Papers

After re-examining a classic article by Hugh Everett, I will discuss the entropy of de Sitter space from the standpoint of an observer living in the space. This provides a nice example of the importance – for certain questions – of taking seriously the fact that in cosmology any observer is really part of the universe that is being observed. (This article has appeared in a special issue of International Journal of Modern Physics A in honor of Gerard 't Hooft.

Counter-storytelling is a powerful mechanism people use to challenge dominant narratives. Unlike other forms of counterspeech that have been widely studied in computational social science, counter-storytelling has largely been overlooked. Counter-stories are difficult to detect automatically; they are relational (defined with respect to expressions of racial stereotypes) and structurally diverse (drawing on stories that describe lived experiences, witnessed events, exemplars, and hypotheticals). We introduce a first framework for detecting and characterizing counter-storytelling against racial stereotypes at scale. This includes (1) a three-dimensional taxonomy grounded in narratology and Critical Race Theory and (2) a multi-stage pipeline that identifies relational pairs of stereotypes and counter-stories in noisy Reddit discourse. Using this pipeline, we annotate 25,549 Reddit posts across 615 communities and identify 1,312 counter-stories. Our analysis shows that speaker identity and post context shape how counter-stories are told. For example, in-group writers favor first-person testimony, often adopting the role of self-reflective insiders. Our work shows how computational methods can scale qualitative approaches to identify and characterize counter-storytelling as a contextual narrative practice, with implications for content moderation, narratology, and racial discourse analysis.

This project explores how multiple cognitive radars can learn to share limited wireless spectrum with other radio users without interfering with one another. Using machine learning (ML), each device independently decides where and how widely to transmit within a fixed 100 MHz band. The system analyzes real or simulated signal activity to detect which parts of the spectrum are currently in use and which are open. Based on this information, the devices adapt their transmission choices to avoid crowded frequencies while making efficient use of available space. The goal is to develop a flexible, scalable approach to spectrum sharing that could support future wireless communication systems. Experimental results using both over-the-air software-defined radio (SDR) recordings and simulated environments demonstrate that the proposed meta-learning approach consistently balances competing objectives better than conventional reinforcement learning (RL) methods in multi-agent spectrum-sharing scenarios. Across five multi-agent benchmark environments, our proposed method achieved the highest average reward among the primary baseline algorithms while simultaneously maintaining low collision rates and stable transmission behavior.

A useful model of AI inference infrastructure must specify the system state, the information available to an observer, and the decisions the model is intended to support. We introduce AID (AI Infrastructure Dynamics), a framework for describing this learning problem across coupled physical, computational, networking, and serving processes. The formulation allows structured and variable-size state, asynchronous observations, multiple physical timescales, and demand that responds to service. We distinguish representations that support prediction under an existing policy from those that preserve service outcomes under changed actions, and separate both from identifying intervention responses. Two analytical results describe a lower bound on prediction error when available observations cannot distinguish models and a sufficient condition for exact controlled state reduction. These results apply established information and state-abstraction principles to AI infrastructure. We then describe a validation protocol for cache representations, workload histories, measurement availability, and imposed actions.

Graphical consoles offer a practical interface for medical acquisition assistance, allowing agents to work through the controls and visual feedback used by human operators. Reliable assistance requires linking on-screen anatomy to acquisition decisions that determine what image evidence becomes available next. We introduce MedImageOSWorld, a benchmark for evaluating this capability in simulated CT, MR, and ultrasound consoles. Using screenshots and mouse-and-keyboard actions, agents configure protocols, plan acquisitions, inspect the resulting images, and make corrective adjustments across seven capability levels, from console operation to feedback-driven control. Evaluation combines task-specific workflow checks with hidden anatomical ground truth to assess procedural completion and acquisition outcomes separately. A common evaluation protocol specifies episode conditions and interaction budgets, while recorded trajectories support analysis of how agents observe, act, and respond to acquisition feedback. Across eleven open-weight agents, success rates range from 3.0 to 25.0 on a 0-100 scale while workflow-progress rates reach 21.5-74.5: agents complete much of the console workflow but rarely acquire the intended anatomy. Success collapses between perception and planning, from 79-90% at the lowest three levels for the best agent to at most 12% for millimetre-level planning and 0% for closed-loop control among open-weight agents. Two proprietary agents reach success rates of about 34 and exceed the best open-weight agent mainly in acquisition quality (61 versus 42). MedImageOSWorld provides a controlled setting for studying whether general-purpose GUI agents can translate visual observations into effective medical acquisition decisions.

