Forskningsradar

Science Journals

Peer-reviewade publikationer — 62343 artiklar

arXiv (fysik & CS) Preprint 2026-06-30
Minimizing cumulative infections in SIS epidemic models over networks via an edge deletion algorithm
arXiv:2606.30142v1 Announce Type: new Abstract: In this paper, we investigate the discrete SIS (Susceptible-Infected-Susceptible) models. We focus on minimizing epidemic spreading over networks by extending an existing edge deletion algorithm to the SIS model. To achieve this, we employ the mean-field approximation to linearize the network dynamics into a deterministic SIS model. We analytically demonstrate that the total number of infections is upper-bounded by a super-modular function, thereby ensuring the efficiency of the edge-deletion approach. To evaluate the proposed method, we conduct experiments on synthetic Erdos-Renyi networks and the real-world dataset collected from BBC Pandemic Haslemere app. Numerical simulations validate our theoretical results, confirming that both configurations converge to the stable, disease-free equilibrium.
arXiv (fysik & CS) Preprint 2026-06-30
Defending Against Harmful Supervision Hidden in Benign Samples
arXiv:2606.30263v1 Announce Type: new Abstract: Existing defenses are effective when harmful content is explicitly mixed into downstream fine-tuning data, but crafted samples can instead hide harmful supervision inside benign tasks. We propose Embedded Attack, where harmful QA pairs are embedded within benign training samples, and show that representative guardrails often fail to detect them at the example level. To address this, we propose Dual-Reference SFT (DR-SFT), which adapts DPO-style contrastive objective design to SFT through token-level regularization, mitigating harmful fine-tuning beyond coarse data filtering.
arXiv (fysik & CS) Preprint 2026-06-30
How do Execution Features Improve Statistical Fault Localization? An Empirical Study
arXiv:2606.30324v1 Announce Type: new Abstract: Automated fault localization helps developers find faults in large code bases. Statistical fault localization (SFL) ranks suspicious lines from pass/fail spectra, but line execution alone misses information like data-flow, values, or branch conditions that explain why a failure occurs. This study evaluates whether augmenting SFL with execution features improves localization accuracy and developer-oriented inspection effort. We extract execution features with EFDD for all Tests4Py subjects, train per-subject random forests, map importances to source lines, and combine the resulting weights with established SFL formulas. The evaluation measures reference-patch accuracy, line- and function-level effort, robustness, and feasibility using a confounder-adjusted mixed-effects model, corroborated by paired statistical tests and outcome-neutral quality checks.
arXiv (fysik & CS) Preprint 2026-06-30
SA-Homo: Scale Adaptive Homography Estimation for Scale Variation Scenarios
arXiv:2606.30408v1 Announce Type: new Abstract: Homography estimation, as one of the fundamental problems in computer vision, remains challenged by scale variation scenarios where image pairs potentially exhibit significant scale discrepancies. Existing deep learning frameworks frequently suffer from a significant performance degradation in such cases, as they rely on limited displacement assumptions and local feature consistency that might not hold under large scale gaps. In this paper, we propose SA-Homo, a novel scale-adaptive homography estimation framework designed to achieve robust alignment across a wide range of scale discrepancy ratios. We adopt a hierarchical scale alignment strategy that transitions from the global perspective with a heavy module to a local perspective with a light module. Specifically, we introduce the Scale-aware Discrepancy Bridging Module (SDBM) for initial alignment, which utilizes a Multi-scale Linear Attention Cascade (MLAC) to capture long-range dependencies and mitigate feature inconsistencies, along with a global Cross-scale Similarity Matrix Block (CSMB) for scale robust correlation representation. Once the initial scale gap is bridged, a lightweight Iterative Homography Estimation Refinement Module (IHERM) progressively polishes the result using local correlations. To facilitate this research, we contribute the HMSA dataset, a high-resolution, multi-modal satellite benchmark specifically tailored for scale-variant challenges. Extensive experiments demonstrate that SA-Homo maintains high precision even under 8$\times$ scale discrepancies, outperforming state-of-the-art methods in both conventional scale-similar scenarios and challenging scale variation scenarios. Code and collected datasets are available at https://github.com/shangxuanx330/SA_Homo
arXiv (fysik & CS) Preprint 2026-06-30
Are Whitepaper Claims Reflected in Market Structure? A Contamination-Aware Pipeline and a Power-Limited Null
