arXiv:2606.08680v1 Announce Type: new Abstract: Fisheye cameras are widely deployed in autonomous driving perception suites for their low cost and full-coverage field of view (FOV), yet their potential remains underleveraged in 3D object detection. Severe radial distortion challenges most BEV detectors by violating the fundamental assumption of uniform sampling. To bridge this gap, we propose Distortion-Aware PETR (DAPETR), a projection-free detector tailored for mixed pinhole-fisheye camera setups. DAPETR incorporates two key learned-adaptive modules: a unified distortion-aware positional embedding that harmonizes positional encodings for image representations with fisheye geometry, and a bidirectional feature-geometry co-modulation module that mutually adapts image features and 3D positional embeddings. In our experiments on a converted KITTI-360 benchmark, we systematically compare our learned adaptive approach against PETR in polar coordinates (PolarPETR). We find that while both methods improve over the baseline, our learned modules achieve superior performance. Crucially, we uncover a negative interaction when combining both strategies, revealing that learned adaptation and explicit geometric reparameterization can conflict. Our final DAPETR model significantly advances the research and benchmark for fisheye BEV detection, providing critical insights into effective distortion-aware 3D perception design other than image rectification.
Science Journals
arXiv:2606.08394v1 Announce Type: new Abstract: Multimodal language models are typically evaluated through external behavior: selecting the correct image--text match, rejecting unsupported captions, or answering visual queries correctly. However, correct behavior alone does not show that the model's internal decision state remains stable under controlled semantic stress. We study this gap through S$^3$E (Structured Semantic Stress Evaluation), a framework for analyzing behavior-internal decoupling in multimodal language models. S$^3$E uses a positive-anchored A/B forced-choice setup in which an image-supported caption is contrasted against semantic stress candidates under both original and swapped option orders, while hidden states are extracted at the pre-answer decision state. We focus on strict-correct trials, where the model consistently selects the correct caption across both orders. Rather than treating arbitrary hidden-state variation as evidence of instability, we measure whether semantic-conflict candidates induce excess decision-state displacement relative to meaning-preserving controls. Across Qwen3VL, Gemma3, and InternVL3, semantic stress consistently produces positive selected-layer excess displacement over lexical controls despite correct forced-choice behavior, while comparisons against random negatives are model-dependent. We interpret this as a scoped decision-state stress-sensitivity signal rather than evidence of downstream failure or hallucination. Our results suggest that forced-choice correctness alone is not a sufficient certificate of invariant internal decision geometry.
arXiv:2606.08400v1 Announce Type: new Abstract: Graduate-level research reading report assessment creates a substantial labor burden for educators. While large language models (LLMs) hold great potential for automating academic grading, their reliability for this specialized task remains understudied, particularly regarding grading consistency, the lack of which represents a primary obstacle to educational fairness. This paper proposes a human-aligned LLM-assisted grading workflow and presents a case study based on 180 student submissions from a graduate advanced software engineering course. We evaluate two mainstream LLMs, Grok and GPT, in terms of grading consistency and alignment with human scores. We find LLMs exhibit distinct levels of intra-model consistency and significant inter-model grading inconsistencies, while simple ensemble approaches cannot improve alignment with human evaluation. Critically, continuous interaction history drives systematic drift in models' grading standards away from human expert scores. Our findings demonstrate LLMs' potential in reducing grading workload for educators in graduate education, while highlighting that indiscriminate LLM grading may introduce systemic unfairness, suggesting that specific operational practices are required to mitigate such disparities.
