Papers

We propose a fast deterministic particle method for the spatially homogeneous Landau equation. The method combines a particle representation of the solution with a Fourier approximation of the nonlinear collision flux. Nonuniform fast Fourier transforms are used to reconstruct the density from the particles and to evaluate the flux and density at the particle locations, while the convolutional structure of the flux enables its efficient computation by FFTs. For a fixed transform tolerance, the cost per time step is $\mathcal{O}(N+M^d\log M)$, where $N$ is the number of particles and $M$ the number of Fourier modes per velocity dimension. We establish a consistency estimate for the reconstructed velocity field on regions where the reference density is bounded away from zero. The estimate separates the spectral truncation error from the particle-density reconstruction error, showing how the latter can dominate for sufficiently smooth densities. Two-dimensional numerical experiments with Maxwellian interactions illustrate the accuracy and efficiency of the method, its sensitivity to particle and spectral resolution, and the effects of filtering. Comparisons with a direct blob implementation demonstrate improved accuracy at a strongly reduced computational cost.

Large language models (LLMs) and agentic AI systems are creating rapidly growing inference energy demands as model sizes grow and reasoning trajectories extend. While in practice, many queries do not require the capabilities of the largest available model, and routinely directing such queries to a high-capability model can introduce unnecessary, considerable computation and energy consumption. In this paper, we investigate whether adaptive routing across a heterogeneous pool of LLMs can reduce this energy burden without substantially compromising task performance. We design a language-model-based router that reads in each query and selects an answer model from a fixed candidate pool. The candidate models are first profiled through an offline tournament that records their correctness, latency, power, and GPU energy for each query. Using these measurements, the router is trained through supervised fine-tuning followed by group relative policy optimization (GRPO) with the tailored paradigms. Results demonstrate that learned routing can selectively allocate expensive model capacity based on query context and improve the accuracy-energy tradeoff in multi-LLM serving. Across seven benchmark tasks, we also observe a sharp accuracy-energy phase transition among routers, providing practical insights into improving energy efficiency while maintaining LLM performance.

Over-the-air federated learning lets edge devices transmit their local updates simultaneously, reducing the communication overhead. The resulting waveform, however, has a peak-to-average power ratio (PAPR) that grows with the model dimension, and keeping the amplifier in its linear range leaves two remedies: clipping the peaks or backing off the transmit power. Neither remedy is without cost: i) the clipping distortion appears at the receiver as a bias that cannot be removed, and ii) back-off keeps the signal intact but degrades the average signal-to-noise ratio (SNR). Independent of this trade-off, the transmission remains uncompressed, spending one channel use per model parameter, which keeps large-model training out of reach. To address these challenges, we propose AirGC-CD, an over-the-air scheme that precodes each local update with a partial Gaussian circulant matrix before clipping. In AirGC-CD, the precoder's output is exactly Gaussian regardless of the update's sparsity, so the clipping function is designed for a known distribution instead of inheriting it from the data. This enables the clipping to be inverted on average by a single scalar Bussgang gain in closed form, and we prove that the resulting aggregate is exactly unbiased, with clipping adding only variance. The clipping ratio is then the only free parameter left, trading the variance of the clipping against the SNR loss from back-off, and we derive its near-optimum in closed form. Since the precoder is linear, it also acts as a compressor, reducing the transmission from the model dimension d to the sketch dimension m at a cost of only O(dlog d) via two fast Fourier transforms, whereas a Gaussian sketch costs O(md). Experiments on five image datasets show that AirGC-CD outperforms baseline over-the-air FL schemes in most settings, particularly at low SNR, while using fewer channel uses per round.

The distinction between the spirit and letter of the law is a central issue across research and everyday life, and a growing concern for building safe, intelligent machines. What is this distinction based on, and how can we develop machines that follow the intention behind a rule? We used targeted adaptation that made large language models prioritize the spirit or letter of the law. With minimal modifications, our method significantly changed LLM behavior across diverse measures, novel vignettes, real-world scenarios, and influential legal cases. An analysis of model internals revealed a low-dimensional space with three interpretable dimensions matching a formal pre-specified framework for the geometry of legal concepts. These findings show how legal thought in LLMs may be organized and directed.

