Papers

Choosing when to translate multilingual documents is a central routing problem in text classification: translation can improve predictions for some languages while degrading others or adding unnecessary computation. Uniform translation and heuristic language tiers do not provide statistically controlled route selection. We introduce a language-level router based on paired e-processes that continuously compares direct and translation-assisted classification before freezing a routing policy. A familywise-controlled threshold of 280 bounds the probability of any false route across 14 eligible languages per dataset by 0.05. On SIB-200 and MASSIVE, the router selects translation for 4 of 15 languages and 14 of 15 locales, improving held-out accuracy over direct classification by 8.14 and 16.70 percentage points, respectively. All 28 decisions remain stable across 50 outcome-independent orderings and relative to the per-group threshold. Our results demonstrate that paired e-processes enable statistically controlled, anytime-valid, and auditable multilingual classification routing.

Alefeld (1981) recast Halley's cubic convergence as Newton's method on $g=f/\sqrt{f'}$, linked to the Schwarzian derivative. We generalize this to matrix gradient fields $F=\nabla\phi$ ($X\in\R^{m\times n}$) via a matrix Schwarzian derivative interpreted through information geometry ($\alpha$-connections). Simplifying Palmore (1994), we construct this operator square-root-free derivative from the third Fréchet derivative of the Newton map. Our main theorem proves local cubic convergence with an explicit error constant, without self-adjointness, commutativity, or gradient-field assumptions. A matrix-free algorithm (Hessian-vector products only) validates the theory. We contrast this with a power-Newton family (whose naive matrix extension fails) and the matrix Laguerre family (which requires an operator square root). A case study on the Oja-type system $\dot X=AXB-XBX^TAX$ reveals that convergence depends on target eigenvalue gaps, cubic gains grow with ill-conditioning, and 70-digit tests confirm exact theoretical orders alongside a working-precision accuracy budget. Finally, we note coupled formulations excel primarily when sequential deflation is inapplicable.

Developing unified physics-based humanoid controllers that can navigate complex 3D scenes and manipulate objects remains a longstanding challenge. Existing approaches are often specialized for either locomotion or object-centric manipulation, or rely on task-specific reward engineering that does not scale well across diverse behaviors. We present CHIP, a unified, physics-grounded framework for learning reusable humanoid interaction skills from heterogeneous motion data. Central to our approach is a conditional interaction prior that models a context-dependent distribution over these skills within a shared discrete space. Our method is trained in three stages. We first learn physics-based motion-imitation policies that acquire grounded teacher behaviors from heterogeneous interaction data. We then distill these behaviors into a context-conditioned interaction prior that captures reusable motion structure across locomotion and manipulation. Finally, we initialize downstream task policies from the pretrained prior and adapt them through prior-regularized online RL post-training. Experiments on a diverse suite of humanoid interaction tasks show that our approach supports scene-aware locomotion, contact-rich object manipulation, and compositional behaviors such as environment-aware object transport and long-horizon skill sequencing, while producing smooth transitions and physically plausible motion.

Mixture of Prompt Experts (MoPE) adapts multimodal transformers through input-dependent prompt composition, while retaining a fixed prompt length. We investigate a layer-wise gating extension in a binary chest X-ray classification pilot study. The controller predicts a retention ratio for each sample, averages these ratios within a mini-batch, and uses the resulting integer length to truncate the static and mixed visual prompts. Retained mixed prompts are also scaled by the individual ratios. In one recorded run per configuration, the gated model reached a best validation accuracy of 0.8996, compared with 0.8969 for the fixed-length baseline; the corresponding final values were 0.8963 and 0.8802. The exported gate statistics imply a retained length of one at all recorded training points, relative to a configured maximum of six. This reduces the complete visual sequence from 210 to 200 tokens, but no direct runtime measurements establish an acceleration benefit. Report-derived labels, report text as input, sequential data partitioning, and the absence of repeated controlled experiments limit interpretation. The findings document prompt shortening under the configured gate penalty; they do not establish sample-specific length allocation, superiority over fixed short prompts, or clinical utility. Code is available at: https://github.com/jingtaolei/mope-dynamic-prompt-truncation.