We prove lower bounds for Fourier transforms of surface measures $$|\widehat{\psi\,d\sigma}(\xi)| \gtrsim_{\psi, K, \sigma, R_0} \sigma(B(p, |\xi|^{-1})), \text{ given }p\in K\Subset\{\psi>0\} ,\ \vec{n}_{p}\parallel \xi,\ |\xi|\ge R_0,$$ on convex hypersurfaces of finite type with at most one vanishing principal curvature at each point. These bounds are uniform in the choice of $p\in K\Subset \{\psi>0\}$ and match the classical upper bound of Bruna-Nagel-Wainger. The key new analytic ingredient is a uniform lower bound for parameter-dependent oscillatory integrals on $\mathbb{R}$ with convex phases. We also give convex examples of finite type with two vanishing principal curvatures for which $$\inf_{p\in K}|\widehat{\psi\,d\sigma}(R \vec{n}_p)|\lesssim_N R^{-N},\ \text{ for arbitrary fixed } \{0\}\Subset K\Subset\{\psi>0\}.$$ Though the estimate above has its own interest, our original motivation is to study Fourier frames for surface measures. In particular, we show that, if $\Gamma$ is a closed convex smooth surface in $\mathbb{R}^d, d\geq 2$, of finite type and at most one principal curvature vanishes at each point, then the surface measure on $\Gamma$ does not admit any Fourier frame. This weakens the non-vanishing curvature condition in a previous result of Iosevich, Lai, the second author, and Wyman. On centrally symmetric surfaces of non-vanishing Gaussian curvature, we obtain more domains without Fourier frames. For example, we show that $\{x\in S^{d-1}:0<x_d<\delta\}, d\geq 3$, does not admit any Fourier frame for all $\delta>0$, which generalizes the result of Chen and the second author on the hemisphere.

Predicting security incidents is a profound task critical for informing proactive defensive measures and cyber-insurance policies. Prior work tackling this problem mainly utilized structured, manually defined features based on network measurements (e.g., protocol misconfigurations). Still, despite leading to promising performance, the network-based features may fail to capture aspects related to adversaries' motives. To fill this gap, our work leverages geopolitical data mined from public sources–which may help capture attacker motives–to forecast security incidents. Specifically, our approach relies on news articles and transcribed podcasts that are fed to large language models to automatically produce rich representations. The representations are then fed to a classifier trained to forecast future incidents based on historical ones. Our evaluation with a large incidents dataset (>15,700 records) demonstrates substantial accuracy (71.4% ROC AUC) with geopolitical data alone. Notably, combining geopolitical data and network measurements outperforms the state-of-the-art technique based on network features alone (81.3% vs.77.5% ROC AUC). Our analysis also helps shed light on when geopolitical data is most helpful and the data sources that are most useful for accurate forecasting.

We present a theoretical framework for quantitative oxidation state mapping of transition-metal and rare-earth elements from near-atomic-resolution electron energy-loss spectroscopy (EELS) in complex oxide heterostructures. The total inelastic differential cross section of transition-metal L$_{2,3}$ and rare-earth M$_{4,5}$ edges is calculated by combining the relaxed Dirac-Hartree-Fock ionization cross section with ligand-field multiplet theory for the discrete transitions. We apply the approach to two correlated-oxide heterostructures, La$_{0.67}$Sr$_{0.33}$MnO$_3$/LaCr$_{0.2}$Mn$_{0.2}$Fe$_{0.2}$Co$_{0.2}$Ni$_{0.2}$O$_3$/SrTiO$_3$ and NdNiO$_3$/La$_{0.67}$Sr$_{0.33}$MnO$_3$/NdGaO$_3$, which include a wide range of transition-metal and rare-earth elements with overlapping features. The method resolves the oxidation states of Cr, Mn, Fe and Co and identifies the Ni contribution within the LaCr$_{0.2}$Mn$_{0.2}$Fe$_{0.2}$Co$_{0.2}$Ni$_{0.2}$O$_3$ layer at near-atomic resolution. We find predominantly Co$^{2+}$ together with mixed Mn$^{2+}$/Mn$^{3+}$/Mn$^{4+}$ character, indicating site-specific valence redistribution within the layer. We also separate overlapping spectral features, such as La-M$_{4,5}$ and Ni-L$_{2,3}$ edges in the NdNiO$_3$/La$_{0.67}$Sr$_{0.33}$MnO$_3$/NdGaO$_3$ heterostructure, thereby recovering the intrinsic Ni-L$_{2,3}$ excitation spectrum. We find that the effective Ni valence in NdNiO$_3$ deviates from the simple ionic picture. More broadly, this work establishes a generic method for quantitative valence mapping from near-atomic-resolution EELS, applicable to a wide range of experimental conditions for the study of local electronic structure of complex oxides.