Enterprise integration traffic in large Enterprise Resource Planning (ERP) systems move through Electronic Data Interchange (EDI) and the Intermediate Document (IDoc), and the operational tools that surround that traffic remain fragmented. Status records are written for a technical audience, exception handling is performed one document at a time, and process health is read manually from long lists. This work proposes an enterprise process control tower that lets an Information Technology (IT) team observe, diagnose, and improve business workflows from a single operational layer. The design links process health observation, semantic translation of technical status text, machine-learning diagnosis of exceptions with model explainability, grouped recovery actions, and key performance indicators inside one governance framework. The proposed framework is evaluated on a synthetic IDoc event dataset of 84,000 records spanning twelve months, of which 15,120 error records are used to train a five-class remediation-routing model that maps each exception to a standard recovery action. A rule-based router that reproduces current production heuristics serves as the primary baseline. The proposed gradient-boosted diagnosis engine reaches a macro-averaged F1 score of 0.883 on the held-out test partition, against 0.742 for the rule-based router, while Shapely-based attributions expose the drivers of each routing decision to operators. Results are derived from synthetic data and are presented as an internal-consistency demonstration rather than real-world validation. The contribution is an integrated tool design that places diagnosis and governance on the same plane as observation.

Stochastic partial differential equations (PDEs) govern critical engineering and geophysical systems but are challenging to use for real-time control under parametric uncertainty. We present a unified framework that couples Physics-Informed Neural Networks (PINNs) with Polynomial Chaos Expansion (PCE) to construct a fast and differentiable surrogate. The PCE representation enables analytical propagation of parametric uncertainty and computation of the corresponding moments without requiring Monte Carlo sampling. We provide an error decomposition for the PINN-PCE surrogate that separates PCE truncation, stochastic quadrature, and PINN approximation errors. Embedding this surrogate into a stochastic model predictive control (SMPC) scheme enables finite-horizon control updates based on analytic mean and covariance predictions. We further show how the surrogate approximation error can be incorporated into tightened probabilistic constraints. The approach is validated on three benchmarks: the Korteweg-de Vries equation, Burgers' equation, and the two-dimensional incompressible Navier-Stokes equations, representing dispersive, convective, and convective-diffusive dynamics. Across all cases, the surrogate enables real-time control updates while maintaining prescribed risk levels and closely matching the corresponding high-fidelity solvers at substantially lower computational cost.

Forecasters often know an event is imminent but not the shape, size, or timing of its effect. We introduce Event Signature Transfer (EST), a training-free, model-agnostic operator that turns a completed past event into an explicit forecast scenario. EST removes a source event's own trend and seasonality, then scales and retimes the remaining event signature onto a native forecast, preserving the forecast's linked structure and reducing to it exactly at zero strength. Because it reads only output quantiles, EST applies to any quantile forecaster, with no training, no model internals, at transfer time. Across twelve real episodes and ten synthetic scenarios on Chronos-2, TimesFM-2.5 and Toto-2.0, manually configured EST reduces real-episode WQL by 21.7-90% in-sample. On Chronos-2, it leads eleven of twelve matched comparisons against covariate conditioning, activation editing and raw replay. The operator builds a scenario; it does not estimate its likelihood.

Recent advances in natural language processing have led to molecular Large Language Models (LLMs) with strong performance across diverse chemistry tasks. However, they still struggle to capture fine-grained structure-property relationships, particularly how small, localized modifications alter a molecule's behavior. To address this limitation, we introduce MolSC, a dataset of substituent contributions, defined as property changes induced by attaching specific substituents to molecular scaffolds. Curated from manually annotated bioactivity records, MolSC spans structural-alert liability, target-specific bioactivity, and physicochemical descriptors, and contains 181K substituent-level examples for training. We further propose MolSC-Bench, a held-out evaluation benchmark of 1,541 examples disjoint from MolSC at the scaffold, substituent, and molecule levels. Our experiments show that existing molecular LLMs and strong proprietary models such as GPT-5.2 and Gemini-3-Flash show limited reliability in substituent contribution prediction. In contrast, training on MolSC substantially improves this ability and achieves strong performance across diverse downstream molecular tasks. These results highlight substituent contribution learning as a key component of fine-grained molecular understanding.