We present a rank-23 algorithm for general 3 x 3 matrix multiplication using 52 additions or subtractions, together with a Karstadt-Schwartz-style basis change [1, 2]. This gives, to the best of our knowledge, a new best-known candidate for the addition-minimization problem for rank-23 general 3 x 3 matrix multiplication.

A shared reward gives agents a common objective, but leaves open when, how and even whether they must cooperate to succeed. We address these questions in the Laser Learning Environment, a multi-agent path-finding environment where cooperation materializes as one agent blocking a laser to let a teammate pass safely. We represent these interactions through temporal cooperation graphs whose timed edges connect helpers to beneficiaries, define six cooperation profiles as overlapping graph predicates, and prove that every cooperative trajectory satisfies at least one. By encoding the environment dynamics and profile predicates as propositional formulae, we distinguish tasks that admit}a profile in some winning trajectory from those that require it in every winning trajectory within a specified horizon. Used as filters, these queries turn a random layout sampler into a generator of tasks with certified cooperation requirements. Experiments with five multi-agent reinforcement learning algorithms show that training diversity improves joint success on unseen tasks when cooperation-free solutions exist. When cooperation is required, greater diversity improves individual-agent exits, but joint success remains near zero. Across five profile-certified pools, final exit rates averaged over algorithms separate the pools into four statistically distinguishable levels but this ordering primarily reflects partial completion: policies collect rewards for individual exits but rarely exhibit the profile required for joint success. Our framework exposes this gap between rewarded partial completion and realized cooperation by certifying what cooperation successful completion requires and using temporal cooperation graphs to reveal what policies exhibit.

Energy-harvesting Internet of Things (IoT) devices must decide whether to transmit each observation immediately or accumulate several observations before reporting to the edge. Early reports make information from initial observations available sooner at the edge but require more transmission actions, whereas larger batches save reports while delaying that information and retaining more state. We develop a unified analytical framework to compare this fixed-batching choice in scenarios with repeated, deadline-constrained inference cycles. Every configuration processes the same ordered observations and yields the same final posterior if all reports arrive, enabling a controlled comparison of the persistent state, actions, and energy costs induced by batching. Using reductions in Bayesian risk under the continuous ranked probability score (CRPS), we express long-run timely value exactly as a statistical-value-weighted sum of expected decision-weighted report availabilities. This separates the inferential value of observations from their availability at application-relevant decision times. We construct exact Markov-reward models accounting jointly for intermittent harvesting, finite storage, energy carry-over, unreliable delivery, retransmissions, and deadlines. We prove that the long-run regime is well defined and derive exact batch-size comparisons, timing benchmarks, and a decomposition of harvesting-law effects into within-cycle and carry-over contributions. Directly evaluable Gaussian and Gaussian-mixture belief specializations show that the preferred batch size can change with the prior, the decision-time profile, link reliability, harvested-energy statistics, and accumulation cost.

Small-object detection remains challenging because limited pixels cause information loss and suppress the scale knowledge encoded in pretrained detectors. Existing approaches mainly improve representations through multiscale training, architecture redesign, or parameter adaptation, implicitly assuming that frozen models lack the required capability. We challenge this assumption and hypothesize that small-object knowledge already exists in frozen detectors but remains underactivated and unstable during query evolution. To test this hypothesis, we propose Counterfactual Query-Trajectory Reliability (CQTR), a training-free framework that elicits latent responses through counterfactual scale interventions and interprets candidate reliability from decoder-internal spatial convergence, semantic persistence, and cross-scale conflicts. A small unlabeled training subset selects the appropriate correction mechanism for each model-data stream, without parameter updates or target-domain annotations. Across 27 combinations of nine frozen detectors and three datasets, CQTR consistently improves average precision (AP) and average precision for small objects (APs). Closed-loop analyses further show that scale intervention activates latent responses, trajectory evidence predicts ground-truth support, and unlabeled routing selects the more effective branch. CQTR therefore reframes small-object detection from external scale augmentation to the activation and reliability assessment of latent scale knowledge.

