic oscillator shapes are considered, such that the singular term is needed to allow analytical solutions for the wavefunctions while preserving binding under adequate conditions. For the tensor sector,the Dirac oscillator potential is employed, which adds another independent binding mechanism to the problem. The exact bound-state solutions are computed by specifically tuning coefficients for an appropriate pair of Ansätze for the radial functions, which leads to wavefunctions in terms of generalized Laguerre polynomials. Although the energy equation cannot provide a general expression for the energy spectrum, specific constraints on the quantum numbers as simple functions of the external potential parameters can be derived for it, which determines the conditions for bound solutions to exist, and of what type: particle, antiparticle or both. It is shown that the results can be simply mapped to the spherically symmetric analogue problem, and this is used to show that the general result encompass several previous particular cases of spherically symmetric harmonic oscillator problems in the Dirac equation available in the literature.
Read the full abstract
Machine Learning
Act More, Decide Less: Skill-Guided Adaptive Action Chunking for Long-Horizon LLM Agents
Large language model (LLM) agents for long-horizon interactive tasks typically follow a ReAct-style protocol, issuing one primitive action per LLM round. While this enables frequent replanning, it is inefficient for long-horizon tasks where many rounds are spent on routine action sequences. A natural alternative is to let the agent emit variable-length action chunks. However, naively training such policies with standard reinforcement learning fails: the agent either collapses to single-action behavior or over-commits to excessively long sequences. Both failures share a common root cause: the inability to learn chunk boundaries. We propose SPACE, which addresses this challenge by distilling chunk-boundary supervision from trajectory-induced programmatic skills. We induce two-level programmatic skills from successful trajectories, where subskill boundaries serve as direct chunk-boundary supervision. This temporal structure is then distilled into a primitive-chunk policy via hybrid on-/off-policy optimization with chunk-aware credit assignment. Experiments on ALFWorld and ScienceWorld show that SPACE improves success rates by 7.0%-31.3% over the strongest baseline in each setting while reducing average LLM decision rounds by up to 78.9%.
Read the full abstract
Combinatorics
Sparse $k$-AP Covering Sets and the Arithmetic Kakeya Conjecture
A subset $A\subseteq \mathbb N_0$ is $k$-AP covering if there exists a constant $n_0$ such that for every integer $x>n_0$, there exists $d\in\mathbb N_0$ such that $x-d, x-2d,\dots,x-(k-1)d$ are all in $A$. Disproving a conjecture of Kiss, Sándor, and Yang, we prove that for every integer $k\geq 6$, there exists a constant $\varepsilon=\varepsilon_k>0$ and a $k$-AP covering set $A$ such that $|A\cap \{0,1,\dots,n\}| < n^{\frac{k-2}{k-1}-\varepsilon}$ for all sufficiently large $n$. We also relate this problem to the Arithmetic Kakeya Conjecture by Katz and Tao.
Read the full abstract
Analysis of PDEs
Quantum Landau Damping with Coulomb Repulsion in the Whole Space
In this paper, we study the large time behavior of solutions to the linearized Hartree equation near a stable equilibrium which is homogeneous in space and investigate the quantum Landau damping. This allows us to get time decay of the electric field, uniformly in the Planck constant, as well as the uniform-in-time convergence of the quantum density towards the classical density as the Planck constant converges to 0.
Read the full abstract
Computer Vision
Test-Time Logit Prompting for Source-Free Missing Modality Adaptation
Vision-language models (VLMs) have achieved remarkable performance by leveraging complementary information from large-scale image-text pairs. However, missing-modality inputs are commonly encountered during real-world deployment, often leading to significant performance degradation. Existing methods primarily enhance model robustness by learning modality compensation strategies from source training data. However, their reliance on source training data makes them difficult to apply when original data are unavailable due to privacy, storage, or accessibility constraints, such as clinical applications and personalized AI services. This raises an important yet underexplored question: can VLMs be efficiently adapted at test time for visual recognition with missing modalities without accessing source training data? To this end, we propose Test-Time Logit Prompting (TLP), a lightweight source-free test-time adaptation framework for visual recognition with missing modalities. To address missing-induced prediction shifts, TLP optimizes logit prompts with uncertainty-aware adjustment and modality-complete consistency regularization, adaptively adjusting prediction confidence while preserving semantic consistency. Extensive experiments across diverse vision-language benchmarks demonstrate that TLP consistently enhances recognition performance under missing-modality scenarios, achieving up to 8% improvements while requiring only hundreds of tunable parameters and a few test-time optimization steps.