This thesis proposes a framework for "responsible intelligence" to address AI's critical challenges in safety, ethics, and cultural sensitivity. It advances three core areas: First, it improves domain adaptation in specialized fields using active learning and graph-based knowledge to reduce hallucinations. Second, it enhances ethical rigor via a novel decoding-time alignment mechanism that proactively blocks harmful text generation in real-time. Finally, it ensures cultural and multilingual safety through language-specific steering that respects diverse linguistic and social norms. Ultimately, this work provides a blueprint for building next-generation AI that is contextually knowledgeable, ethically sound, and culturally adaptable.

Referring surgical video instrument segmentation (RSVIS) aims at segmenting the instrument in a surgical video, given a textual description. Despite recent progress, current models are trained and assessed on relatively small-scale benchmarks, hindering the development of more general RSVIS. In addition, existing benchmarks support only the single-target expression that refers to one instrument in the video, while overlooking multi-target and no-target referring expressions, restricting the applicability of RSVIS in practical scenarios. Addressing these issues, we propose LD-RSVIS, a new benchmark aiming to facilitate more robust and general RSVIS. Specifically, LD-RSVIS consists of 3,536 surgical videos with 1.09 million frames and covers a broad set of 30 instrument classes from 25 various procedures. By including abundant videos and classes, LD-RSVIS could benefit large-scale training and evaluation of more general RSVIS methods. Besides, unlike existing datasets, LD-RSVIS offers diverse referring settings, including no-target, single-target, and multi-target expressions, which enables the development of more practical RSVIS models in real applications. In order to ensure high-quality annotations, all videos in LD-RSVIS are manually labeled with multiple rounds of inspection and refinement. To our knowledge, LD-RSVIS is the largest and most diverse benchmark for RSVIS. To analyze LD-RSVIS and to provide comparison for future research, we evaluate 12 representative methods, and the results reveal that more efforts are required for improvements. To encourage future research, we present a simple yet effective RSVIS method, dubbed Cascade-RSVIS, that first mines target-specific cues using the complementary multi-cue text information and then employs such cues and textual information for segmentation, achieving promising performance. Our benchmark and code will be released.

Chain-of-thought (CoT) can sound plausible yet be unfaithful to the model's underlying reasoning. Most prior work probes CoT faithfulness through input–output behavior or input attributions, leaving internal computation largely underexplored. We instead cast faithfulness as internal concept grounding: Does a large language model's (LLM) CoT reasoning engage the same internal concepts that support the LLM's direct prediction, and do the shared concepts causally drive its answer? Encoding a prediction pass and a CoT pass with a single shared sparse autoencoder (SAE), a reliable approximator of the latent concepts LLMs use, makes their internal concepts directly comparable. We introduce three correlational metrics of concept-level alignment and a causal metric, $\Delta p$, which ablates the shared concepts and measures the drop in answer probability. Across five LLMs and four datasets, concept alignment is generally high, as indicated by the correlational metrics; yet these only identify which concepts are shared, not how much they causally contribute. $\Delta p$ fills this gap: causal faithfulness varies substantially with model depth, peaking at mid-to-late layers rather than the final ones, and model scale reshapes the layer-wise profile. Moreover, causally important shared concepts are not always verbalized in the CoT. These dissociations suggest that faithfulness cannot be reliably assessed from surface-level or representational correspondence alone; assessing it requires causal tests of whether the internal concepts underlying a CoT actually drive the model's prediction.

Understanding the physical world requires more than object recognition, scene description, and short-term visual prediction, as real-world physical systems involve multiple continuous fields, latent causal mechanisms, partial observations, and intervention-sensitive dynamics. We propose FireWorldBench, a benchmark for evaluating complex physical world intelligence in multimodal large language models and agents through coupled-field fire dynamics. Fire provides a canonical stress-test environment, where multiple interacting physical fields jointly shape observable states and temporal dynamics. FireWorldBench is organized along two complementary axes, a physical capability axis and a fire scenario task axis, jointly covering physical-state understanding, temporal dynamics, causal mechanisms, and intervention reasoning. The benchmark comprises 520 fire-world entries, including 494 controlled simulation worlds and 26 real-world-aligned event groups, spanning 47 scene archetypes across 7 environment families. These entries combine structured textual observations, multiple 2D physical-field visualizations, and 3D event-level scene modeling, yielding 9,074 text-image interleaved question-answer pairs across choice-based and open-ended report-generation formats. FireWorldBench evaluates whether models can infer latent physical states, explain underlying mechanisms, forecast coupled-field evolution, and assess intervention consequences from multimodal partial observations, providing a challenging testbed for complex physical world intelligence.