arXiv:2601.20336v5 Announce Type: replace-cross Abstract: Do the functional narratives in cryptocurrency whitepapers correspond to how their tokens behave in markets? We develop a content-verified, contamination-aware pipeline for measuring structural correspondence between project narratives and market structure, and report two results. The first is a cautionary one. An apparent entity-level signal in an earlier version of our corpus -- specialised tokens appearing to align more strongly than broad infrastructure tokens -- was entirely an artifact of corpus contamination: roughly a quarter of the documents were failed-download stubs or wrong-document whitepapers (for example, a "Cosmos" entry that was in fact Binance Smart Chain text), and the apparent ordering does not survive content verification: on the clean corpus no token registers as helping alignment. We therefore report it as a contamination diagnosis, not a finding. The second is an honest null. Combining zero-shot NLP classification of 43 content-verified whitepapers across 10 semantic categories with seven cross-sectional market-structure statistics computed from hourly data (17,543 timestamps, 2023-2024), and aligning the two spaces with Procrustes rotation and Tucker's congruence coefficient ($\phi$), we do not detect a significant claims-market alignment in this $n = 43$ sample (dimension-matched $\phi = 0.303$, zero-padded $\phi = 0.223$; both non-significant). A positive-control and power analysis shows the binding constraint is the low reliability of the text instrument: the minimum detectable effect is $\phi \approx 0.66$, well above the observed $\approx 0.22$. This is absence of evidence for alignment, not evidence of its absence -- we can reject strong alignment ($\phi \geq 0.70$) but cannot distinguish weak alignment ($\phi \approx 0.3$) from none.
arXiv (fysik & CS) Preprint 2026-06-30
Sustainable Hybrid Document-Routed Retrieval for Financial RAG: Resolving the Robustness-Precision Trade-off
arXiv:2603.26815v3 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) systems for financial document QA typically follow a chunk-based paradigm: documents are split into fragments, embedded, and retrieved by similarity. In structurally homogeneous corpora such as regulatory filings, this suffers from cross-document chunk confusion. Semantic File Routing (SFR), which uses LLM structured output to route queries to whole documents, reduces catastrophic failures but sacrifices targeted-chunk precision. We identify this robustness-precision trade-off on the FinDER benchmark (1,500 queries across five groups): SFR achieves higher average scores (6.45 vs. 6.02) and fewer failures (10.3% vs. 22.5%), while chunk-based retrieval (CBR) yields more perfect answers (13.8% vs. 8.5%). To resolve it, we propose Hybrid Document-Routed Retrieval (HDRR), a two-stage architecture that uses SFR as a document filter followed by chunk retrieval scoped to the identified document(s), eliminating cross-document confusion while preserving chunk precision. HDRR achieves the best performance on every metric: an average score of 7.54 (25.2% above CBR, 16.9% above SFR), a 6.4% failure rate, 67.7% correctness (+18.7 pp over CBR), and a 20.1% perfect-answer rate (+6.3 pp over CBR, +11.6 pp over SFR), simultaneously attaining the lowest failure rate and highest precision across all five groups. Beyond accuracy, HDRR is also the most efficient of the high-quality systems: it preserves CBR's compact per-query token budget (~5K-15K, an order of magnitude below SFR's ~50K-200K), incurs no indexing-time LLM spend (versus the one-time ~$100 cost of contextual indexing), and uses fewer per-query LLM calls than self-correcting agentic baselines, translating directly to lower API spend and inference-time energy at deployment scale.
arXiv (fysik & CS) Preprint 2026-06-30
COHORT: Collaborative Orchestration for Hardening via Offensive Replay on Emulated Topologies
arXiv:2606.30479v1 Announce Type: new Abstract: Mitigating an observed adversary in an enterprise network typically takes weeks of expert work: an analyst derives a mitigation tailored to that adversary, validates it without breaking production, and verifies it disrupts the specific attack. The procedure relies on expert judgment and cannot safely be exercised against the production network. COHORT is the first end-to-end framework to automate this procedure for deployable mitigations. A role-decomposed multi-agent LLM workflow proposes candidates, implements them as real device commands, and refines them through a critique loop, all on a high-fidelity GNS3 emulator running real vendor firmware (firewall, switch, router). Each candidate is evaluated by offensive replay: re-executing the original adversary on the mitigated network for a paired comparison against the unmitigated baseline, rather than the reward-signal or expert-judgment proxies used in prior simulation, hybrid, and configuration-generation work. Two further checks complement replay: a connectivity-regression check (LAN ping and internet HTTP probe) rejects mitigations that disrupt legitimate LAN or internet connectivity, and a cumulative evaluation stacks approved mitigations onto a persistent state to surface compound effects. Across three topologies and four attack scenarios (ransomware, lateral movement, DNS exfiltration, data theft), 46.7% of generated mitigations both disrupt the attack and preserve connectivity under replay, 4.4 times the rate of a single-agent baseline using the same model and tool access. A demo video walking through the framework is available with our released artifacts.