Read the full abstract
Optimization and Control
A Hereditary Property of Cutting Plane Procedures
Let $K'$ denote the closure of a convex set $K$ under a given cutting-plane procedure. The procedure satisfies the hereditary property if $F' = K' \cap F$ for every face $F$ of $K$. The property underlies inductive proofs of finite-rank and the polyhedrality of closures, and it is an admissibility requirement in abstract frameworks for cutting-plane procedures. Yet, it has not been studied systematically across well-known cutting-plane procedures. We establish two sufficient conditions for the property. The first applies to procedures that can be expressed as closures under intersection cuts from a family of lattice-free convex sets, when a single family realizes the closure of $K$ and of each of its faces. It yields the hereditary property for the split, lift-and-project, Lovász–Schrijver, Sherali–Adams, and Lasserre closures, applied over general convex sets and for faces that need not be exposed. The second applies to procedures whose cuts are valid inequalities for Gomory's corner polyhedron, and it yields the hereditary property for the closure of Dantzig cuts derived from all bases. We complement these results with four classical procedures for which the hereditary property fails, namely the closure of Gomory's fractional cuts (derived from either all bases or feasible bases only), the mixed-integer Chvátal closure, the $+$-cut closure, and the closure of Dantzig cuts derived from feasible bases.
Read the full abstract
High Energy Astrophysical Phenomena
Are Repeaters Prevalent Among the Known Fast Radio Burst Sources?
Fast radio bursts (FRBs) are millisecond radio pulses of unknown origin. Despite extensive follow-up observations, only $\sim3\%$ of FRBs have been confirmed as repeaters. It remains unclear whether the rest are truly one-off bursts, or essentially repeating sources that have only been detected once due to limited monitoring time. Using the second CHIME/FRB catalog, we test this debate by comparing non-repeaters with two repeater-based subsamples: the first-detected bursts of repeaters and their highest-fluence bursts. A non-parametric method that accounts for selection effects is employed to derive the energy functions and event rates of these samples. All samples are well described by broken power-law energy distributions with comparable break energies ($\sim 5\times10^{38}$ erg), but with significantly different slopes between repeating and non-repeating populations. Their event-rate evolution also differs significantly. Assuming $\rho(z) \propto (1+z)^B$, we have $B = -5.57^{+0.15}_{-0.15}$ for non-repeaters and $B = -8.63^{+0.46}_{-0.41}$ and $-9.10^{+0.55}_{-0.56}$ for the two repeater samples. Size-matched resampling shows that the repeater event-rate indices lie far outside the 5$\sigma$ range expected from non-repeater subsamples, ruling out sample size as the reason for the observed difference. These results indicate that at least a subset of one-off FRBs are intrinsically non-repeating, implying that repeating sources may represent a distinct and possibly less common population.