Striped jets, characterized by magnetically dominated regions with alternating polarity, can enable fast magnetic energy dissipation in global current sheets formed in between the stripes, and are thus considered as a viable source for the high-energy emission observed in blazars and gamma-ray burst prompt emission. We perform the first global axisymmetric GRMHD simulations that follow the evolution of striped jets from their launching, by accretion of poloidal magnetic loops with alternating polarity onto a spinning black-hole, to their energy dissipation at large distances. The accretion of each loop launches a Poynting-flux-dominated mini-jet carrying a helical magnetic field whose polarity is inherited from the loop, and the collection of which forms the jet core. The relativistic core is surrounded by a non-relativistic sheath composed of returning fieldlines from previously launched mini-jets in a nested configuration. Efficient mini-jet launching requires poloidal loops in the accretion disk to form close to the black hole ($r \lesssim 50\,r_g$); otherwise, they lose coherency due to turbulence. Once launched, the mini-jets accelerate electromagnetically, similar to steady-state jets, with the Lorentz factor increasing linearly with the jet cross-sectional radius as long as lateral causal contact with the jet axis is maintained. Current sheets form when two trailing mini-jets carrying oppositely directed magnetic fields come into contact. Early sheet formation is facilitated by the longitudinal expansion of the mini-jets, driven by internal pressure before the flow becomes super-fast-magnetosonic. We measure an energy dissipation rate of a few tens of percent of the jet Poynting power in the observer frame, within our simulation box of $10^4\,r_g$. Our results can account for the bright superluminal features observed in BL Lac type blazars.

Event-informed forecasting requires translating reports and historical responses into changes to a numerical forecast. We propose TIMBRE (Temporal Integration of Memory-Based Responses and Evidence), which combines source-aware representation, state-conditioned response transfer, and reliability-guided fusion before a frozen forecast head. A separate readout adjusts interval widths while preserving the median. In a single-seed, one-epoch development study of 13 tasks, TIMBRE improves MAE over ordinary fusion on eight tasks but over native Chronos-2 on only two. Disabling response transfer in the trained model reduces BTC and AULF MAE by 47.04% and 6.81%, respectively. These findings identify sensitivity to learned response transfer rather than a general forecasting advantage. Missing development-set scores and the absence of retrained ablations limit attribution to individual evidence mechanisms.

An agent with long-term memory can answer from a record it should no longer use, such as a plan the user later cancelled. We ask what a memory store must expose for an agent to know this before retrieving anything, and we score that judgment on its own, as metamemory monitoring. Readers see only a value-free summary: record counts by lifecycle state and a list of attribute names. Each of 148 questions is asked against three versions of one store that differ in one attribute, so wording cannot give the answer away. What a model adds depends on the length of that list. On the short lists the benchmark's ground truth produces (2.6 names on average), none of five language models reliably beats a cosine-similarity lookup over the names, and the three stronger ones are equivalent to it within 0.05. Under a control that fixes counts and list length, none is reliably above it and none is shown equivalent. On lists longer than the stores we tested write, padded to 60 names, the lookup loses 0.18; GPT-5.6 Luna and Sol lose 0.06 to 0.07 and lead it by 0.13 to 0.14, while the other three fall with it. The lead holds for Sol against a sibling padded to the same length and counts. Reworded to share no word with the names, the lookup loses 0.02 and one stronger reader edges 0.04 ahead. On the short lists the three stronger readers lead only where the summary counts a state without naming it, and one count feature closes that lead within a question, though not when items are pooled. The backends we tested omit or misstate this information: under FR-Bank's own metadata, GPT-4.1 mini, Haiku 4.5 and the lookup fall from 0.73, 0.82 and 0.75 to 0.60, 0.71 and 0.59. In these tests, what the store wrote down limited every reader we gave it to. All stores are synthetic; control, whether a better judgment yields a better answer, is left to a later study.