arXiv (fysik & CS) Preprint 2026-06-30
Offline accuracy is not enough: closed-loop instability and stabilisation of a wall-sensor neural estimator in opposition control
arXiv:2606.30484v1 Announce Type: new Abstract: Opposition control reduces skin-friction drag by opposing the wall-normal velocity on a near-wall detection plane, but the detection-plane velocity it requires is not available from wall-mounted sensors. Wall data can reconstruct inner-flow quantities accurately when assessed offline on a fixed flow state, and we ask whether such a reconstructed field can instead serve as a live surrogate sensor inside the feedback loop. We train a recurrent estimator to infer the detection-plane velocity from the two wall-shear-stress components in opposition-controlled turbulence. Offline it performs extremely well, reaching a correlation of 0.99 and near-unity coherence across the energetic scales; yet the same estimator fails in closed loop, decorrelating from the true field within a few viscous time units as the control collapses. The failure is not one of accuracy but of distribution shift induced by the controller itself: small closed-loop errors carry the flow off the attractor represented in the training data, while unresolved high-wavenumber errors enter through the wall boundary condition and return as out-of-distribution inputs. Standard remedies such as low-pass filtering and exponential averaging only delay numerical breakdown while accelerating decorrelation. Stable wall-only control is recovered by imposing spectral consistency on the deployed actuation and retraining the estimator on its own closed-loop data, giving a controller that holds much of the drag reduction of ideal opposition control from wall quantities alone. The obstacle is not whether the near-wall flow can be reconstructed offline, but whether that reconstruction stays dynamically consistent when allowed to modify the flow it senses.
arXiv (fysik & CS) Preprint 2026-06-30
Sequential Planning via Anchored Robotic Keypoints
arXiv:2606.30613v1 Announce Type: new Abstract: We present Sequential Planning via Anchored Robotic Keypoints, SPARK, a training-free neurosymbolic manipulation system that reaches 43.7% on six LIBERO-PRO position \& task cells, more than doubling CaP-Agent0 and Vision-Language-Action (VLA) baselines. CaP-Agent0, a multi-turn code-generation agent, achieves 18.2% by re-querying an LLM at every turn, but its restart-from-scratch solution proves costly against minor policy failures. Perception is the layer that fails most under position and task changes so SPARK spends its computation there. A single Gemini call composes the plan as a typed behavior tree (BT) of composable primitives, each already containing the low-level control (motion, grasping, depth geometry) a code-generation agent would otherwise regenerate on every trial. The rest of the budget goes to perception: a second Gemini call proposes three alternative text prompts per object, SAM3 evaluates each, and we keep the prompt$\to$label pair with the most confident detection and a recovery loop then retries a failed primitive against freshly detected objects, with no new LLM call. The alternative prompts add +27.7 points on the spatial suite and +10.0 on the object suite, with the recovery loop adding +5.0 overall. SPARK runs the same primitives on three robot families (UR10e, Franka FR3, bimanual Franka) across nine unique tasks at twenty trials each, averaging 68%. Since the detector, planner, and controller modules sit behind the typed plan, they swap independently without training, and each primitive's checkable post-condition traces a failure to the corresponding module or a kinematic limit. Every trial logs a verified, labeled trajectory, so a training-free planner that already beats VLAs can supply the data those policies need without teleoperation. Project page: https://cwru-aism.github.io/spark-page/
arXiv (fysik & CS) Preprint 2026-06-30
Hard-constraint physics-residual networks for hydrogen crossover prediction and high-pressure extrapolation in PEM water electrolysis