Read the full abstract
Computer Vision
SelfLift: Accelerating Few-Step Diffusion via Self-Recovering Resolution Transition
Few-step diffusion models substantially compress temporal computation, making the spatial cost of each model evaluation an increasingly dominant source of inference latency. Progressive-resolution inference reduces this cost by performing early denoising at low resolution and reserving high-resolution computation for refinement. However, existing methods typically lift intermediate latents directly and rely on subsequent steps to absorb the induced distribution mismatch. In the few-step regime, the limited recovery budget leaves these errors as visible artifacts, constraining how late the transition can occur and, consequently, how efficiently it can be performed. We introduce SelfLift, a self-recovering progressive-resolution framework that derives both transition-repair signals and trajectory-aligned supervision from the generative model itself. SelfLift-zero proposes a training-free Artifact-Aware Consistency Lift, using disagreement between direct latent lifting and pixel-VAE re-encoding as both a localized artifact-risk signal and a model-native correction direction. It enables reliable late transitions without external super-resolution, extra denoiser evaluations, or sampling-schedule modifications. Building on this robust transition, SelfLift-rich performs On-Policy Self Recovery on student-visited states, transferring dense high-resolution guidance from an internal self-teacher while remaining aligned with the altered progressive-resolution dynamics. Across FLUX.2-Klein and Z-Image-Turbo, SelfLift reduces end-to-end latency by 41.5% and 44.1%, respectively. Combined with timestep distillation, it delivers overall speedups of 29.61x and 19.21x over the corresponding 50-step models while preserving competitive generation quality, establishing a stronger speed-quality frontier for few-step diffusion.
Read the full abstract
Cryptography and Security
Implicit Manipulation for Skill Selection in LLM Agents with Semantic Matching
Skill selection is a key stage in LLM-agent workflows, determining which installed skill should handle a user request. Existing attacks on this stage primarily rely on explicit prompt injection or instruction-level steering, which can expose recognizable manipulation signals. In this work, we identify a new implicit attack surface for skill selection: even when the user prompt and skill description appear benign in isolation, their semantic relationship can still be strategically shaped to favor an attacker-chosen skill. Based on this observation, we present Implicit Skill-Selection Manipulation via Semantic Matching (ISM), which jointly shapes target-skill metadata and reusable prompts to manipulate skill selection without explicit selection instructions. Specifically, we develop a three-stage strategy to broaden semantic coverage, strengthen target distinctiveness, and preserve natural prompt wording. Across four task domains and eight selector models, ISM increases the average target-selection rate (TSR) from 15.2% to 63.5%. In a matched comparison, ISM achieves a 73.5% TSR, only 9.8 percentage points below Explicit Steering. Human reviewers block ISM in only 2.9% of judgments, versus 91.4% for Explicit Steering, while five LLM-based inspectors pass ISM at an average rate of 82.9%, versus 37.4% for Explicit Steering. Moreover, ISM remains effective against PPL-W, Llama Prompt Guard 2, and PIGuard.
Read the full abstract
Atmospheric and Oceanic Physics
Kilometer-Scale AI Downscaling of Atlantic Hurricanes with Generative Ensembles
This study presents an AI-based dynamical downscaling system for Tropical Cyclones (TCs). The system incorporates an AI-based limited-area model that downscales 3-hourly low-resolution boundary forcings into hourly high-resolution fields autoregressively, and a diffusion model that converts the outputs into ensembles of hazard-relevant variables. The system is trained on the regridded CONUS404 data with ERA5 forcings, and is evaluated on 20 TCs in 2020–2024. Verification shows stable downscaling performance across Atlantic hurricane seasons, with energy spectra closely matching the CONUS404 reference. The system is also verified to produce skillful TC-relevant weather extremes, largely improved over a deterministic AI baseline. The system performs well with forcing data from other models (GDAS/FNL) and can produce detailed eyewall, rainband, and landfall structures in TC case studies. The study provides a good example of how AI-based dynamical downscaling systems can be designed to resolve small-scale extreme weather events.
Read the full abstract
High Energy Physics (Experiment)
Search for the baryonic decay $ D_{s}^{*+} \to \ p \bar{n} $
The first search for the baryonic decay $ D_{s}^{*+} \to \ p \bar{n} $ is performed using $e^+e^-$ collision data taken at center-of-mass energies between 4.128 and 4.226 GeV, collected by the BESIII experiment and corresponding to an integrated luminosity of 7.33 fb$^{-1}$. No significant signal is observed, and an upper limit on the branching fraction is set to be $1.3\times 10^{-4}$ at the $90\%$ confidence level.