Recent incidents show that AI agents sometimes reach measured goals through unsanctioned means. This study applies self-control, general strain, anomie, neutralization and routine activity theory to reward hacking in generative AI models, and it treats the measures as behavioral analogues. Study 1 (2,310 conversations, seven models) measured delay discounting with the Kirby Monetary Choice Questionnaire and stated willingness to take shortcuts. Pressure raised the discount rate k 2.8-fold in fresh conversations but 12.6-fold when the same sentence followed a baseline answer, which indicates a response to conversational cues rather than a stable trait. Models chose a shortcut in 1 of 700 dilemmas when answering as themselves and in 64 of 700 when asked to assume human impulses, each step of pressure raised the odds by 40%, and shortcut answers contained far more techniques of neutralization (rate ratio = 146). In the preregistered Study 2, five models worked on 20 coding tasks whose tests contradicted their specifications. Two Claude models never cheated. GPT-5.6, Qwen and DeepSeek cheated in 86%, 69% and 65% of episodes and clearly disclosed the conflict in 27%, although their reasoning recognized it in 95%. GPT-5.6 had never endorsed a shortcut in Study 1. The registered effects of pressure and of an auditor cue did not survive correction for multiple testing. In exploratory analyses, two further models cheated in 69% and 100% of episodes, and one sentence stating that the specification takes priority eliminated cheating in all 280 episodes. Therefore, stated refusal does not guarantee compliant agent behavior.

Missing modality remains a longstanding challenge in multimodal learning. Existing methods typically address this issue through modality recovery or adaptive strategies. However, they overlook models' internal cross-modal dependencies formed during multimodal training, which later impair robustness. We systematically characterize a counterintuitive deployment-time failure mode: models trained on full modalities can underperform unimodal models when one modality is missing at inference time. This pattern appears across diverse architectures, such as fusion models, CLIP-style two-tower models, and vision-language models. We show that such degradation is closely associated with learned cross-modal dependencies in the principal parameter subspaces. Multimodal training induces structured rotations of these subspaces, particularly in cross-modal interaction layers. These rotations are associated with reduced task-aligned margins and larger task-aware representation harm under missing-modality inputs. We propose Geodesic Unlearning (GU), a lightweight parameter-editing method that leverages Grassmannian subspace geometry for structured subspace correction to improve missing-modality robustness. It rotates the principal input subspace toward a unimodal reference along a geodesic path. We prove that this correction minimizes the distance to the reference within a fixed subspace-distance budget. Experiments across architectures and datasets show that GU improves performance under missing-modality inference while preserving full-modality accuracy, outperforming strong missing-modality robustness baselines. These findings support a geometric view of deployment-time missing-modality degradation and suggest localized subspace editing as a practical route for robustness correction.

At the core of modern prompting techniques is contextual sensitivity, the ability of large language models to adapt their predictions based on inference-time context. Despite its central role, inference behavior under strong contextual influence remains poorly understood, particularly at the level of internal inference dynamics. We introduce a theoretical framework for analyzing contextual influence through inference dynamics, enabling quantitative characterization of inference behavior beyond output-level answer changes. Our analysis shows that inference dynamics do not exhibit unbounded drift under repeated contextual assertions. Instead, predictive representations converge to stable, query-dependent regimes that fundamentally constrain whether contextual signals can alter a model's prediction. This leads to a surprising finding: Repeated contextual assertions do not act as accumulating evidence during inference and may therefore fail to alter a model's prediction even under unbounded repetition, while in other cases a prediction change becomes inevitable. We empirically validate our theoretical predictions, demonstrating strong alignment between theory and observed inference behavior. These contributions offer a principled pathway toward characterizing the limits of contextual influence during inference, providing practical implications for model development.

We present a Bayesian framework for assessing the calibration of nested uncertainty sets from simulation studies. Calibration is typically evaluated by repeatedly simulating data from a known model and comparing empirical coverage probabilities with their nominal values, an approach that often requires hundreds or thousands of independent model evaluations to obtain reliable estimates. Such computational demands can be prohibitive when the forward models are expensive to evaluate, for example in data assimilation using a large Earth-system model, or may not be applicable, for exmaple when the output of the simulation is a single or multiple uncertainty sets rather than a full distribution. We present a method that leverages the nested structure of uncertainty sets across nominal coverage levels to assess the calibration of these models. The method, which is based on a Dirichlet-Multinomial model, allows for testing both exact and region-based notions of model calibration in an efficient manner using Bayes factors. Unlike methods that require a full predictive distribution, this method works for procedures that directly produce prediction or confidence intervals. We demonstrate the method on a data assimilation problem containing model form error and show that it is capable of detecting miscalibration even when there are only a limited number of simulation studies available.