We investigate stochastic simple bilevel optimization with smooth and possibly nonconvex upper- and lower-level objectives. Existing stochastic extensions of dynamic barrier gradient descent (DBGD) either obtain fast convergence under an unverifiable trajectory-dependent “rare-visit” assumption, or remove this assumption at a substantially higher oracle cost. We show that a simple denominator-only regularization of the DBGD multiplier eliminates the need for such an assumption while preserving fast convergence rates. Specifically, our method achieves $(\varepsilon, \varepsilon)$-stationarity in $O(\varepsilon^{-2})$ iterations using $O(\varepsilon^{-4})$ upper-level and $O(\varepsilon^{-7})$ lower-level stochastic gradients, which improves upon the best assumption-free complexities. We additionally derive anytime parameter schedules.

Machine learning methods, and especially neural networks, are now routinely used for malware detection in network traffic. Though very effective, systems based on such methods often (i) are purely data-driven, ignoring the substantial body of available knowledge about the tactics, techniques, and procedures (TTPs) possibly used, and, consequently (ii) are not precise, since they either cannot correlate malicious activity with TTP usage, or if they do, they are unable to explain which TTP has been maliciously used. In this paper we demonstrate that it is possible to precisely detect malware by (i) providing the neural network model with information about the TTPs used by any given sample, and (ii) teaching the neural network to detect not just the malicious activity as a whole, but which specific TTPs are maliciously used. We show that our approach consistently outperforms the three alternative models, which either do not exploit TTP information, or which are not taught to detect the malicious usage of TTPs, or both. Moreover, we show that our approach (i) is particularly beneficial in detecting malware that utilises rarely-used TTPs, a scenario which is particularly challenging for the other systems; (ii) allows for TTP by TTP tuning, further improving its ability to detect the malicious usage of TTPs; (iii) consistently outperforms other systems across a wide-range of scenarios, including when relying on limited training data or when subjected to adversarial attack.

World Action Models (WAMs) extend Vision-Language-Action (VLA) models by incorporating future visual dynamics into action generation. However, existing WAMs often utilize imagined futures with limited adaptation to evolving execution progress, potentially introducing distracting or unreliable predictive cues. This limitation arises from two empirically identified forms of non-uniformity in future utility: (i) at the inter-progress level, the utility of imagined futures varies across execution stages as control demands change; and (ii) at the intra-progress level, individual future latents exhibit heterogeneous relevance within the same progress state. To address these limitations, we propose ProWAM, a Progress-Conditioned World Action Model that introduces execution progress as an explicit intermediate representation for adaptive imagination utilization. ProWAM comprises two tightly coupled components: (1) To obtain a reliable representation of execution progress, we propose the Self-Supervised Dual-Temporal Progress Encoder (SS-DTPE). SS-DTPE couples short-term action-observation interaction modeling with long-term recurrent progress aggregation to capture recent execution feedback and accumulated task history. (2) Conditioned on the progress representation from SS-DTPE, we propose the Hierarchical Progress-Conditioned Imagination Modulation (HPIM) to adapt imagination utilization to execution progress. HPIM operates at two complementary levels: an inter-progress global modulation mechanism adapts future utilization across execution stages, while an intra-progress relevance mechanism differentiates individual future latents within each progress state. Extensive experiments demonstrate consistent gains over strong VLA and WAM baselines.

Recent studies report that automated red-teaming finds more vulnerabilities, at lower cost, than human red-teaming on standard AI safety benchmarks, and some read this as evidence that human evaluators are becoming dispensable. The comparison measures one thing and the conclusion claims another. A benchmark measures how thoroughly an attacker searches a predefined set of harms, fixed in advance by the developers, and a harm left out of that set is invisible to any attacker working inside it, automated or not. The same blind spot appeared in academic cryptography and in clinical drug trials, where an evaluation that was internally valid stayed silent about the population it was never pointed at. We call the AI-safety version the threat-model coverage gap, and find that it persists in a current open-weight model, where harms surface in non-English prompts that English benchmarks miss. Closing it requires evaluators whose deployment context differs from the developers'. The case for those evaluators is methodological, grounded in coverage, and the existing evaluation frame is unlikely to produce them on its own.