Small-object detection in UAV imagery is challenged by weak visual evidence, ambiguous boundaries, dense object distributions, and complex backgrounds. Effective detection therefore requires long-range contextual information for target-background discrimination while preserving explicit local two-dimensional structures for accurate localization. These requirements arise at different stages of the detection pipeline and are not naturally addressed by a uniform feature-processing strategy. We propose Hybrid Dual-domain Mamba-YOLO (HDMamba-YOLO), a stage-wise heterogeneous SSM-CNN detector organized according to a perception-reconstruction-alignment-interaction rationale. EfficientVMamba-based EVSS establishes long-range contextual perception in the backbone, while PhasePatchMerging2D provides phase-aware hierarchical transitions. DST-Wrapper and Native C3k2-ASSAF then perform perception-to-reconstruction transition and repeated local two-dimensional reconstruction during FPN/PAN aggregation. DySample provides content-adaptive cross-scale resampling, while OS-CVTIA introduces macro-micro interaction and task-specific modulation for localization and classification. On VisDrone2019, HDMamba-YOLO-B achieves 42.737% mAP50 and 25.713% mAP50:95 with 10.042M parameters and 29.879 corrected GFLOPs. HDMamba-YOLO-Lite achieves 41.140% mAP50 and 24.741% mAP50:95 with 5.344M parameters. Under the unified AI-TOD evaluation protocol, HDMamba-YOLO-B obtains 21.621% AP and 47.881% AP50. Controlled ablations further support the stage-wise allocation of state-space perception, convolutional reconstruction, dynamic alignment, and task interaction for UAV small-object detection.

We study the online deployment of mobile ad hoc networks in unknown orthogonal environments, formalized as the Partially Observable Cooperative Guard Art Gallery Problem. We give a full proof that CADENCE algorithms achieve full coverage while maintaining a connected visibility graph using at most $n/2 + h - 2$ agents in orthogonal worlds with $n$ corners and $h$ holes. We further evaluate deployment-order heuristics that reduce agent count and deployment time in practice.

Agent runtimes that plan before acting run every step as soon as it is ready. Neither these runtimes nor their benchmarks distinguish a step that is ready from a step the request needs, and the difference is large: across 303 requests on 154 production workflows, an eager runtime that stops once the goal is met spends 62.2% of its CPU on work the request never needed. LazyAgent is, to our knowledge, the first complete formulation of demand-driven execution for agent runtimes. Its rule: run a ready step only if the current request needs it, and keep that answer exact as the plan changes. A demanded set derived from the program decides, in one traversal with no model calls, so LazyAgent spends nothing outside the request. Because demand is defined exactly, its behavior is proved: builders know before any run where demand must save and where it must tie. The bill is known before a plan runs, so LazyAgent can act as a query optimizer to choose among candidate workflows. Measured live against that eager runtime, LazyAgent saves 51.7% and 52.1% of container time on release gates where a frontier LLM writes the patches, 12.8% of tokens on DataSciBench plans at matched quality, and 24.0% of dollars inside a live LangGraph application, with no losses. When the LLM re-plans after real code failures, it solves every solvable goal with 11.7% fewer tokens per goal. On the production workflows it saves 42.0% of CPU, ties exactly where a request needs the whole graph, and adds at most 0.43% overhead. The same set steers every physical operation: projection saves 80-89% of a shared CI step, where the identical eager control saves 0.0%. Public agent benchmarks are designed the same way as eager runtimes, so they cannot show this difference. We compose our test programs from their tasks and call on industry to build benchmarks that measure how efficiently agent workflows run.

Agentic retrieval-augmented generation (RAG) enables language models to adapt retrieval based on previously retrieved evidence, but it remains unclear whether such adaptive orchestration consistently outperforms well-designed static pipelines. We conduct a controlled comparison of agentic and static RAG for Taiwanese historical question answering, sharing the same generator and hybrid retrieval backend. Despite similar aggregate performance, the two pipelines differ on 70.83% of questions, with their advantages largely canceling out when averaged. An oracle that selects the better response per question improves the composite score by 0.2417 over the better individual pipeline, revealing substantial headroom for question-level selection. We therefore introduce a post-hoc selector that compares the two responses and their cited evidence, significantly outperforming either individual pipeline and recovering 60.34% of the oracle headroom. These results show that aggregate comparisons can obscure meaningful question-level differences between retrieval strategies, suggesting that exploiting their complementarity may be more fruitful than seeking a universally superior pipeline.