The adjoint method is a standard approach for computing gradients in PDE-constrained optimization and central to parameter estimation, model calibration, and optimal control. This tutorial presents a discrete, implementation-oriented introduction to the method with an emphasis on conceptual understanding. Starting from an inverse problem governed by a partial differential equation, it develops the corresponding optimization problem and derives the adjoint gradient from the perspective of iterative descent methods. The roles of the state, the parameters, the objective, and the constraint are made explicit, and the adjoint equation is obtained directly from the gradient computation rather than postulated. Particular attention is given to explicit derivations, implementation details, and the connection between the mathematical development and the resulting computational algorithm. The tutorial also examines the influence of discretization on the adjoint construction, explains the origin of the backward-in-time adjoint recursion, and presents practical procedures for gradient verification. A vibrating-string inverse problem serves as a running example, illustrating the complete workflow from model formulation to adjoint-based optimization. The presentation is intended for advanced undergraduate students, graduate students, and practitioners seeking a step-by-step understanding of adjoint gradient computation.

No-signalling constraints on isolated outcome probabilities do not determine whether an apparatus preserves the joint law of its accessible records under repeated readout, network composition, or reuse. We study these extensions through a conditional-record criterion and four explicit models. For a wrapped-Cauchy retrocausal source, protection under two conditionally independent continuous readouts forces an angle-independent response; three readouts suffice without continuity. A finite angular-bandwidth assumption gives a quantitative leakage bound, whereas continuity alone gives no uniform stability. At a specified two-source Bell junction, the required correction matrix has full nonnegative rank on every even uniform angular grid. A Gaussian controller can cancel branch-mass dependence while leaving a distinguishable joint record, for which we derive exact noise and law-change tradeoffs. Finally, a fixed singlet with shared classical noise reproduces the variable-width isolated outcome family, but retained and refreshed controllers generate different temporal statistics. We prove a finite-resource obstruction and construct positive finite-controller approximations. Numerical studies comprising 8.3 million independent simulated histories support the model-specific predictions. Together, the results identify conditions that must accompany marginal no-signalling when extending an apparatus to richer records and interventions.

Neural Architecture Search (NAS) has proved to be a strong alternative to manual network design, but applying it to 3D medical image segmentation is limited by two well-known costs, large annotation budgets and multi-GPU clusters. Thus, this paper introduces Active-DiNTS (Active Differentiable Network Topology Search), an approach that embeds pool-based Active Learning (AL) into a bi-level differentiable topology search to perform architecture discovery and label curation jointly. At each query round, unlabeled MRI volumes are ranked by one of three uncertainty signals (Entropy, Variance, or Standard Deviation), and only the top-ranked volumes are sent to an oracle for annotation. The new labels feed two interlocked stages. Network weights are updated in an outer loop, while the macro/micro topology of a U-Net-style backbone is refined in an inner loop. Three AL regimes (weights-only, topology-only, joint) expose the speed-accuracy trade-off. Evaluations on the Medical Segmentation Decathlon (MSD) Task01 BrainTumour benchmark showed that Active-DiNTS surpasses DiNTS, C2FNAS, and nnU-Net in Dice, with gains of about 10 percentage points on Edema and 5 points on Non-Enhancing core, on a single GPU and using a fraction of the labeled volumes. The discovered architectures are denser and more FLOP-heavy than prior baselines, but remain competitive in trainable parameters and peak memory; the fastest search variant finishes in under 0.25 GPU-days, over 27x faster than the eight-GPU DiNTS search. Together, these results indicate that pairing differentiable NAS with active data acquisition is a practical recipe for accurate 3D segmentation under realistic constraints.

Time-series foundation models buy zero-shot forecasting with large temporal corpora; a vision model needs none, since a natural image implicitly embeds the patterns a forecaster must model, and an ImageNet-pretrained masked autoencoder forecasts a series by inpainting a rendering of it. A rendered series is not a natural image, however, and closing that gap takes temporal-aware adaptation. We show that the rendering geometry - how the series is folded and drawn - is a controllable, mixable axis for it. Kaleido detects the dominant periods, renders a rule-generated set of fold geometries, combines the inpaintings with a convex per-position gate fit on validation only, and fuses the result with the zero-shot output at one fixed share, with no per-dataset hyperparameter beyond the baseline's published settings. Training only LayerNorm (0.05%), Kaleido lowers MSE by 13% against the published zero-shot baseline on LTSF and, frozen, by 6.6%; on GIFT-Eval it improves the baseline by 7.4% in MASE and 19.3% in CRPS.