Bivariate bicycle (BB) codes of Bravyi et al.~\cite{Bravyi2024} are quantum low-density parity-check codes with weight-$6$ checks, exemplified by $[[144,12,12]]$ with $kd^2/n=12$. We develop the algebraic structure theory of BB-type codes with weight-$8$ checks (weight-$4$ generator polynomials) and use it, together with an exactly validated search pipeline, to construct and certify new codes. We prove an exact dimension formula $k=2\dim R/(A,B)$ (forcing even $k$), a $4\ell m$-element symmetry group on generator pairs, an $X/Z$ distance equality $d_X=d_Z$, and a family of subgroup-coset kernel vectors giving rigorous distance upper bounds and a design rule for high-distance constructions; all distances are computed exhaustively by a cross-validated bit-mask verifier. At $n=144$ the pipeline returns a census of $53$ codes whose strongest members surpass the BB benchmark: $[[144,6,d\ge 15]]$ exceeds the benchmark distance $12$ (certified $d\ge 15$), $[[144,10,12]]$ reaches it with weight-$8$ checks, and $[[144,16,10]]$ encodes a third more logical qubits at $kd^2/n=11.11$ ($7.4\%$ below benchmark) while decoding no worse. At $n=72$, $[[72,14,8]]$ attains $kd^2/n=12.44$—more than twice the same-length BB code—and decodes better; a circuit-level memory experiment places our weight-$8$ codes at $\approx 0.1\%$ pseudo-threshold versus $\approx 0.4\%$ for the BB reference under an identical model, quantifying the threshold cost of the heavier checks. All structural statements are verified numerically on the whole census.

We show that for every prime power $q$ and $\rho\in(0,1-\frac1q)$ there exists $\varepsilon_0(q, \rho) > 0$ such that for all positive $\varepsilon < \varepsilon_0(q, \rho)$, a random linear code over $\mathbb{F}_q$ of rate $1-h_q(\rho)-\varepsilon$ has list size $\lfloor h_q(\rho)/\varepsilon\rfloor$ or $\lfloor h_q(\rho)/\varepsilon\rfloor+1$ with probability $1-o(1)$. Before this work, the list size was known only up to a constant factor for $q\ge3$, and for $q=2$ it was known to be either $\lfloor h_2(\rho)/\varepsilon\rfloor-1$, $\lfloor h_2(\rho)/\varepsilon\rfloor$ or $\lfloor h_2(\rho)/\varepsilon\rfloor+1$. We also show that the same two-point concentration holds for the list size of random LDPC codes. This follows from a general theorem establishing that random LDPC codes and random linear codes have the same threshold rate for distance, list decoding, list recovery from errors and any other convex property, which answers a question of Mosheiff, Resch, Ron-Zewi, Silas and Wootters.

Social deduction games (SDGs) require agents to reason under partial observability by maintaining relational beliefs about hidden roles and team alignments. While recent LLM-agent approaches improve gameplay through prompting and preference optimization, they often optimize actions and in-game speech without explicitly grounding them in such beliefs. This frequently leads to strategically inconsistent behavior, especially for compact LLM agents. We introduce Multi-Agent Relational Belief Optimization (MARBO), a belief-grounded preference optimization framework that leverages relational beliefs to guide strategic decisions and in-game speech. MARBO provides preference feedback only when behaviors are supported by reliable relational beliefs and lead to strategically favorable social outcomes, encouraging more consistent learning under uncertainty. Experiments on representative SDGs show that MARBO enables compact LLM agents to consistently outperform existing baselines. The Code is available on https://github.com/PleaseTakemeAway/MARBO.

Sparse Bayesian learning is widely used for sparse linear inverse problems, yet its large-system stationary behavior remains poorly understood because all variance hyperparameters are estimated from the same data. We study classical sparse Bayesian learning, formulated as evidence maximization (type-II maximum likelihood), for underdetermined linear models with sensing matrices having independent and identically distributed Gaussian entries, Gaussian measurement noise, and an unknown deterministic signal sequence. Hyperparameter reoptimization induces a nonvanishing feedback term: a typical coordinate obeys a reoptimization-corrected scalar Gaussian law whose signal coefficient is governed by the normalized adaptive response rather than the frozen resolvent trace. A one-coordinate leave-one-out construction gives an exact conditional Gaussian law, which is transferred to a selected full stationary branch without assuming asymptotic closeness of the reduced and full stationary vectors. Combining this law with the Karush–Kuhn–Tucker conditions of the evidence objective yields a generally set-valued scalar relation and three branchwise large-system consistency relations. If the model noise variance is jointly estimated by evidence maximization, interior joint stationarity yields an exact finite-dimensional equality between the normalized residual energy and normalized resolvent trace. When the limiting signal law has nonzero mass at zero, this identity further yields a parameter-free asymptotic chi-square null law. Under an additional differentiability condition on the selected scalar branch, the large-system characterization also gives a closed relation for the reconstruction error of the posterior mean. The analysis is stationary-point based and permits multiple stationary branches.