Read the full abstract
Statistical Methodology
COSTA: Covariance-Optimized Design and Causal Inference under Network-Temporal Interference
Experiments on networks observed over time face network spillovers, temporal carryover, and dependence deliberately introduced by the design. We propose COSTA—Covariance-Optimized Spatiotemporal Treatment Allocation—a joint Bernoulli design for unit–time assignments. Under common treatment marginals and a nonnegative linear network–temporal exposure model, Horvitz–Thompson bias for the sustained all-treated versus all-control contrast is exactly the negative expected weight of an assignment cut. A covariance-level variance envelope yields an MSE bound that can be optimized directly over assignment covariance. To scale this design, we introduce a thresholded-Gaussian Kronecker parameterization that mirrors the network and temporal exposure operators while preserving valid Bernoulli marginals. We next develop inference theory for the joint effects of designed treatment dependence and interference-induced outcome dependence. Canonical correlations between latent blocks generating separated HT contributions supply the coefficients required by graph-$\psi$ central limit and network-HAC theory; a spectral-floor and far-row-mass condition gives a primitive sufficient check. The framework covers sparse, block, Kronecker, locally factored, and other structured covariance sequences satisfying these conditions. Semi-synthetic RetailRocket and MovieLens experiments show substantial default-setting RMSE reductions and well-calibrated model-assisted design-centered intervals across linear, nonlinear, and demand-substitution outcome surfaces.
Read the full abstract
Theoretical Economics
Equilibrium Architecture in Multi-Battle Contests with Count-Dependent Prizes
We study how aggregating local victories shapes participation and equilibrium uniqueness. Two contestants with possibly different marginal costs compete across identical Tullock battlefields with discriminatory power at most one. A nonconstant, nondecreasing, symmetric schedule divides a fixed prize by victory count. We establish existence and show that every realized equilibrium allocation divides effort equally across battlefields. Equilibrium is unique through six battlefields, with at most two effort levels per contestant; seven first permits multiple equilibria or randomization by both contestants. The prize schedule and technology, not the cost ratio, determine whether equilibria use infinitely many effort levels; in that case, cost asymmetry changes only the higher-cost contestant's participation probability. Under majority rule, we characterize the unique equilibrium for every battlefield count and cost ratio.
Read the full abstract
Game Theory
The Complexity of Justified Representation with Additive Utilities
We study the computational complexity of satisfying proportional representation – in particular proportional, extended, and fully justified representation (PJR, EJR, and FJR) – in participatory budgeting and committee elections with additive utilities. First, we give a complete picture of the complexity of the axioms for a constant number of voters or voter types. Second, we show that even for committee elections with integer utilities bounded above by a small constant, satisfying FJR is intractable, giving the first strong NP-hardness result for a justified representation axiom. Third, we extend the Expanding Approvals Rule to committee elections with additive utilities and show that it satisfies PJR. Lastly, we show that no sequential voting rule can improve on the known positive result, thus proving that novel, substantially different voting rules are needed to surpass these boundaries. Beyond their theoretical merit, our results carry practical importance, as multi-winner voting with additive utilities has recently been gaining prominence in online deliberation and real-world participatory budgeting.