Discrete diffusion models now match autoregressive language models on several benchmarks, while the question of how best to train them has received far less attention: the optimizer is inherited from one paper to the next and never compared. New optimizers, meanwhile, are validated almost exclusively on autoregressive pretraining, a different objective on a different loss surface. We present a controlled optimizer benchmark across four diffusion formulations, to our knowledge the first for discrete diffusion: seven optimizers (AdamW, Lion, Muon, SOAP, MARS, MARS-M, Schedule-Free) on masked diffusion (text8), uniform diffusion (QM9, and LM1B through the Gaussian duality) and Gaussian diffusion on images (CelebA-64), each on a task with published reference values. Every optimizer receives the same search protocol, and every winner is retrained at the full budget with three seeds. AdamW is a strong default but not always the right choice: it is beaten by a resolved margin on two of the four tasks, and the winner changes with the formulation, so the optimizer deserves the same care as the rest of the training recipe. Notably, methods validated on autoregressive language model pretraining transfer well: Muon, MARS-M and SOAP each beat the tuned AdamW on at least one diffusion formulation. The benchmark, all runs and every figure are reproducible end to end from the released code at https://github.com/armanbolatov/diffusion-baselines.

The Time-Evolving Natural Gradient (TENG) method evolves neural approximations of time-dependent partial differential equations by projecting time-discrete target states onto the neural-network manifold through local natural-gradient iterations. For stiff diffusion-dominated problems, however, explicit Runge–Kutta methods face severe stability restrictions. High-frequency diffusive components can drive stage targets beyond the locally reachable region of the neural manifold, worsening the conditioning of local least-squares projections and destabilizing parameter updates. We propose TASE–TENG, which applies the highly stable explicit operator $T_p(hW_0)$ to each Runge–Kutta stage increment before projection, where $W_0$ approximates the dominant stiff diffusion operator. To construct this surrogate in a mesh-free setting, we develop a learnable operator whose local interaction coefficients are generated by a shared geometry-dependent neural rule. A structured factorization enforces discrete self-adjointness, dissipativity, and constant preservation. The operator is trained offline and remains fixed during online integration, where it is used only within the TASE stabilization operator. Numerical experiments on stiff diffusion and reaction–diffusion problems demonstrate improved stability and accuracy at the same time-step size, with reduced errors from high-frequency diffusive modes. The learned rule also allows the stabilization operator to be reconstructed at different spatial point resolutions without retraining.

A RAG system can hand you the right answer and cite a source it did not actually use. Models output these unfaithful citations via post-rationalization: they write the answer first and then attach a citation to whatever passage looks close enough. Search agents are now trained with reinforcement learning from verifiable rewards (RLVR), which pays them for getting the answer right. We asked whether that training also teaches them to cite honestly. Improving an existing methodology with a required control, we compared an instruction-tuned model against three RLVR agents trained from it, on four question-answering datasets, using only free-tier Kaggle GPUs. Post-rationalization is everywhere: on Wikipedia-based questions roughly one citation in seven is unfaithful. RLVR does not fix it. The agents post-rationalize at their base model's rate, and one lands slightly worse. Rewarding correct answers buys nothing in citation faithfulness, so faithfulness has to be trained and measured on its own terms.

Large-scale scene reconstruction is a critical foundational technology in robotic autonomous systems such as 3D mapping and autonomous driving. In recent years, 3D Gaussian Splatting (3DGS) has demonstrated remarkable advantages in both visual quality and computational efficiency, making it a promising representation for large-scale scene reconstruction. However, it still faces challenges in large-scale scenes, including excessive memory consumption and uneven viewpoint coverage caused by UAV acquisition, limiting its real-world applications. To address this, we propose VDGS, a novel 3DGS framework that incorporates camera distribution into scene modeling. VDGS introduces visibility-driven statistics for scene anchors to quantify supervision strength. These statistics are further leveraged for scene partitioning and for gradient compensation in under-optimized regions, thereby promoting balanced optimization across different regions. Extensive experiments on multiple large-scale aerial scene datasets demonstrate that, under imbalanced viewpoint distributions, VDGS consistently outperforms existing methods, while maintaining competitive performance in scenarios with more uniform view distributions.