Large language models (LLMs) are often considered fragile under aggressive sparsification, and maintaining reliable performance typically requires sticking to moderate sparsity levels. However, recent studies suggest that LLMs are more resilient to high sparsity than previously thought, reframing the problem as a design challenge rather than a fundamental limitation. In this work, we challenge the perceived limits of unstructured post-training LLM pruning by revisiting elementary pruning strategies that have remained relatively underexplored at this scale. Through a progressive sparsification framework with second-order saliency and continued training coordinated with sparsity progression, we show that pretrained LLMs can retain strong performance far beyond commonly studied sparsity regimes. Across LLaMA-2 and Qwen-3 model families, our approach improves perplexity and downstream accuracy up to 99% sparsity, surpassing both the current state-of-the-art and representative baselines. Precisely, on LLaMA-2-7B, our approach achieves WikiText-2 perplexities of 13.48 and 19.67 at 95% and 99% sparsity, respectively, while delivering 3.23$\times$ decoding speedup and 6.21$\times$ memory savings at 95% sparsity. Taken together, our results show that LLMs can be pushed into extreme sparsity while retaining strong performance, providing a foundation for further improving sparse models in this regime.

Mobile vendors and application developers increasingly deploy LLMs on smartphones for diverse prefill-only services. Yet current systems rely mainly on dense models whose regular computation maps efficiently to mobile NPUs, leaving more capable MoEs underused. MoE prefill does not fit mobile NPUs: NPU graphs are fixed at compile time, yet MoE picks experts at runtime; and one request touches most experts, more than a phone can hold in memory. We present EStream, which resolves both by separating what the NPU must fix from what MoE decides at runtime. A single compiled expert graph serves every expert, with each expert's routed tokens and weight address bound at call time, so dynamic MoE execution runs entirely on the NPU without padding or CPU/GPU fallback. Expert virtualization keeps the expert pool in UFS flash storage and pages it through a fixed-size NPU-addressable arena, group by group, with loading hidden behind computation, so memory is bounded by the arena rather than by the model. It further introduces a hardware-aware configuration algorithm that automatically configures the UFS–NPU pipeline and maximizes loading–computation overlap. Across 18 comparative settings covering three 7B–16B MoEs and 256–4,096-token prompts, we evaluate EStream on a commercial Snapdragon smartphone. Compared to the fastest baseline at each setting, EStream achieves a 2.25–27.57X pure-prefill TTFT speedup and reduces peak physical memory by 1.19–12.29X. EStream further scales to MoE models with up to 46.7B parameters.

Real-time 3D mapping is fundamental for autonomous robotic navigation, with Euclidean Signed Distance Fields (ESDFs) serving as the standard representation for online motion planning. While recent advancements in non- projective distance fields yield highly accurate maps, their computational overhead remains a severe bottleneck. Conventional integrators redundantly re-fuse millions of depth pixels every frame, even long after the corresponding voxels have converged, wasting significant computational resources in environments dominated by large planar surfaces. In this paper, we present SHIFT (Surface-aware High-speed Integration For TSDFs), an efficient mapping framework designed to reduce this per-frame update cost. By exploiting structural redundancy directly from 3D depth geometry, SHIFT compresses flat local regions into weighted super-rays and freezes flat-voxel gradients. A compact ESDF voxel layout further reduces the memory footprint of the remaining wavefront. Extensive evaluations across various RGB-D and LiDAR sequences show that SHIFT cuts TSDF cost by 1.42 to 4.07 times, while holding mesh error within millimeters, and reduces ESDF-layer memory by up to 28%