Camera-to-motion-capture (mocap) calibration is essential for using mocap as ground truth in robotics, AR/VR, and other computer vision tasks. However, the calibration can drift after deployment, while calibration residuals and visual inspection provide limited independent verification. We present Lollypop, a fiducial-mocap reference target for independent calibration verification. The target couples an ArUco fiducial with a mocap marker constellation so the visual center and tracked centroid represent the same physical point. Given a candidate calibration, verification projects the mocap point into the image and measures its disagreement with the detected fiducial center. Experiments show sub-pixel nominal error, sensitivity to controlled extrinsic perturbations, and increasing error during an illustrative mixed-handling sequence.

Traffic classification (TC) is crucial to secure Internet of Things (IoT) networks, whose edge nodes often operate under privacy, bandwidth, and energy constraints. Yet, encrypted payloads and limited computing power make accurate, real-time TC a challenging task. Existing learning-based TC approaches often fix the input configuration a priori, even though it directly influences both predictive performance and computational cost. This paper presents RepTC, a representation-aware hardware-constrained strategy that addresses session-level TC through joint model and input optimization. The method co-optimizes network architecture, session length, and header preprocessing; this enables joint control of model complexity, input scale, and data representation within a unified resource-constrained design space. The proposed approach enforces microcontroller-class constraints on memory, model size, and computation, and yields compact models deployable on low-power edge devices. A gateway monitors traffic by aggregating sessions and either performs inference locally or offloads to a low-power edge node. Both scenarios are validated on heterogeneous embedded hardware, including a Raspberry Pi 3B+, STM32 Nucleo-F401RE, and XIAO ESP32-C3, spanning Cortex-A, Cortex-M, and RISC-V processor architectures; measured inference latency ranges from 0.63 to 18.59 ms, with MCU-side inference energy between 0.50 and 1.41 mJ per session. RepTC yields configurations that achieve high accuracy on a variety of established benchmarks: 96.21% on ISCX VPN-nonVPN, 99.62% on USTC-TFC2016, and 99.97% on Edge-IIoTset; at the same time, model size and computational requirements were reduced by up to three orders of magnitude compared with state-of-the-art methods. The results show that representation-aware optimization can improve efficiency while preserving competitive classification performance.

Counterfactual explanations are most useful when they can be generated with low latency, remain close to the factual input, and satisfy input-domain, categorical, and actionability constraints. Achieving these objectives simultaneously is challenging for deep neural networks. Heuristic methods are often fast but may return invalid counterfactuals, whereas exact methods can certify global proximity but may not finish within practical time limits. We introduce DASH, a batched heuristic search method for finding close, valid, and actionable counterfactuals for deep neural networks under $\ell_1$, $\ell_2$, and $\ell_\infty$ objectives. DASH uses directional Lipschitz bounds and local affine models to generate anchors, then ranks and expands promising regions with batched network evaluations. We compare DASH against nine prior heuristic methods and time-limited exact mixed-integer baselines on four tabular datasets, with network depths from 2 to 32, and evaluate scalability on PBMC3k. Across 9,000 tabular query-norm cases, DASH returns a valid counterfactual within $5\%$ of the best heuristic-observed valid distance in $94.6\%$ of cases, with a median CPU search runtime of $0.061$ s. PGD-bisect, the baseline with the highest pooled within-$5\%$ coverage, meets this criterion in $41.1\%$ of cases, with a median runtime of $0.298$ s. These results show that the proposed search maintains high valid proximity across norms while keeping its search runtime practical.

A search for beyond the standard model effects in high-mass dielectron and dimuon final states with the CMS detector is performed, with a specific focus on Higgs boson compositeness. The search uses data collected at the LHC at 13 and 13.6 TeV center-of-mass energies, corresponding to a total integrated luminosity of 207 fb$^{-1}$. Compositeness would manifest as distortions in the observed dilepton mass spectrum and the presence of bound states, such as the Higgs boson and new massive vector boson resonances. Assuming the existence of new boson states with masses well above the TeV scale, an effective field theory approach provides stringent bounds on the oblique parameters $\mathcal{W}$ and $\mathcal{Y}$, which are zero in the standard model. One-dimensional fits, in which only one of the oblique parameters is allowed to deviate from zero, provide the following limits: $-$3.9 $\times$ 10$^{-4}$$\lt$ $\mathcal{Y}$ $\lt$ 3.5 $\times$ 10$^{-4}$ and $-$2.1 $\times$ 10$^{-4}$ $\lt$ $\mathcal{W}$ $\lt$ 1.6 $\times$ 10$^{-4}$ at 95% confidence level. More stringent compositeness limits at the TeV scale are set in several new physics scenarios under the assumption that a new Z' resonance is physically present within or slightly above the visible dilepton mass spectrum. In this context, cases with large resonance decay widths are explored for the first time.