arXiv:2605.26108v3 Announce Type: replace Abstract: Recent advances in few-step diffusion distillation have enabled efficient image generation, yet aligning these models with human preferences remains challenging. We propose Reward-Tilted Distribution Matching Distillation (RTDMD), a two-stage framework that unifies distribution matching distillation with reward-guided reinforcement learning for few-step flow generators. We show that minimizing the KL divergence to a reward-tilted teacher distribution naturally decomposes into a distribution matching term and a reward maximization term. In the first stage, we introduce Ambient-Consistent Distribution Matching Distillation (AC-DMD), which performs subinterval-wise distribution matching and augments the fake score objective with a consistency regularizer to help the fake score model track the shifting generator distribution under limited updates. In the second stage, we jointly optimize both terms: for the reward maximization term, we derive a hybrid policy gradient that combines a GRPO-style estimator for the stochastic intermediate transitions with direct reward backpropagation through the deterministic final step, and further introduce step-subset GRPO (SubGRPO) to reduce variance. Experiments on SD3, SD3.5, and FLUX.2 demonstrate that RTDMD establishes new state-of-the-art results across preference, aesthetic, and compositional metrics with only 4 inference steps, outperforming previous few-step text-to-image generation methods. Code and models are available at https://github.com/Harahan/RTDMD.
arXiv:2606.01869v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly asked not only to write static interfaces, but to construct executable interactive worlds from natural language. Browser-native 3D, commonly built with Three.js, is a natural next frontier: generated programs must integrate assets, obey spatial and physical constraints, and keep user-facing controls synchronized with hidden runtime state. Existing web-generation benchmarks and evaluators, however, largely observe only pixels or DOM nodes, while the mechanics of a Three.js world unfold inside an opaque . We introduce WorldCoder-Bench, a benchmark for autonomous, physically grounded 3D world synthesis. WorldCoder-Bench contains 2,026 expert-curated tasks across Simulation, Rendering, and Application scenarios, with optional .glb assets and hidden behavioral contracts. We further propose StateProbe, an execution-based protocol that probes generated programs in a sandboxed browser and verifies hidden, mutation-hardened contracts over runtime states and transitions. Beyond verification coverage, we report Return on Automation and Time Efficiency Multiplier to measure correctness-adjusted cost and time savings. Across nine frontier models, the best system reaches only 27.8% verification coverage on WorldCoder-Core and 19.9% on WorldCoder-Robust, with failures dominated by state-schema drift and broken interaction chains rather than missing scene elements. Utility metrics further show that cheap or fast models can still provide substantial value on easier domains. WorldCoder-Bench is available at https://anonymous.4open.science/r/WorldCoder-Bench/.
arXiv:2606.02274v2 Announce Type: replace Abstract: End-to-end manipulation policies, combined with web-scale pretrained Vision-Language Models (VLMs), show the promise for generalizable and dexterous robotic manipulation. However, they inherit two key limitations from 2D foundation models: 1) the reliance on 2D RGB inputs that ignores the intrinsically 3D nature of manipulation; and 2) the lack of spatial 3D alignment between input-output spaces as well as across diverse robot embodiments, camera setups, and trajectory datasets. In this paper, we present a series of contributions to address these issues. First, we introduce aligned vertex map and vertex spectrum -- a pixel-wise 3D representation that elevates 2D visual inputs to 3D, using camera calibration and optional depth. This novel input representation marries 3D awareness with the generalization of 2D large VLMs. Then, we propose to align the inputs and outputs of manipulation policies by expressing per-pixel 3D information of each camera view and robot actions to a shared coordinate. Based on this, we designate a canonical Bird's-Eye-View (BEV) alignment frame and innovatively propose to construct BEV images, producing a view-invariant representation robust to camera pose variations. To enable training and evaluation at scale, we develop a comprehensive data processing pipeline to perform such alignments; we also introduce a novel temporal alignment scheme for trajectories across diverse robots, human operators, and datasets. These contributions collectively mitigate input and output spatial-temporal misalignments, improving the consistency and generalization for real-world manipulation. Pretrained checkpoint, source code and data processing pipeline are available in https://hnuzhy.github.io/projects/Dex-BEV.