arXiv:2511.05879v5 Announce Type: replace Abstract: Hydrogen crossover is a critical safety and efficiency constraint in high-pressure polymer electrolyte membrane water electrolysis (PEMWE), but accurate prediction remains difficult because data are limited, transport physics are strongly coupled, and industrial operation requires reliable extrapolation beyond observed conditions. This study develops a hard-constraint physics-residual network (PR-Net) for hydrogen crossover prediction in PEMWE and compares it with a purely data-driven neural network (NN) and a soft-constraint physics-informed neural network (PINN). PR-Net embeds Henry's, Fick's, and Faraday's laws as a deterministic backbone and learns only a residual correction for unmodelled nonlinear effects. The benchmark includes 184 observations from eight peer-reviewed sources across six membrane types, covering 1-200 bar, $25-85{\deg}C$, and $0.05-5.0 A cm^{-2}$. PR-Net achieves $R^2 = 99.57 \pm 0.16%$, with 9-fold lower prediction variability than NN and PINN. In pressure-axis extrapolation, PR-Net attains $R^2 = 94.02 \pm 0.92%$ at 200 bar, 2.5 times beyond the training pressure range, compared with $68.06 \pm 5.52%$ for PINN and $58.00 \pm 8.60%$ for NN (p < 0.001). Residual analysis indicates that the learned correction captures part of the high-pressure gas-phase non-ideality and recovers a transport-regime transition near $0.23 A cm^{-2}$ between Fickian diffusion-dominated and Faradaic production-dominated transport. With a computation time of $1.08 \pm 0.34 ms$ on low-power embedded hardware, PR-Net provides a practical framework for real-time crossover monitoring, adaptive process control, and safer high-pressure green-hydrogen operation.
arXiv (fysik & CS) Preprint 2026-06-30
Dual Magnetic and Electric Dipole Symmetry: Pseudo Angular Momentum in Parity Space and the Electric Land\'e $g$-Factor
arXiv:2511.07692v3 Announce Type: replace Abstract: Electric dipole moments (EDMs) are sensitive probes of fundamental symmetries and central to searches for physics beyond the Standard Model. We present a symmetry-based, Zeeman-analogue operator framework that places magnetic and electric dipole physics on parallel footing under electromagnetic duality, and introduce a polar-sector pseudo-angular-momentum degree of freedom in parity space together with an associated electric Land\'e factor that organizes induced orbital dipoles. Following Ohanian's effective-current formulation of the Zeeman effect, we construct its electric dual: the wavefunction's microscopic polarization admits an equivalent effective magnetic probability-current representation, providing a field-equivalence description of parity-mixed charge displacement. In this notation the total EDM expectation takes the unified form $\langle \hat{\vec d}_{\rm tot}\rangle= d_B(g_E\,\frac{\langle \hat{\vec J}_p\rangle}{\hbar}+ g_E^{e}\,\frac{\langle \hat{\vec S}\rangle}{\hbar})$, with $g_E^{e}=\frac{2d_{\rm int}}{d_B}$, where $\hat{\vec J}_p$ captures Stark-induced pseudo-angular momentum and $\hat{\vec S}$ encodes any intrinsic (spin-aligned) EDM $d_{\rm int}$ from symmetry-violating interactions. We define a natural electric dipole unit (the ``Bohr EDM'') as $d_B \equiv e a_0=\frac{2\mu_B}{c\alpha}$ ($a_0$ the Bohr radius and $\mu_B$ the Bohr magneton). As a canonical analytic benchmark, we show in the hydrogenic problem that a static electric field couples within a fixed $n$ manifold through the scaled Runge--Lenz structure, yielding a compact Land\'e-like description and reproducing the Stark doublet (e.g.\ $|\langle d_{\rm orb}\rangle|=3d_B$ for the $2s$--$2p_{m=0}$ mixing).
arXiv (fysik & CS) Preprint 2026-06-30
Robust Sparse Identification of Nonlinear Dynamics via Least Trimmed Squares
arXiv:2606.28584v1 Announce Type: cross Abstract: In this work, we propose a robust Sparse Identification of Nonlinear Dynamics (SINDy) pipeline for handling datasets corrupted by noise and outliers. The method decouples outlier filtering from sparse regression by combining Iterative Least Trimmed Squares (ILTS) with Sequentially Thresholded Least Squares (STLS). Unlike standard approaches that treat all observations uniformly within a single regression stage, the proposed ILTS-SINDy framework first applies an ILTS procedure that iteratively minimizes the sum of the smallest squared residuals to identify the most reliable observations without prior knowledge of outliers, after which STLS is used to recover a parsimonious governing model. Extensive numerical experiments show that ILTS-SINDy can significantly outperform existing robust SINDy variants across a range of outlier contamination levels, with performance maintained even under settings with up to $20\%$ corrupted observations.