Compressed manipulation policies can pass offline evaluation while failing in closed-loop execution; this dissociation is established in prior work and is not our claim. We contribute a causal anatomy of one naturally occurring case. An 8-layer distillation of Octo-Base retains 86% of parameters, passes every offline check we applied (0.996 and 1.000 teacher-ratios on the family's own validation metrics), and collapses in closed loop: 0/72 vs. the teacher's 40/72 on a simulated WidowX pick-and-place task. The collapse is structured, not diffuse: early task stages degrade gradually (the student moves the object at 90% of the teacher's rate and grasps at 55%), while transport-to-target fails categorically, at 0% in every training variant. Paired action-trace forensics isolate the signature: a negative, late-heavy $z$ residual, roughly 10x its post-repair magnitude, and persistent across the base distillation and both continuation branches. Four standard therapies fail under matched controls: continued training and in-domain offline data leave success at zero, even though the latter measurably improves marginal action statistics; command-level compensation recovers nothing at any offset, although the same perturbations degrade healthy policies; clamping the symptom in the command channel preserves grasping, yet success stays at floor. A minimal-pair intervention that substitutes half of the training stream with deployment-distribution teacher rollouts, with every other setting held fixed, restores parity with the teacher (18/36 vs. 17/36 held-out), eliminates that signature, and recovers a teacher-like perturbation-response profile. We claim existence, not universality. Operationally, offline gates, including a family's own validation metrics, are insufficient acceptance tests for compressed policies; a few dozen closed-loop trials sufficed to find what they missed.

We establish the asymptotic-preserving property of the exponential Euler method for stochastic reaction-diffusion-advection equations in $\mathbb R^2$ under fast Hamiltonian advection with multiple critical points. In the fast-advection limit, the dynamics reduces to a stochastic partial differential equation on a noncompact metric graph. A key ingredient is the strong convergence analysis of the limiting exponential Euler scheme, which is complicated by the graph's noncompactness, its multiple edges and vertices, and the nonuniform ellipticity and vertex degeneracy of the graph operator. Using a weighted $L^2$ framework and analytic-semigroup smoothing estimates, we prove a temporal convergence order arbitrarily close to $1/2$ for weighted $L^2$ initial data and exactly $1/2$ under a half-order fractional-domain condition. Combining these estimates with the fast-advection limits yields the asymptotic-preserving property in the multiple-critical-point setting. Numerical experiments illustrate the asymptotic behavior and confirm the predicted temporal convergence rate.

Large Language Models (LLMs) enable open-ended dialogue in interactive games, but their non-deterministic outputs make it difficult to preserve authorial control, factual consistency, and the intended sequence of information disclosure. These challenges are particularly significant in detective games, where premature revelation or fabricated details can undermine the logic of player progression. We present a Structured Knowledge Tree architecture coupled with a tri-agent LLM pipeline for controlling dialogue in an open-ended interrogation game. The system separates knowledge retrieval, dialogue generation, and response verification to ensure that the virtual suspect reveals only information permitted by the current narrative state. We evaluate the approach through The Interrogation of Adrian Gale, a playable detective-game testbed, and a formal user study examining hallucination reduction, adherence to authored disclosure sequences, and perceived logical progression. Our results demonstrate that the structured architecture reduces critical hallucinations by 64.78% and entirely prevents premature narrative disclosure. While the strict mechanical constraints introduced usability trade-offs regarding forced conversational reveals, the system successfully enforces rigorous epistemic pacing and provides players with a clear, subjective sense of progression toward solving the case.