In an image latent space, the embeddings of high-resolution, natural, and sharp images form a manifold. Degradation of high-resolution images pushes their embeddings off this manifold. Real-world super-resolution (SR) then becomes the task of mapping the degraded embedding back onto this manifold — not anywhere on the manifold, but to the point that preserves what the input still carries, both its semantics and pixel details. Every published method implements this mapping in a reconstruction-oriented latent space or pixel space. We claim these spaces are the wrong substrates for SR. Low-resolution and degraded images are embedded far from the manifold, making the mapping difficult and expensive. The lack of semantic information in these substrates also makes it difficult to navigate to the faithful point on the manifold, causing severe hallucination when degradation is heavy. Thus, restoring in a suitable latent space is crucial to the SR task. We show that the latent space of 23 fused layers of a frozen DINOv3-L is one such space that makes the SR task easier. Degraded images are embedded near the manifold. Moreover, this substrate contains a hierarchy of information, from pixel record to degradation robust semantics, guiding the model to find the faithful point on the manifold. On this substrate, a 415M decoder is trained under reconstruction and adversarial objectives to map the degraded embeddings back and decode to pixel space in one pass. The resulting model, RAESR, attains the best fidelity–perception trade-off among state-of-the-art adversarial and diffusion-based restorers on RealSR, DRealSR, LSDIR and DIV2K-Val, at 37 ms per 512 by 512 image on a single H20 GPU. Swapping the substrate for a VAE latent under an identical recipe loses on every metric.

We present a spectral analysis of diffuse Galactic background emission using SPHEREx data in a pilot Galactic-plane field, where uncertainty in the estimation of the diffuse infrared background is a major source of systematic uncertainty in extinction studies at near- and mid-infrared wavelengths. We construct channel-resolved background maps across the 0.75–5~$\mu$m range using a Large Median Filter approach and then carry the extinction analysis forward only for the cleaner long-wavelength subset at $\lambda \geq 3~\mu$m. In the present field, comparison against an earlier masked-median background model gives a median channel-by-channel background difference of about 6% and a maximum change of $|\Delta\tau_{\rm norm}|\simeq 7.4$ near 3.205~$\mu$m, underscoring the importance of wavelength-dependent background modeling. These results demonstrate that spectral mapping with SPHEREx provides a practical framework for controlling diffuse background systematics and enables improved constraints on dust opacity and extinction in the Galactic plane.

Concurrent AI coding agents create a coordination problem in which cheap signals may prioritize work, but only an executable checker can establish the property being claimed. We study this separation through a property-scoped verification contract and MERGEGYM, a three-track benchmark for open-time scope forecasting, replay- conditioned conflict resolution, and scheduling. On a stratified 715-pair lineage set (167 textual conflicts), 79 conflicts (47.3%) occur despite disjoint authored file sets. This is an operationally important proxy mismatch expected from three-way merge: authored PR diffs are measured against PR-specific bases, whereas the checker compares both heads to their common merge base. A standalone lineage-union rule reaches held-out AUROC 0.877; a 12-feature logistic model reaches 0.882 [0.845, 0.917] with PR-AUC 0.661, and an untuned random forest on the same decision-time features reaches 0.902 [0.875, 0.928] with PR-AUC 0.701. At a 33.3% held-out replay budget, the logistic and random-forest models recover 81.2% and 82.9% of conflicts, respectively. Patch reconstruction succeeds for 48/79 zero- overlap conflicts and all 48 become clean; the other 31 cases are inconclusive, so this check validates the expected three-way-merge explanation rather than claiming a new Git mechanism. In T1, a zero-shot LLM reaches AUC 0.704 and LLM-plus- metadata fusion 0.740. In T3, a decision-time gate de-overlaps a median 91.7% of labeled scope collisions at 65.0% makespan inflation under frozen-label replay. Because local git merge-tree replay is already cheap in our logs (median 0.02 s), we do not claim that lineage triage saves this checker alone: when exact replay is cheap, verify everything. All empirical guarantees in this paper remain limited to textual mergeability or the explicitly stated frozen-label scheduling target.