Although machine learning can be used to predict the evolution of physical systems from data, a formulation that learns only the system state at each time leaves the temporal and dynamical structure of the solution to be resolved within a broad hypothesis space. We introduce Learning$^2$, a representation-level framework that structures this space by coupling a primary representation to a second representation through a known physical transformation. The resulting cross-representation constraint restricts the effective hypothesis space and provides an ante-hoc, physically interpretable criterion for excluding solutions that satisfy the primary representation alone. We instantiate Learning$^2$ through EuLaNet, an Eulerian–Lagrangian representation for fluid dynamics. Given the velocity state $u(\mathbf{x},t)$, EuLaNet constructs its induced Lagrangian flow map $X(\mathbf{a},t)$ through $\dot{X}(\mathbf{a},t)=u(X(\mathbf{a},t),t)$, from which material transport and finite-time deformation are derived. The resulting representation couples the predicted state to the dynamical consequences it induces, providing a second consistency criterion beyond state-level agreement. We formalize this construction through an effective hypothesis space $\mathcal{H}_{L^2}\subseteq\mathcal{H}$ and define the conditions under which a consequence representation provides discriminative constraints on candidate solutions. EuLaNet is implemented as a model-independent representation module, separating the physical constraint from the downstream learning architecture. This construction provides an ante-hoc mechanism for physically interpretable constraint in scientific learning and offers a basis for developing and evaluating broader classes of Learning$^2$ architectures. The implementation is open-sourced to support the development and extension of the architecture across scientific domains.

Large language models (LLMs) are increasingly used to generate media, but whether their content perpetuates gender stereotypes is unknown: standard benchmarks rely on selection-based formats rather than long-form generation, and surveyed baselines for local gender associations are scarce outside the West. We collect gender associations for 22 occupational and domestic roles from 695 respondents across the United States, India, Kenya, and Nigeria, and evaluate eight LLMs under two regimes: direct questioning and media generation. Models track the surveyed associations under direct questioning but skew substantially more male under media generation in major local-language cells, consistent with the male bias documented in human-produced media. Outside the US, the shift is much smaller and non-significant under English prompting, so English-only or country-agnostic evaluation would miss this bias in the languages where these models are most deployed. Instruction prompting reduces the shift directionally, but trades off against alignment with the surveyed associations. Evaluating LLM gender bias for global deployment therefore requires generation-format testing, local-language prompting, and locally-collected human baselines.

Enterprise AI is evolving into an Enterprise Operating System where autonomous AI agents can plan, reason, use memory, invoke tools, execute workflows, and collaborate with other agents. This shift creates a new governance challenge: existing authorization, security, guardrails, and compliance mechanisms are fragmented and are not designed to govern autonomous AI as a unified system. This paper introduces the Unified Policy Architecture (UPA), a governance architecture for Enterprise AI Operating Systems. UPA provides a unified policy model for governing AI and agents, tools, workflows, memory, enterprise resources, and agent-to-agent interactions and enterprise business rules. It extends policy control beyond authorisation to include runtime obligations, human approvals, compliance, audit evidence, and governance evaluation. We present UPA's governance model, declarative policy language foundations, policy evaluation semantics, extensible plugins, industry policy packs, and an evaluation framework for enterprise governance. We also identify extensions for multi-agent coordination, provenance-aware policies, and stateful runtime governance. UPA provides a foundation for building secure, accountable, and governable Enterprise Operating Systems for autonomous AI.

The evolution of modern-day communication networks towards enhanced capacity and robustness is driving interest in topological light wave as a next-generation information carrier. However, harnessing its potential is hindered by the lack of an effective topological reader, particularly when strong disorder renders conventional readout intractable. Here we address this by introducing TOPO$^{2}$, an artificial intelligence framework that leverages on topological learning to recover topological information from light. We demonstrate it on highly distorted optical fields with embedded Skyrmion topologies, showing that we can extract the topological numbers in a single shot with just a conventional camera. We showcase its capability in a TOPO$^{2}$-based communication scheme, sending and recovering information encoded in a topological alphabet, confirming that it outperforms conventional approaches. Our connection of topological photonics with topological AI as a powerful paradigm for robust high-capacity information encoding and processing opens a new path towards topological information optics.