arXiv:2606.08411v1 Announce Type: new Abstract: Block-wise semi-autoregressive decoding is the standard inference paradigm for diffusion large language models (DLMs), but it imposes a strict dependency between blocks: the next block cannot begin until the current block is fully decoded or its denoising budget is exhausted. We observe that once a block exposes a reliable delimiter boundary or stable semantic prefix, continuation generation need not wait for every residual token to be resolved. We propose AsyncLane, a training-free decoding scheduler that decouples refinement from advancement. AsyncLane forks a generate lane at observed delimiter boundaries into a refine lane and a continuation generate lane: the prefix remains editable, while the continuation advances before prefix refinement finishes. The resulting lane tree records decoding dependencies and output order, while execution proceeds over the active lane set. To make this asynchronous schedule efficient under bidirectional attention, AsyncLane combines shared-prefix lane batching, lookahead draft reuse, cascading termination, and compact cache refresh with refresh-logit reuse, preventing model-call cost from scaling directly with the number of lanes. AsyncLane is a drop-in replacement for block-wise DLM samplers and requires no retraining. Experiments on mathematical reasoning and code generation show that AsyncLane consistently improves throughput while maintaining competitive quality. Across LLaDA and Dream backbones, AsyncLane achieves the highest TPS in all evaluated benchmark-length settings; relative to the fastest competing baseline, it reaches peak speedups of 2.95x on LLaDA and 3.04x on Dream, with especially large gains under longer generation budgets.
arXiv:2606.02802v2 Announce Type: replace Abstract: Large language models (LLMs) exhibit strong natural-language reasoning abilities for clinical decision support, but struggle to effectively model structured longitudinal electronic health records (EHRs). In contrast, EHR foundation models can learn predictive patient representations, yet lack interpretable language-based reasoning. To bridge this gap, we propose ChatHealthAI, a multimodal reasoning framework that aligns structured EHR representations from a pretrained EHR foundation model with the semantic space of a frozen LLM through a task-aware resampler. By integrating longitudinal patient representations with refined clinical event descriptions, ChatHealthAI enables clinically grounded natural-language reasoning while maintaining accurate patient prediction. We evaluated ChatHealthAI on three clinical predictive tasks from the EHRSHOT benchmark. Results show that ChatHealthAI improves reasoning quality and interpretability while preserving competitive predictive performance. These findings highlight the potential of integrating EHR foundation models with pretrained LLMs for interpretable clinical prediction.
arXiv:2606.03258v2 Announce Type: replace Abstract: We investigate the linear stability of a rotating fluid annulus subjected to localized heating at the outer periphery of the bottom surface and uniform cooling at the inner cylindrical wall through a rigorous stability analysis. The localized forcing generates a non-axisymmetric base state, invalidating the classical normal-mode decomposition. We employ Floquet-Bloch theory in the azimuthal coordinate combined with a BiGlobal eigenvalue formulation in the meridional plane. The non-axisymmetric base state is expanded in azimuthal Fourier harmonics; perturbations are expressed as quasi-periodic Bloch modes that couple all azimuthal wavenumbers through base-state harmonics. Full linearised perturbation equations, the BiGlobal block-operator structure, pressure elimination, solenoidal projection, and the modal energy budget are derived. Instability is driven by cross-modal baroclinic energy release and shear production - mechanisms absent in classical axisymmetric theory.
arXiv:2606.03787v3 Announce Type: replace Abstract: Robots solving generalist tasks need to be able to ground instructions in their past experience, since humans may refer to notable past events when giving a task (e.g., ``Take me to where the chemical spill happened yesterday''). Since memory limits make storing all past events infeasible, long-term robot memory must be selective, ideally retaining only those episodes with high utility for future tasks. However, future tasks are not typically given a priori for generalist robots. To select generically useful memories, we propose Bayesian surprise as a gating mechanism for memory formation. We present an approach to compute surprise in a semantically rich deployment-agnostic latent space provided by V-JEPA-2. Using our gated episodic memory to augment 4D scene graph-based spatial memory, we show a consistent improvement over state-of-the-art benchmarks in robot question answering, outperforming prior robot memory methods by $\geq12\%$ for temporal, spatial, and binary questions, and surpassing the performance of supervised and non-causal methods with an unsupervised causal method in event segmentation tasks.