Read the full abstract
Artificial Intelligence
HeadWiseKV: Budgeted Per-Head Cache Residency for Hybrid Long-Context Language Models
Long-context inference retains a growing key–value (KV) cache during decoding, which consumes substantial GPU memory and can reduce generation throughput. This bottleneck remains in hybrid language models because their residual global-attention layers can dominate context-dependent cache demand. We study how to allocate this state under an aggregate KV-residency budget. We introduce HeadWiseKV, a training-free framework that compresses the residual global KV caches of hybrid language models while preserving their native local, recurrent, and linear paths. It assigns each physical KV head a static, multilevel history window, making cache demand predictable before serving. We formulate this allocation as a restricted operational rate–distortion problem and propose SeqCalib as the core policy-generation algorithm in HeadWiseKV. SeqCalib processes layers in execution order and conditions each decision on the lower-layer policy used at deployment, thereby accounting for interactions across depth. A grouped-cache runtime materializes the selected policy as actual per-head KV residency rather than a mask over a full cache. We evaluate downstream quality across four hybrid long-context models and study physical residency and serving behavior on Qwen3.6-27B. HeadWiseKV retains near-Full-KV RULER and LoCoMo quality across the evaluated models. In the fixed-model systems study, it reduces sampled peak device memory by 8.59% at a 112K context length and extends the largest verified successful context from 114K to 161K.
Read the full abstract
Computer Vision
Detecting Object Hallucinations in Large Vision-Language Models via Cross-Modal Attention Drifts and Mask-Based Verification
Despite recent advances in large vision-language models (LVLMs), object hallucination remains a major barrier to their reliable deployment. Existing detection methods often characterize visual grounding using attention from individual layers, leaving its evolution across layers underexplored. We propose CADMP, a lightweight object hallucination detection framework that combines adjacent-layer cross-modal attention drift with prediction sensitivity to targeted visual masking. During decoding, CADMP quantifies distributional changes between consecutive cross-modal attention maps to capture abrupt transitions in visual grounding. It then selects the transition with the largest drift, locates the corresponding visually relevant regions, and measures the change in prediction probability after masking these regions. These two signals provide complementary evidence: attention drift characterizes the stability of internal visual grounding, while probability variation verifies whether a prediction truly depends on the identified visual evidence. A lightweight detector integrates both signals to identify hallucinated predictions. Experiments on multiple benchmarks and representative open-source LVLMs demonstrate that CADMP achieves consistently competitive detection performance. Ablation studies further confirm the complementary contributions of adjacent-layer drift modeling and mask-based grounding verification.
Read the full abstract
Performance
Multi-Turn LLM Conversations under the Least-Recently-Used Policy: Mean-Field Asymptotics and Hit Ratio Approximation
The major workloads in modern large language model (LLM) serving systems have shifted from single-shot LLM calls to multi-turn conversations, where new responses are generated based on the whole conversation history across all previous turns. The hit ratio, i.e., the average fraction of KV caches accessed directly from existing caches stored in high-bandwidth memory (HBM), is hence a crucial metric that governs system performance. Estimating the hit ratio is a highly nontrivial task due to the complex system dynamics, where the KV cache prefixes grow with turns and some must be evicted due to finite memory capacity. We formulate the system as a multi-turn conversation model under the least-recently-used (LRU) policy. Through a mean-field asymptotic framework, we prove that as the conversation arrival rate and the memory capacity grow proportionally to infinity, the hit ratio converges to a closed-form limit. Based on the characterization of the limit, we further propose a practical hit ratio estimator, and validate its accuracy by real LLM serving experiments on the Qwen3-8B model implemented on Ascend NPUs. Our results provide a theoretical foundation for the analysis of multi-turn LLM serving systems and a practical guideline for memory capacity provisioning.