Patient world models are increasingly proposed for longitudinal prediction, intervention-aware reasoning, and clinical-trial simulation. Causal or clinical intervention validity is distinct from predictive generalization and reliability; before making stronger claims, the underlying predictive state should generalize across patients, survive realistic shifts and missing observations, and expose failure through meaningful reliability signals. We evaluate these prerequisites in a deliberately narrow setting: short-horizon digital-biomarker forecasting from PhysioNet GaitPDB, comprising 165 participants, 306 recordings, and 51,129 context-future pairs. Using persistence, ridge, MLP, GRU, Transformer, and a compact JEPA-style predictor, we build an evaluation ladder that progressively removes raw temporal overlap, same-recording familiarity, and same-patient familiarity before testing unseen-patient generalization. For GRU, NMSE rises from 0.1227 under random-window splitting to 0.1393 after eliminating raw train-test overlap and to 0.1961 under patient holdout. Among 54 participants with repeated recordings, exposure to a different recording from the same patient improves GRU NMSE from 0.2177 to 0.1556, while a recording-excluded identity hypothesis is not supported at the participant level. Under participant-held-out evaluation, MLP and Transformer are statistically indistinguishable. Study shift, a four-times-longer prediction gap, and partial observation further degrade performance; under 50% temporal masking, Transformer NMSE rises to 0.611 while MC-dropout predictive variance falls. We do not claim a longitudinal or intervention-aware simulator. Instead, the results support a prerequisite evaluation stack of patient separation, repeated-measure controls, shift, missingness, and uncertainty validation before stronger patient-world-model claims are trusted.

Inflammatory bowel disease (IBD) requires frequent imaging-based assessment, yet interpretation of modalities such as wireless capsule endoscopy (WCE) and intestinal ultrasound remains heavily dependent on specialist expertise. Vision-Language Models (VLMs) demonstrate significant potential in multimodal medical image analysis, but their clinical adoption is hindered by their insufficient domain-specific reasoning, susceptibility to hallucination, scarcity of high quality training data in fine-grained diagnostics and limited interpretability. We introduce ARISE (Adaptive Agentic Reasoning with Image-grounded Self-Evaluation), an autonomous planning framework that models few-shot medical image understanding as a sequential agentic workflow. ARISE structures agent execution into a transparent 5-stage workflow: hypothesis generation, image-grounded evidence summarization, evidence-conditioned refinement, symbolic verification, and final diagnosis. We apply ARISE to IBD assessment across two independent patient cohorts: wireless capsule endoscopy (WCE) images for Crohn's disease and B-mode ultrasound data for ulcerative colitis. ARISE consistently improves diagnostic performance over baseline VLMs while exposing where reasoning succeeds or fails, providing a more interpretable basis for clinical decision support and realistic deployment.

This paper studies distortion-aware clustering for uplink fronthaul-limited cell-free MIMO networks employing maximum ratio combining (MRC). While MRC is appealing for its low-complexity, its performance is limited by interference, fronthaul distortions, and channel imperfections. This motivates clustering strategies that compensate for these impairments through coordination of access points. Since instantaneous small-scale fading-based optimization is not scalable in large systems, and is impractical due to frequent channel variations, we instead optimize an objective depending only on large-scale fading coefficients. To this end, asymptotic analysis is used to derive deterministic equivalent expressions for the average network sum rate, with focus on quantization distortion, leading to a quadratic-over-linear objective. Although maximizing such an objective under binary constraints is non-convex, we exploit its structure to develop a polynomial time scheme that attains the global optimum. Numerical results show that the proposed clustering method provides performance gains over improved variants of existing literature, and remains competitive with the small-scale fading-based global optimum obtained via exhaustive search.

In this paper, we present a distributed cascade force control system (DCFC) for multiple robots with the aim of pushing a rigid object towards a desired moving target without their inter-robot communication. These mobile robots are equipped with 360-degree vision-based soft tactile sensors utilized to determine contact location and resultant impact force. By investigating the dynamics of moving rigid objects on the flat, we proposed a distributed cascade control. The inner loop control incorporates contact force and positioning, ensuring the robots' pushing contact and application of the desired force to the object. The outer loop control coordinates the robots to push the object in a desired direction without inter-robot communication, regardless of the unknown object mass. The stability and convergence of the system are verified using the Lyapunov stability theory. We also conducted simulation and real-word experiments to validate the performance of the proposed control method, and the experimental results showcase the successful coordination of multiple robots in pushing an object towards a moving desired direction.