arXiv:2606.04029v2 Announce Type: replace Abstract: Reinforcement Learning (RL) has received increasing attention and adoption in real-world use cases. Most of these systems follow a train-then-fix paradigm, where trained agents do not learn while interacting with the world until performance degrades and retraining becomes necessary. In this position paper, we argue that deploying an agent that is incapable of optimality, but receives an evaluative reward signal, is inherently a continual RL problem. We identify four sources of non-stationarity after deployment that necessitate never-ending learning, and highlight why the best deployed agents never stop adapting. We analyze successful examples of continual RL in the real world, and present the community with the advantages and measures to move away from the current train-then-fix paradigm.
arXiv:2606.04109v2 Announce Type: replace Abstract: Context-augmented language model systems often wrap supplied content with labels such as Reference:, Evidence:, Instruction:, Note:, or Example:, but the effect of these labels on reader-model behavior remains underexplored. We introduce a paired fixed-content probe over 500 MMLU-Pro items: each item receives the same misleading answer-bearing assertion under different discourse-role labels, and adoption is measured by whether the model outputs the injected wrong option. Across GPT-5.5, DeepSeek V4 Pro, Llama-3-8B-Instruct, and Qwen2.5-7B-Instruct, Misleading Adoption Rate shifts by 56-84 percentage points. Binding or source-like labels such as Instruction: and Reference: produce high adoption, whereas Example: consistently suppresses it. Paired tests, bootstrap intervals, final-instruction ablations, and Qwen final-step log-probability probes support a label-conditioned candidate preference. Boundary probes show where the effect weakens or persists: arithmetic tasks reduce adoption, passage-shaped external context preserves smaller label gaps, short-answer evaluation rules out option-letter copying, and nested-label conflicts suggest that illustrative framing can delimit adoption scope. A 200-case single-author manual audit confirms that the short-answer contrasts are stable under conservative adjudication. The resulting claim is bounded but practical: context-utilization and reader-side RAG benchmarks should report and control wrapper labels, because presentation choices can change measured reliance on supplied context.
arXiv:2606.08413v1 Announce Type: new Abstract: We present a structured analysis of how contemporary clinical AI systems integrate electronic health record (EHR) data and the extent to which they support longitudinal clinical reasoning. Drawing on a curated corpus of clinical natural language processing (NLP) and EHR-integrated systems, we develop a coding framework that captures both technical integration strategies and reasoning-relevant representational features, such as trajectory modeling, cross-encounter synthesis, longitudinal analysis, and absence reasoning. We also elicited the experiences of three physicians in their EHR use, including what strengths and weaknesses they found with their institution's current EHR system(s). Our analysis shows that while many systems incorporate EHR data, they predominantly operate on encounter-level or aggregated representations, with limited support for explicit temporal reasoning across patient histories. Reasoning-relevant structures are inconsistently represented, and evaluation paradigms remain largely focused on predictive performance instead of longitudinal interpretability. We argue that current approaches treat EHR data as a static input rather than a substrate for ongoing clinical reasoning, and we outline a framework for understanding how future systems might more effectively align with the temporal and interpretive structure of clinical practice.
arXiv:2606.08968v1 Announce Type: new Abstract: Surface-level reactor antineutrino experiments usually have substantial cosmic ray induced neutron backgrounds, particularly with shallow overburden. The Array of Lattice for Anti-neutrino Reactor Monitoring (ALARM) is a plastic scintillator based experiment designed for reactor power monitoring. It will be deployed about 44 m from the core of a reactor at the Taishan Nuclear Power Plant. Placed at a depth of 9.6 meters below the surface, cosmic ray induced fast neutrons constitute a significant background, making an effective neutron shielding system essential for the experiment. For the shielding design of ALARM, we tested the shielding performance of three materials water, HDPE, and 40\% boron-doped HDPE (BHDPE) against both fast and thermal neutrons. A thermal neutron detector composed of an EJ426 scintillator setup was first used to measure the shielding efficiency of these materials at various thicknesses using neutrons from an Am-Be source. A 30-cm thickness of BHDPE achieved a shielding efficiency exceeding 95\% for both fast and thermal neutrons. Monte Carlo simulations of the EJ426 setup yielded results consistent with the experimental data. Simulation results for the shielding performance of the full ALARM shielding assembly are also presented.