Read the full abstract
Signal Processing
Pinching-Antenna Systems-enabled Secure ISAC: A Two-Timescale Optimization Framework
A novel two-timescale optimization framework is proposed for pinching-antenna systems (PASS)-enabled secure integrated sensing and communications (ISAC). Specifically, a base station (BS) equipped with pinching antennas (PAs) transmits signals to a legitimate user under the existence of an eavesdropper (Eve), while employing leaky coaxial cables (LCXs) for receiving echo signals to track Eve's mobility states, i.e., locations and velocities. Considering the practical PAs activation overhead, the pinching beamforming and baseband processing are optimized in the large and small timescales, respectively. The multiple-waveguide scenario is first considered, where the BS can transmit the artificial noise together with communication signals for both jamming and sensing purposes. A joint baseband and pinching beamforming design problem is formulated to maximize the average secrecy rate. To address this problem, an alternating optimization algorithm is first invoked for jointly optimizing the pinching and baseband beamforming with predicted Eve's mobility states. With determined PAs positions, the baseband beamforming is updated with refined Eve's states obtained from real-time echo signal processing. The single-waveguide scenario is then considered. Since a single waveguide carries at most one independent data stream, an ISAC framework with separate communication and sensing phases is proposed. The element-wise algorithm proposed for the multiple-waveguide scenario is extended to solve the resultant pinching beamforming problem. Numerical results demonstrate that: 1) Eve's velocities and positions can be accurately tracked with the proposed two-timescale framework in both multiple- and single-waveguide scenarios; and 2) PASS achieves superior secrecy rate compared to conventional multiple-antenna benchmarks.
Read the full abstract
Mesoscale and Nanoscale Physics
Complete Topological Classification with P and T Symmetries: Revealing a Topological Invariant Invisible to K-Theory
The K-theoretic framework provides a complete topological classification of the tenfold symmetry classes and has been generalized to incorporate crystalline symmetries. Here, we show that this classification is incomplete even in the elementary case of spinless systems possessing both P and T symmetries. We obtain the complete classification through a first-principles analysis of the topological classes of P- and T-symmetric bands, namely, by classifying the corresponding symmetric clutching data. We identify a topological invariant that is invisible from the K-theoretic perspective when the occupied states at each inversion-invariant momentum have uniformly positive or uniformly negative parity. In special cases, such as when all inversion-invariant momenta have uniformly positive parity, this invariant can be interpreted as the second Stiefel–Whitney class or the Euler number defined over an inversion fundamental domain, namely, half of the Brillouin zone. Our work not only reveals a new topological invariant that cannot be determined from the parity spectra at inversion-invariant momenta under P and T symmetries, but also demonstrates the existence of crystalline topological phases that are absent from the K-theoretic classification.
Read the full abstract
Statistics Theory
Spectral Analysis of Gram Matrices with Missing at Random Observations: Convergence, Central Limit Theorems, and Applications in Statistical Inference
Motivated by the statistical inference using the Gram matrix in the context of missing at random observations, this paper investigates the spectral properties of the random matrices $\mb S_n=\frac{1}{n}\mb Z\mb Z^*$, where $\mb Z=\mb D\circ(\boldsymbol{\Sigma^{1/2}}\mb X)$ represents a Hadamard random matrix with entries determined by independent Bernoulli variables $\mb D$. Operating within the high-dimensional framework, we establish the convergence of the empirical spectral distribution of $\mb S_n$ to a well-defined limiting distribution. In addition, we explore the impact of the missing mechanism on the second-order properties of the spectral distribution of the Gram matrix $\mathbf{S}_n$. We establish the central limit theorem for the linear spectral statistics of $\mathbf{S}_n$, shedding light on their fluctuations. Surprisingly, our analysis reveals that even in the ideal Gaussian distribution scenario, the fluctuations of statistics generated by eigenvalues are influenced by the eigenvectors of the population covariance matrix in the missing-at-random case. This discovery uncovers a remarkable phenomenon that starkly contrasts with the classical case. Subsequently, we demonstrate the practical application of our central limit theorem in hypothesis testing for the population covariance matrix.