Pre-deployment secret scanning operates only on source code, never on what a production application serves. We document two exploitation chains in which Azure AD client credentials and APIM subscription keys from production JavaScript bundles enabled account takeover and mass data exposure. An authorized engagement covered approximately 2,000 enterprise web assets in one organization; 113 (5.65%) served live credentials. To quantify the shift-right gap, we built an independent Ground Truth (GT-194) of 194 secret-grade credentials through Claude Opus 4.7 extraction and manual analyst review, with the 247 LLM-extracted candidates independently validated by GPT-5.5 (Brennan-Prediger kappa = 0.676). The principal finding is structural: 13.9% of GT-194 (27 of 194) is surfaced only by manual analysis and recovered by none of the nine evaluated production scanners, a tool-agnostic blind spot the ground-truth model also misses. CryptoJS encrypted configuration separately defeats every static scanner: the credential exists only after decryption with a co-located key, reached only by runtime-aware detection. Combined coverage plateaus at 86.1%. Among the nine scanners, the best static scanner recovers 36.6% and the best runtime-aware scanner 77.8% (F1 = 0.818, McNemar p < 0.001); the ground-truth model is reported separately as a reference comparator, not an evaluated detector. On 63 of 86 secret-exposed applications (73.3%), the full Azure AD token-mint chain is co-located in one bundle, reachable from browser code. We characterize five paths by which credentials reach production undetected and present a layered runtime detection methodology and remediation framework. Recall is scoped to a single-organization Azure-heavy corpus.

We consider a nonlinear model motivated by polychromatic computed tomography (CT). Here, $W$ distinct $d$-dimensional signals $x^*_1, \ldots, x^*_W \in \mathbb{R}^d$ must be recovered from $n$ measurements $(a_i, y_i)_{i = 1}^n$ that obey the nonlinear model $\mathbb{E}[y_i|a_i] = h(\langle a_i, x^*_1 \rangle, \ldots, \langle a_i, x^*_W \rangle)$, where $h: \mathbb{R}^W \to \mathbb{R}$ is a known nonlinearity that models a certain type of exponential attenuation law. Even when there is no noise in the measurements, the sample size $n$ (for any measurement ensemble $\{a_i\}_{i = 1}^n$) must exceed the number of unknowns $Wd$ to guarantee that the underlying signals are identifiable. We construct a measurement ensemble that ensures perfect signal recovery almost surely provided $n \geq Wd + W - 1$, thereby isolating the injectivity threshold up to an additive factor $W - 1$. We also study computational complexity of signal recovery in this model with a general measurement ensemble $\{a_i\}$. We show that if $W \geq 2$, then there is a measurement ensemble for which deciding whether there exist signals consistent with the measurements is NP-hard. In particular, this implies that polychromatic, multimaterial CT reconstruction is NP-hard in general. This finding stands in sharp contrast to the single-material setting, for which a polynomial-time algorithm can provably perform signal recovery for any measurement ensemble.

LLM-based AI systems answer political questions for hundreds of millions of people. Current audits measure what they say to an average user, but their behavior is dynamic. I argue that their political behavior is a set of policies over whom to answer, what to say, and whether to engage at all, conditional on the topic and what the system knows about the user. I call these policies the system's speech regime, which is how a developer settles the tradeoff between answering, accommodating the user, and refusing, each of which carries a cost that varies by topic. I derive a typology of five regimes from two dimensions, engagement and stance. I test six AI systems (OpenAI, Anthropic, xAI, Google, Mistral, DeepSeek) in a preregistered experiment of 7,500 multi-turn conversations that randomly assign the user's political identity across five topics: abortion, Catalan independence, climate change, Nazism, and a zero-stakes control (pineapple on pizza). Two LLM judges from different developers score every answer, validated against human coding, and refusal is treated as an outcome rather than missing data. Every system accommodates the user on the control topic, showing that political restraint is a policy. On contested topics the systems fall into different regimes: on abortion, GPT engages and mirrors every user, Gemma refuses everyone, Claude answers strongly conservative users 35 percent of the time and almost no one else, and Grok accommodates conservatives only. On settled topics such as climate change and Nazism, five systems hold firm for every user. The systems also infer the user's overall ideology, so accommodation can spill over to topics not yet discussed. A comparison of two Grok releases shows the regime changing between versions in a way current audits miss. Speech regimes matter for alignment research and for polarization, political knowledge, and the quality of democracy.