arXiv (fysik & CS) Preprint 2026-06-30
Classical Coherence and Biological Aging
arXiv:2606.28838v1 Announce Type: new Abstract: In previous work it was argued that the cells of a multicellular organism form a classically coherent system and that such coherence is essential for life. Here we make this claim precise by introducing an explicit classical formalism in which a many-cell system is represented by a single state vector in an abstract DNA code space. Using Dirac's bra-ket notation purely as a compact representation of classical states, we construct an analogue of the center-of-mass coordinate that encodes the organismal identity and show how a common genetic code shared by all cells corresponds to a coherent phase in this space. We then map this structure onto DNA sequence space by introducing a classical Biological Hamiltonian whose generalized coordinates encode DNA codes and their cell-wise distribution, so that the organismal identity is represented by a global code state rather than by individual molecular constituents. Within this framework we define a time-dependent maintenance operator with code-correcting and code-breaking terms, weighted by coefficients $A(t)$ and $B(t)$, which captures the balance between restorative dynamics and environment-induced damage to the code. Aging is described as a slow drift in these control parameters: as $A(t)$ decreases and $B(t)$ increases, the identity state becomes less stable and the organism moves from robust code coherence to stochastic code variability. In this picture, death appears as a transition in which the global identity state can no longer be maintained.
arXiv (fysik & CS) Preprint 2026-06-30
DLGStream: Dynamic Language-embedded Guassian Splatting for Open-vocabulary Enabled Free-viewpoint Video Streaming
arXiv:2606.28840v1 Announce Type: new Abstract: 3D Gaussian Splatting~(3DGS) has emerged as a promising paradigm for reconstructing streamable free-viewpoint video~(FVV) from multi-view videos. However, 3DGS-based FVVs typically lack user interaction and editing capabilities, which diminishes the immersive experience. Recent research has integrated language features from CLIP into 3DGS via distillation, enabling open-vocabulary queries and supporting many downstream applications. Nevertheless, the stringent requirements of FVV, low frame size and high FPS, make current language Gaussian representations unsuitable for language-embedded FVV. In this paper, we propose DLGStream, a novel language-embedded FVV representation that streams time-varying language features alongside Gaussian attributes to support 4D environment interaction, scene editing, and spatial intelligence. Specifically, we propose a dual-opacity dynamic language Gaussian representation, which maintains two opacity attributes for color and language features to deal with performance degradation that occurs when colors and features are jointly optimized. Furthermore, we introduce an interpolation-based deformation field to reduce temporal redundancy. This deformation field can also be used for 4D frame interpolation, boosting FVV sequences from low to high FPS. Experimental results demonstrate that DLGStream achieves superior performance in both on open-vocabulary segmentation and reconstruction quality with an average frame size of merely 43 KB. The code is available on \href{https://github.com/kkkzh/DLGStream}{https://github.com/kkkzh/DLGStream}.
arXiv (fysik & CS) Preprint 2026-06-30
An Integrated Machine Learning and Hierarchical Variance Decomposition Pipeline for Student Performance Prediction and Metacognitive Calibration on Multi-Signal Telemetry
arXiv:2606.28881v1 Announce Type: new Abstract: Predicting student performance and characterizing metacognitive calibration are essential for personalization in intelligent tutoring systems. Prior research treats performance prediction, calibration error calculation, and variance decomposition as separate pipelines, preventing unified interpretation. I propose the Unified Behavioral Prediction and Calibration Analysis Pipeline (UBP-CAP), an integrated framework processing student pre-execution behavioral telemetry through three linked modules: (1) a LightGBM classifier with SHAP for binary correctness prediction, (2) formal calibration metrics (ECE, MCE, and Brier score decomposition) to evaluate metacognitive alignment, and (3) a crossed Generalized Linear Mixed-Effects Model (GLMM) for decomposing calibration deviations. I introduce the Predictive-Explanatory Divergence Index (PEDI), which quantifies structural divergence between predictive and explanatory feature profiles. Evaluated on 1,195 interaction records (27 students, 45 tasks), Logistic Regression achieves AUC-ROC = 0.903, outperforming LightGBM (0.878). Student naive ECE (0.109) significantly exceeds model ECE (0.068), confirming systematic miscalibration. The crossed GLMM yields ICCStudent = 0.123, showing calibration is situational rather than dispositional. PEDIcos = 0.081 (p = 0.327) indicates structural alignment between prediction and explanation on shared behavioral features.