Large language models (LLMs) are increasingly deployed in safety-critical applications, yet jailbreak attacks can conceal harmful intent through role-playing, fictional scenarios, or seemingly benign motivations. Existing inference-time defenses may miss disguised attacks or excessively refuse legitimate requests. We propose SRD-GUARD, a parameter-free, black-box defense framework that exposes concealed intent through semantic rewriting and consensus-based risk assessment. Given an input prompt, SRD-GUARD generates five semantically related rewrites that preserve the underlying objective while removing unnecessary contextual packaging. The original prompt and rewrites are jointly evaluated by multiple independent LLM-based safety scorers on a continuous risk scale. A decision module combines absolute risk thresholds with relative risk changes between the original and rewritten prompts to adaptively intercept, preserve, or warn on requests. We evaluate SRD-GUARD against UNIATTACK, CIPHER, and DeepInception on Llama-3-8B-Uncensored and DeepSeek-V4-Flash using AdvBench and OR-Bench-Hard. SRD-GUARD achieves average DSRs of 91.44% and 100%, with ORRs of 8.00% and 12.00%, respectively. Compared with evaluated baselines, it provides a more favorable DSR–ORR trade-off. Ablation studies show that rewriting exposes concealed harmful intent, joint scoring improves robustness to individual evaluator behavior, and risk-adaptive decision making enables selective handling of ambiguous inputs. These results demonstrate that semantic intent exposure, consensus-based risk assessment, and relative-risk-aware routing provide an effective and selective approach to black-box jailbreak defense. The artifact is available at https://anonymous.4open.science/status/CICD-Guard-D648.

Typing passwords on smartphones in public places exposes users to video-based side-channel attacks, in which an adversary records the typing session and reconstructs the entered password by analyzing finger movements. Prior video-based keystroke inference attacks have targeted free-form text on tablets, numeric PINs, pattern locks, and passwords under unrealistically simplified conditions. We present the first general video-based keystroke inference pipeline that recovers rule-based alphanumeric passwords from smartphone QWERTY keyboards, covering all four keyboard layouts and requiring no assumptions about the victim or their device. Using the built-in camera of the popular Meta Ray-Ban smart glasses, our pipeline tracks the device and typing fingertips at sub-pixel resolution, predicts keystrokes via a self-supervised ensemble of neural networks, dynamically estimates the on-screen keyboard layout, and computes per-key probability distributions. These video-derived probabilities are combined with prior password-typing distributions based on 2.19 billion leaked passwords to estimate the rank, and therefore the crack time, of the observed password. In a user study using passwords that adhere to a typical enterprise password policy, human-chosen passwords of up to 16 characters can be cracked in hours to days, and password-manager-chosen random passwords of up to 13 characters in under an hour on a modern GPU, reducing password entropy by up to 60 bits relative to the prior baseline. For human-chosen passwords, combining video evidence with prior statistics yields an additional statistically significant reduction of approximately 20 bits compared to either source alone. The attack succeeds at distances up to 1.8 m, across diverse attacker-victim postures and viewing angles, and generalizes across three popular smartphones.

Recent developments have shown that eigenvalue localization regions can often be significantly sharpened by identifying and excluding subsets that are guaranteed to contain no eigenvalues, leading to the concept of inclusion-exclusion localization sets. In this paper, we introduce a new family of inclusion-exclusion eigenvalue localization sets based on the S-strict diagonal dominance framework and the corresponding CKV localization regions. For an arbitrary nonempty subset S of the index set, we construct novel exclusion regions and derive the associated S-SDD inclusion-exclusion localization sets. We prove that every eigenvalue of a complex matrix belongs to the proposed regions and define a refined localization set obtained by intersecting the corresponding inclusion-exclusion regions over all nonempty subsets S. The proposed framework unifies and extends several existing results. In particular, we show that the classical Gershgorin inclusion-exclusion localization set and the Dashnic-Zusmanovich inclusion-exclusion localization set arise naturally as special cases of our construction. Consequently, the new localization set provides, in general, a sharper eigenvalue inclusion region than both of these previously known localization regions. The theory is further extended to matrix polynomials, yielding corresponding inclusion-exclusion localization results for polynomial eigenvalue problems. Several numerical examples illustrate the effectiveness of the proposed approach and demonstrate the improvement achieved over existing localization sets.