Spatial reasoning is essential for vision-language models (VLMs) to understand and act in the physical world. Reasoning in dynamic environments requires VLMs to perceive local state transitions caused by object motion and viewpoint changes and integrate them over long trajectories to maintain an updated spatial state, yet existing VLMs remain limited in both capabilities. Current spatial training primarily focuses on static questions about object attributes and spatial relations, providing limited direct supervision for state transitions; in contrast, interaction trajectories naturally connect a preceding observation, an action, and a subsequent observation, offering direct supervision for local state transitions, while complete trajectories reveal dependencies among consecutive transitions. We therefore introduce Spatial-Interactor, a framework that trains VLMs to model physical-world state transitions through interaction, organizing this learning process into a three-level curriculum covering L1 passive world-state transitions, L2 active self-state transitions, and L3 long-horizon interaction trajectories. Accordingly, we construct the Learning from Spatial Interaction dataset (LSI-108K) from simulated and real interaction trajectories, with tasks aligned with the objective of each level. Our two-stage training strategy applies Supervised Fine-Tuning (SFT) to L1 and L2 for local transition modeling, and On-Policy Distillation (OPD) then uses privileged self-distillation: a teacher branch given segment-level transition descriptions supervises the student's on-policy CoT, helping the student learn to integrate consecutive transitions over L3 long trajectories. Experiments across multiple VLMs and spatial benchmarks show consistent gains in local transition modeling and long-horizon integration.

We present a fast numerical method for safe continuous-time motion planning under Temporal Logic (TL) specifications. The method generates smooth continuous trajectories that remain collision-free while robustly satisfying temporal and logical task requirements. A central component of our method is the formulation of nonconvex safety and logic constraints as unions of convex sets where associated discrete decisions are encoded in a joint feasibility graph. This graph representation allows Euclidean projection onto the feasible set and proximal robustness maximization to be reformulated as shortest- and widest-path problems, respectively. Building on this structure, we develop a nonconvex splitting method based on the Alternating Direction Method of Multipliers (ADMM), which decouples smooth spatio-temporal trajectory optimization from nonsmooth discrete constraint handling within the optimization. The resulting algorithm exhibits reliable convergence across benchmarks and scales to large-scale motion-planning problems, providing a 4.7x average speedup over the state of the art on discrete and continuous-time logic problems.

Approximate nearest neighbor search (ANNS) underpins large-scale vector retrieval in search, recommendation, and retrieval-augmented generation. Graph-based indexes have demonstrated state-of-the-art search performance for ANNS. They connect each corpus vector to a small set of nearby or navigationally useful vertices and answer queries by traversing the resulting graph. Because these edges are selected using construction-time distances, the graph index is tied to the embedding model. Re-encoding a corpus with a new model may change distances and neighborhoods of the vectors. Reconstructing the graph for the new embedding vectors incurs substantial construction cost and delays deployment. When the embedding model changes, we observe a phenomenon in the old graph index that we call residual reachability. Specifically, although derived from different models, the vectors describe the same underlying objects and often retain part of their similarity structure. These shared relations are reflected in the connectivity of the old graph index, leaving many exact new-model neighbors reachable within a few hops in the old graph index. Motivated by this observation, we develop an index-migration approach that utilize the residual reachability in the old graph index to faster construct the new graph index for the new embedding vectors. Our method, Drift-Guided Migration (DGM), provides two migration paths. DGM-Local performs parallel shallow expansion over the inherited graph index and screens second-hop candidates with packed position sign codes before exact evaluation. DGM-Search uses hop-bounded beam traversal to explore beyond shallow expansion. Across eight text and image migrations, our DGM methods can achieve up to 17.43 times speedup on constructing the new graph index than the fastest degree-matched reconstruction method while keeping competitive recalls.

Looped language models repeatedly apply a weight-shared block to increase effective depth without increasing parameter count, but the resulting T sequential recurrent-block calls per generated token substantially increase decoding latency. To address the issue, we introduce Wavefront Decoding (WFD), a training-free self-speculative decoding framework designed for looped language models. WFD exploits two properties of these architectures: intermediate recurrence outputs provide effective draft predictions, and weight sharing allows token states at different positions and recurrence depths to be processed in one batched recurrent-block call. WFD organizes these mixed-depth states into a diagonal wavefront, continuously drafting new positions at shallow depth while advancing earlier positions toward full-depth verification. Unlike the phase-separated draft-then-verify schedule, WFD therefore concurrently batches drafting and verification within the same recurrent calls, while rejected drafts are corrected using full-depth predictions. Across six Spec-Bench task categories, WFD achieves 2.42x speedup on Ouro-2.6B and 3.54x on Huginn-3.5B over autoregressive decoding, consistently outperforming draft-then-verify. Cross-recurrence KV sharing further reduces wavefront KV traffic and increases WFD's speedup to 4.81x on Huginn-3.5B. The code is available at https://github.com/summerbro-hhj/wavefront-decoding.