arXiv:2606.08712v1 Announce Type: new Abstract: Purpose: Spatial transcriptomics (ST) enables gene expression measurements within the tissue context. However, these measurements are often noisy, low-resolution, and sparsely sampled, which limits the recovery of fine spatial structure. Deep neural networks have become powerful tools for expression imputation from histology, but their performance remains constrained by limited sample sizes and a lack of biologically informed augmentation. Most of the existing augmentation strategies for learning are designed for classification tasks rather than regression, which neglect spatial and transcriptomic relationships, leading to biologically implausible interpolations that hinder prediction performance. Approach: To address these limitations, we propose SNR-ST-Mix, a geometry- and expression-aware data augmentation framework designed specifically for ST data. It constrains mixing to a spot's k-nearest spatial neighbors and adaptively weights interpolation coefficients based on expression similarity, generating augmented samples that preserve local biological structure while ensuring spatial smoothness. This dual conditioning yields synthetic examples that expand the effective training manifold, promote generalization, and enhance prediction stability under sample-specific training. Results: Extensive experiments with various tissue types demonstrate that SNR-ST-Mix consistently outperforms conventional augmentation methods without requiring architectural changes or additional computation. Conclusions: SNR-ST-Mix provides an effective and biologically principled augmentation strategy for spatial transcriptomics regression tasks. By explicitly leveraging spatial geometry and transcriptomic similarity, it expands the effective training manifold and improves predictive performance without increasing model complexity.
arXiv:2606.09495v1 Announce Type: new Abstract: Modern object detectors achieve strong performance on standard benchmarks, yet their robustness to contextual variation remains insufficiently understood. Prior evaluations largely rely on aggregate metrics such as AP on uncontrolled distribution shifts, which can obscure how performance degrades under context change. We introduce ContextShift, a controlled benchmark that systematically manipulates object--context relationships while preserving object appearance. Built on COCO 2017, it isolates context as an independent variable through geometric transformations and synthetic and natural background substitutions, including a continuous compatibility axis based on normalized pointwise mutual information (NPMI). Across diverse detector architectures, we observe a consistent degradation pattern: false negatives increase by up to 227% and prediction volume decreases by up to 44%, while false positives remain stable or decline. This suppression behavior is not captured by aggregate metrics such as AP, which can mask substantial recall loss and changes in prediction dynamics. Further analysis suggests that degradation is driven less by reduced confidence than by a reduced formation of valid detection candidates. Moreover, performance along the statistical compatibility axis is non-monotonic, peaking at intermediate NPMI and degrading toward both extremes, indicating that statistical co-occurrence does not correlate linearly with effective visual context. Finally, we show that context-aware augmentation improves robustness: every augmented variant outperforms the dataset-only baseline on both original and manipulated test images, partially recovering performance lost to prediction-suppression failures by exposing models to object--context decoupling during training.
arXiv:2606.08421v1 Announce Type: new Abstract: Multi-contrast brain MRI provide complementary soft-tissue characteristics that aid in the screening and diagnosis of diseases. However, limited scanning time, image corruption and various imaging protocols often result in incomplete multi-contrast images. While current approaches excel in image synthesis, they often struggle to synthesize critical tumor regions and exploit contextual information in multi-contrast brain MRI effectively. To address this issue, we propose a synthesis-centric, segmentation-assisted closed-loop framework with retrieval augmentation synthesis. Our method overall takes a generative adversarial architecture, which aims to synthesize missing contrasts from any combination of available ones with a single model. To explicitly capture tumor semantics and focus synthesis on tumor regions, we add an auxiliary segmentation branch that predicts tumor masks and feeds them back as semantic conditioning in synthesis branch, thereby learning tumor-aware representations in the model and improving synthesis fidelity. Furthermore, we propose a dual-bank retrieval augmentation strategy. It dynamically queries two external knowledge bases, namely a tumor masks memory bank for crucial tumor context and cross-image contrast feature memory bank for global style information, to augment synthesis. Verified on two public multi-contrast magnetic resonance brain datasets: BraTs2020 and UCSF-BMSR, the proposed method is effective in handling medical brain images synthesis tasks and shows superior performance compared to previous methods. Code is available at:https://github.com/iBizzard/SSCF.git
arXiv:2606.08426v1 Announce Type: new Abstract: We present CritLens, a visual analytics system that helps users build personalized multi-criteria decision models from review text. In everyday decisions -- choosing equipment, hotels, or restaurants -- evaluation criteria are either preset by platforms or generated by LLMs, leaving users unable to discover, adjust, or verify them against the underlying evidence. This is problematic because many preferences are latent: they surface only upon encountering specific reviews, and any fixed framework risks overlooking low-frequency but decisive details. CritLens addresses this gap by using LLMs to transform reviews into an initial AHP decision model, then supporting iterative, human-in-the-loop refinement. Through coverage gap detection in the embedding space, users discover criteria missed by the initial model; through interactive weight adjustment under AHP consistency constraints, they express personal priorities; and through a multi-level scorecard and exportable decision report, they trace every ranking back to the original review text. Two case studies, an eight-participant user study, and a quantitative consistency-repair experiment demonstrate the system's effectiveness.