Sequential sounding in UAV fluid-antenna systems (FASs) provides additional spatial information but delays transmission, causing earlier channel observations to age. This paper addresses the resulting information–freshness tradeoff by jointly determining where to probe and when to stop under blockwise hardware constraints. A dynamic Karhunen–Loève estimator aligns asynchronous measurements with the transmission state, covariance information gain selects feasible probe locations, and an age-discount identity with a local one-more-slot condition characterizes the covariance-level balance between information and freshness. Paired simulations show that temporal alignment recovers most of the lower-tail reliability lost by static stacking, covariance-aware probing ranks highest numerically among the evaluated policies, and optimized one-slot sounding becomes statistically competitive with two-slot alternatives at the highest tested mobility. Within the evaluated schedule family, increasing mobility shifts the competitive operating region toward shorter scans, supporting joint probe-placement and sounding-duration design.

We introduce 3DHarnessBench, a benchmark that evaluates the agentic ability of frontier vision-language models (VLMs) to recover 3D geometry as Blender Python code from a variety of inputs. Unlike previous frameworks that prompt the VLMs with a fixed input (e.g., a single rendering or a text description), 3DHarnessBench evaluates four separate harness settings that progressively enable active agentic exploration, facilitated by recent Blender MCP functionality. Our hierarchy from Single-view, Multi-view, Active Visual (arbitrary viewpoint access), and Full 3D Interaction (complete access to the target object through Blender function calls) probes the models' abilities in both visual perception and active inference, tool calling, and self-correction. We observe that the ability of all frontier models to recover 3D geometry improves significantly with richer function call access, although the improvements are strongly model-dependent, revealing highly uneven agentic 3D-to-code capabilities. We will release the benchmark, code, outputs, and agent trajectories for reproducible 3D evaluation.

Continual learning is a crucial capability for Graphical User Interface (GUI) agents to adapt to evolving applications while retaining knowledge acquired from previous applications. Such application streams pose a challenging knowledge modeling problem: new applications often share underlying knowledge with past ones, yet also introduce specific knowledge that must not interfere with historical knowledge. In this paper, we propose activation-conditioned selective knowledge control, a lightweight method that achieves selective knowledge retention via neuron-level gradient manipulation. Our method maintains a compact historical knowledge state to protect highly activated MLP neurons that preserve previous knowledge. When a new application arrives, it performs real-time gradient surgery conditioned on forward activation. Concretely, the protected neurons are categorized into two types: unactivated neurons holding specific knowledge, whose gradients are truncated to prevent interference; and activated neurons holding shared knowledge, whose gradients are orthogonally projected to preserve stability while enabling adaptation. After each application stage, newly identified critical neurons are merged into the historical state for future learning. Empirical evaluations on multi-app sequential benchmark demonstrate that our method effectively mitigates catastrophic forgetting on prior applications while sustaining robust adaptation to new ones.

Large language models (LLMs) have shown strong potential for translating natural-language (NL) requirements into PL/SQL programs, attracting increasing attention from the database community. However, existing NL-to-PL/SQL efforts primarily focus on directly generating PL/SQL from complete NL requirements. In practice, PL/SQL development involves diverse scenarios, such as from-scratch development, code modification, debugging, and optimization, and may require either direct generation or multi-turn interaction. Yet, no comprehensive benchmark evaluates multi-scenario, direct and interactive, and multi-dialect NL-to-PL/SQL development. In this paper, we present ProcArena, an execution-based benchmark covering both Direct and Interactive modes. ProcArena comprises 3,998 executable tasks over 157 databases, spanning nine development subscenarios in PostgreSQL and Oracle. We construct challenging Direct tasks through Iterative Logic Enhancement and scenario-specific adapters, and derive paired Interactive tasks through Knowledge Integration and Requirement Perturbation while preserving executable targets. We further design a controlled Solver-User Simulator protocol that allows models to clarify user intent and inspect the database environment without exposing hidden execution feedback. Evaluating seven language models, we find that the best average scores are only 62.2% and 57.8% in Direct and Interactive, respectively, demonstrating that realistic NL-to-PL/SQL development remains challenging, particularly in interactive settings.