arXiv (fysik & CS) Preprint 2026-06-30
Unit-Circle Moment Closure
arXiv:2606.28894v1 Announce Type: new Abstract: Moment closure is a central problem in reduced descriptions of stochastic, kinetic, and quantum dynamics, where equations for low-order observables are coupled to an unresolved hierarchy of higher-order moments. Existing closures usually impose a prescribed form on the distribution or directly truncate the hierarchy, which can become inaccurate or unstable for strongly non-Gaussian states. Here we introduce unit-circle moment closure, which recasts the problem as analytic continuation. Raw moments are mapped to bounded unit-circle moments, whose unresolved tail is reconstructed by a Takagi-Prony procedure from the effective pole structure of a mapped generating function. The resulting continuation yields stable higher-order moments without assuming a fixed distributional ansatz. Illustrative static and dynamical examples demonstrate accurate reconstruction of non-Gaussian distributions and stable evolution of moment hierarchies. Our approach provides a general perspective for moment closure based on analytic structure rather than direct truncation.
arXiv (fysik & CS) Preprint 2026-06-30
MedEvoEval: Evaluating Continual Evolution of Doctor Agents through Simulated Clinical Episodes
arXiv:2606.28900v1 Announce Type: new Abstract: Doctor agents are moving beyond single-turn answer generation toward evolving clinical decision systems. Within an outpatient episode, they acquire evidence, use examination and consultation resources, and decide when to finalize a diagnosis and management plan. Across episodes, their behavior may change through memory, retrieval, reflection, or other update mechanisms. Current evaluations only partially cover this setting. Fixed-input medical QA benchmarks score final answers from complete inputs, whereas many interactive benchmarks still focus on individual encounters or fixed runs, providing limited support for evaluating how episode-level decisions interact with cross-episode experience. We introduce MedEvoEval, an executable longitudinal evaluation framework based on action-gated simulated outpatient episodes. Each source case is converted into role-specific patient, examination, and manager views; evidence is revealed only through valid actions; and each episode records a structured trace that links observations, actions, final outputs, manager scores, and optional experience write-back. We release a runnable E&D artifact with 700 processed episodes, provenance notes, schemas, an episode runner, scoring scripts, configurations, example logs, analysis code, and trajectory- and step-level derivatives. Experiments show that episode traces expose process costs hidden by final-answer scoring, show how MDT-style consultation reallocates resources, and support longitudinal analyses of memory maturation, held-out transfer, update-stage response, and backward retention. Together, these results show that MedEvoEval provides a concrete basis for evaluating whether doctor agents improve through experience, transfer useful behavior, and retain earlier capabilities over time.
arXiv (fysik & CS) Preprint 2026-06-30
Pressure-robust ALE space-time DG method for the Stokes equations on moving domains
arXiv:2606.28914v1 Announce Type: new Abstract: We propose and analyze a space-time discontinuous Galerkin method for the incompressible Stokes equations on moving domains within the arbitrary Lagrangian-Eulerian setting. We use a contravariant Piola map in the definition of the discrete velocity space to preserve the pointwise divergence-free property on the discrete level. We show that the method is inf-sup stable, with no constraints on the spatial mesh or the time partition. We also establish a priori error estimates in the energy norm for arbitrary degrees of approximation in space and time. For piecewise-constant and piecewise-linear approximations in time, we show that the method is also robust at low viscosity regimes, and provide numerical evidence suggesting that this property extends to high-order cases as well. We present several numerical experiments to validate our theoretical findings.