Read the full abstract
Representation Theory
A negative answer to a question on tilting objects and two-term complexes
Let $M$ be a silting object in an idempotent complete algebraic triangulated category $\mathcal T$. Put $B={\rm End}_{\mathcal T}(M)$, and let $\mathbb{P}_M\colon {\rm pr}(M)\to K^{[-1,0]}({\rm proj}B)$ be the presentation functor associated with $M$. It was recently asked whether $\mathbb{P}_M(T)$ must be tilting whenever $T\in{\rm pr}(M)$ is a tilting object. We answer this question in the negative by giving an explicit finite-dimensional example. Namely, for $$ \Lambda=k(1\xrightarrow{\alpha}2\xrightarrow{\beta}3\xrightarrow{\gamma}4)/(\alpha\beta\gamma),$$ we construct a silting object $M\in K^b({\rm proj}\Lambda)$ and a tilting object $T=\Sigma\Lambda\in{\rm pr}(M)$ for which $$ {\rm dim}_k{\rm Hom}_{K^b({\rm proj}B)}\bigl(\mathbb{P}_M(T),\Sigma^{-1}\mathbb{P}_M(T)\bigr)=1.$$ Thus $\mathbb{P}_M(T)$ is a two-term silting complex but not a tilting complex.
Read the full abstract
Probability
Well-Posedness for SDEs with Logarithmical Critical Distributional Drifts
We study the stochastic differential equation $$d X_t=b(t,X_t)d t+\sqrt{2}d W_t$$ on $\mathbb R^d$, where $b$ is a time-dependent, divergence-free distributional drift of critical Hölder–Besov regularity $-1$, strengthened by an iterated-logarithmic correction. For every initial probability law, we construct a weak solution by smooth approximation and realize the singular drift as an additive functional. The main analytic ingredient is the Schauder estimate with a logarithmic smallness factor. Combined with uniform logarithmic Krylov estimates and a stochastic substitution formula for distributional test functions, this estimate allows us to apply a Zvonkin transformation and prove uniqueness in law among weak solutions satisfying the corresponding Krylov bounds. For solutions starting from deterministic points, we further show that their time-marginal distributions admit densities satisfying two-sided Aronson-type Gaussian estimates.
Read the full abstract
Representation Theory
Measures on partial orders
We determine the measures (in the sense of Harman–Snowden) on the Fraïssé class of partially ordered sets: the space of measures is a union of a plane, eight lines, and 15 isolated points. This is the first case where the space is not equidimensional, and the first primitive case in which it has dimension at least two. ChatGPT was used to obtain many arguments. The writing was done entirely by the author.
Read the full abstract
Robotics
Real-Time Dynamics-Based Torque-Sampling MPPI for Compliant and Force Aware Manipulation
This study proposes a novel Model Predictive Path Integral (MPPI)-based task-space control framework. The proposed framework explicitly solves rigid-body dynamics within a real-time MPC formulation and enforces safety constraints, enabling accurate motion and force control that yields compliant behaviors for safe and effective physical interaction of robotic manipulators in unstructured environments. By leveraging MPPI, the proposed framework efficiently handles nonlinear dynamics that are difficult to solve with conventional MPC approaches in real-time. Furthermore, we develop a torque-sampling-based control architecture that enables efficient exploitation of GPU-based parallelization, resulting in effective compliant and force-aware behaviors. As a result, the proposed framework achieves a solver update rate of over 166 Hz with a 0.18 s prediction horizon, and its performance is validated through real-world experiments on a 7-DoF manipulator.
Read the full abstract
Analysis of PDEs
Pseudo-differential noise and nonlocal singularity formation in the stochastic Córdoba–Córdoba–Fontelos equation
We study the stochastic Córdoba–Córdoba–Fontelos equation driven by multiplicative Stratonovich noise. The noise amplitude is allowed to be a pseudo-differential operator whose leading part is nearly skew-adjoint. This class contains classical transport noise and also permits genuinely nonlocal perturbations. We first develop a local-in-time theory for maximal classical solutions in Sobolev spaces, proving existence, uniqueness, and a blow-up criterion. We then consider the special case of Stratonovich transport. For sufficiently large initial nonlocal steepness at a global maximum, we prove finite-time blow-up with arbitrarily high prescribed probability and obtain an explicit upper bound on the lifespan. Finally, on the event that the nonlocal steepness at the transported maximum diverges, we establish a conditional Type-I upper bound. When the terminal Cesàro average of the normalised nonlocal energy converges, we further identify the exact leading-order blow-up rate in terms of its limiting value.