arXiv (fysik & CS) Preprint 2026-06-30
ML-Powered LDAP Reconnaissance Detection using Weak Supervision
arXiv:2606.28917v1 Announce Type: new Abstract: Lightweight Directory Access Protocol (LDAP) is a protocol that allows users to query and modify Active Directory (AD) data. By default, all users have read access to all AD data through LDAP, making it a common initial tool for reconnaissance when a threat actor first compromises an identity. To capture threat actors early in the reconnaissance phase, we developed two machine learning frameworks to detect LDAP reconnaissance: an ML classifier to predict malicious LDAP queries and an ML-based data-mining method to extract malicious query signatures. By correlating LDAP queries with endpoint detections, the first framework uses weak supervision to label a massive dataset and classify LDAP queries as malicious or benign. For immediate deployment, a second technique was developed on top of this approach to employ a rigorous statistical hypothesis-testing framework for mining novel, malicious LDAP signatures. While this weakly supervised approach is limited compared with manual human labeling, it is more practical for this use case because it leverages large-scale automated corpus construction, reducing costs and time. Ultimately, both the LDAP classifier and the ML-based LDAP signature mining method achieved performance benchmarks, with the classifier achieving up to a 65\% True Positive Rate (TPR) on the holdout set while limiting false positives, and mined signatures demonstrating 81.48\% field precision with CrowdStrike's Managed Detection and Response team.
arXiv (fysik & CS) Preprint 2026-06-30
DLR: Zero-Inference-Cost Latent Residuals for Low-Rank Pre-Training
arXiv:2606.28932v1 Announce Type: new Abstract: Large language models have driven recent progress in language and multimodal AI, yet pre-training them at scale is prohibitively expensive. Low-rank pre-training, which factorizes each weight matrix into a rank-r product to reduce both parameters and FLOPs, is a promising response but typically lags full-rank training in quality. We propose Duplicated Latent Residual (DLR), a training-only, parameter-free, foldable plug-in for low-rank pre-training. DLR augments the standard low-rank output Bz with a fixed structured residual alpha/sqrt(K) * Expand_K(z) that replicates each latent coordinate K = ceil(d_out/r) times across the output. With alpha fixed, DLR adds zero learnable parameters per layer; after training, it is absorbed into the up-projection in closed form, B* = B + alpha/sqrt(K) R^T, so deployment parameter count, FLOPs and memory match the underlying low-rank backbone exactly. Across LLaMA models from 60M to 7B parameters, DLR strengthens low-rank pre-training on C4 validation perplexity in most settings, with the clearest gains at 130M and above; folded checkpoints transfer cleanly to supervised fine-tuning on standard benchmarks.
arXiv (fysik & CS) Preprint 2026-06-30
EVLA: An Electro-Aware Multimodal Assistant for Physically-Grounded Driving Reasoning and Control
arXiv:2606.28938v1 Announce Type: new Abstract: Modern vision-language models (VLMs) for driving assistants typically treat vehicle dynamics as a black box, resulting in decisions that lack awareness of the vehicle's real-time electro-mechanical state. To bridge this gap, we introduce the Electro-Visual-Language Assistant (EVLA) -- a novel framework that combines multi-modal scene understanding with real-time perception of the electrified powertrain state (e.g., motor torque, battery SOC). Our approach features two key innovations: first, a Unified Co-State Encoder (UCSE) that fuses visual, textual, and vehicle-state inputs into a shared latent representation, augmented with an Energy-Efficiency Field to model spatial energy costs; and second, an Electro-aware Structured Reasoning Chain (ESRC), which replaces external chain-of-thought prompting with an internal, deterministic reasoning process grounded in physical constraints and optimization objectives. Trained end-to-end with a physics-guided joint loss, EVLA learns to generate context-aware and energy-optimal driving decisions. Extensive evaluations on a driving QA benchmark demonstrate that EVLA substantially outperforms strong fine-tuned VLM baselines, improving the final score by +0.0871 and accuracy by +5.6\%. Ablation studies validate the necessity of each component, and efficiency analyses show that EVLA achieves 36\% faster inference than multi-stage pipelines. This work underscores that integrating vehicle-state awareness and structured physical reasoning is crucial for developing next-generation, physically-grounded driving assistants.
arXiv (fysik & CS) Preprint 2026-06-30
ReGuide: From Test-Time Guidance to Self-Improving Diffusion Policies
arXiv:2606.28939v1 Announce Type: new Abstract: Behavior-cloned diffusion policies are expressive but remain vulnerable to covariate shift: small deviations from demonstrated states can compound into task failure. Existing methods address this either by expanding the training distribution through expert corrections or synthetic augmentation, or by steering a frozen policy at test time with guidance from a learned model. The former can be expensive or assumption-dependent, while the latter discards the corrected trajectories after execution. We introduce ReGuide, a self-improving framework that treats guided rollouts as reusable on-policy recovery data. ReGuide first uses Phase-Conditioned Guidance (PCG) to generate corrective rollouts: it constructs phase-specific latent targets, applies guidance only in the drifted-but-recoverable regime, and guides through the estimated clean action to match the dynamics model's training distribution. Successful guided rollouts are then absorbed back into the policy through ReGuide-FT, which fine-tunes the current checkpoint, or ReGuide-FS, which retrains from scratch on the augmented dataset; the two can also be composed and iterated. On Robomimic Can, Square, Transport, and Tool Hang, ReGuide improves base-policy success by $1.3$--$7.7\times$, outperforms LPB in the test-time-only setting, and matched-data ablations show that the gains come from guided recovery data rather than additional rollouts alone.
arXiv (fysik & CS) Preprint 2026-06-30
Modeling and Analysis of Sensing Assisted UAV Networks for Urban Vehicular Communications
arXiv:2606.28940v1 Announce Type: new Abstract: Urban vehicular networks (VNs) demand seamless connectivity and situational awareness within road-constrained environments, motivating the deployment of unmanned aerial vehicles (UAVs) platforms capable of simultaneously sensing vehicles and establishing communication with them. In this paper, we present a sensing-assisted UAV network that provides connectivity to the vehicles in an urban area. The road network of the urban area is modeled as Manhattan Poisson line process (MPLP), and the random location of vehicles on each road is modeled as one dimensional Poisson point processes (PPPs). UAVs are distributed in the urban area at a fixed altitude and provide connectivity after sensing the vehicles. Their locations are modeled as a two-dimensional homogeneous PPP. Combined with the fixed altitude, this results in a three-dimensional spatial configuration. We incorporate an elevation dependent blockage model and define the sensing radius based on detection probability (DP), showing that it is jointly limited by signal strength and blockage effects. We derive the DP and characterize the typical UAV's sensing region within the reliability requirements. We also derive the Laplace transform (LT) of aggregate interference accounting for directional patterns and sensing-driven activity, and analyze the resulting coverage probability (CP). Finally, we obtain the rate coverage (RC) of sensed vehicles falling within the UAV's sensing zone. Numerical results shows that increasing altitude degrades sensing and coverage performance, whereas RC exhibits a non-monotonic trend, first decreasing and then increasing with altitude.
arXiv (fysik & CS) Preprint 2026-06-30
Clustering Unsupervised Representations as Defense against Poisoning Attacks on Speech Commands Classification System
arXiv:2606.28953v1 Announce Type: new Abstract: Poisoning attacks entail attackers intentionally tampering with training data. In this paper, we consider a dirty-label poisoning attack scenario on a speech commands classification system. The threat model assumes that certain utterances from one of the classes (source class) are poisoned by superimposing a trigger on it, and its label is changed to another class selected by the attacker (target class). We propose a filtering defense against such an attack. First, we use DIstillation with NO labels (DINO) to learn unsupervised representations for all the training examples. Next, we use K-means and LDA to cluster these representations. Finally, we keep the utterances with the most repeated label in their cluster for training and discard the rest. For a 10% poisoned source class, we demonstrate a drop in attack success rate from 99.75% to 0.25%. We test our defense against a variety of threat models, including different target and source classes, as well as trigger variations.
arXiv (fysik & CS) Preprint 2026-06-30
FlipGuard: Defending Large Language Models Against Quantization-Conditioned Backdoor Attacks
arXiv:2606.28962v1 Announce Type: new Abstract: Model quantization is essential for the efficient deployment of Large Language Models (LLMs), but introduces a critical vulnerability: Quantization-Conditioned Backdoor (QCB) attacks. In these attacks, malicious behaviors remain dormant in full-precision models and activate only after specific quantization distortions, bypassing standard security audits. To mitigate this, we introduce FlipGuard, a proactive defense framework that selectively perturbs model weights prior to quantization. By breaking the adversary's precise alignment between weight patterns and quantization boundaries, FlipGuard suppresses backdoor activation without requiring access to training data or trigger samples. We further propose the Defense Effectiveness Ratio (DER), a unified metric to jointly evaluate security gains, utility preservation, and computational cost. Extensive experiments across seven LLMs (including StarCoder and LLaMA-family models) and three quantization schemes (INT8, FP4, NF4) demonstrate that FlipGuard effectively neutralizes QCBs across three scenarios, i.e., vulnerable code generation, content injection, and over-refusal, achieving high security with negligible performance degradation.