arXiv:2606.08431v1 Announce Type: new Abstract: This paper studies neural ordinary differential equations (neural ODEs) from a control-theoretic perspective using controllability and observability concepts. The neural ODE is represented in a control-affine form to facilitate analysis using tools from nonlinear and linear time-varying (LTV) systems. Controllability is examined through trajectory linearization, where the LTV controllability Gramian provides a local, first-order measure of input influence along a nominal trajectory. Observability is analyzed through output linearization, where the LTV observability Gramian characterizes the local ability to reconstruct system states from output measurements. Koopman-based lifting is considered to extend the analysis to a higher-dimensional representation, and its limitations under multiple equilibria and basin-dependent behavior are discussed. The proposed framework is illustrated on a series RLC circuit. The learned neural ODE reproduces system trajectories and generalizes to unseen initial conditions. The computed Gramians are numerically full rank along the tested trajectories, indicating local controllability and observability of the linearized dynamics.
arXiv:2606.08432v1 Announce Type: new Abstract: On-policy distillation (OPD) has become a central post-training tool for large language models (LLMs), providing dense per-token teacher supervision along the student's own rollouts. In this work, we identify a common structural cause underlying OPD, which we call prefix failure. Under prefix failure, dense per-token supervision induces a bimodal teacher mixture and fragmented gradients that token-level loss truncation or reweighting fail to address. This observation motivates us to move beyond token-level loss interventions toward trajectory-level output corrections. We thus propose Trajectory-Refined Distillation (TRD), a trajectory-level correction method that revises the student's rollout under the teacher guidance while within on-policy support. By correcting problematic prefixes before distillation, TRD mitigates prefix failure at its source. Moreover, TRD improves the exploration by exposing the student to alternative valid derivations under teacher guidance, even when the original rolls are already correct. TRD can also be applied to on-policy self-distillation (OPSD), a parameter-sharing variant that uses the student model conditioned on privileged informations as the teacher. Across a wide range of benchmarks and base models at multiple scales, TRD consistently outperforms prior baselines, improving single-attempt accuracy and broadening reasoning coverage. Code is available at https://github.com/louieworth/trd
arXiv:2606.08433v1 Announce Type: new Abstract: This paper reads six engine-level measurements together -- 1.1 host attack surface, 1.2 information leakage, 1.3 defense-in-depth stackability, 1.4 public CVE history, 1.5 patch cadence, and 1.6 upstream fuzzing posture -- to describe how five AI-sandbox products isolate guest code from the host kernel. No single axis is a sufficient basis for a comparative judgement; the cross-axis reading is the load-bearing analysis. Three high-level findings: (1) engine classes (microVM, userspace kernel, OCI container) separate cleanly on every architectural axis, but products within a class do not; (2) product pin policy is the dominant operator-facing variable -- engine-side patch latency aggregates to ~0 days for coordinated disclosures, while downstream lag spans 0 days to 471+ days to "opaque" to infinity; (3) fuzzing investment splits into three tiers, and the strongest combination -- microVM x continuous public fuzzer -- is unoccupied in this set, leaving the "0 published CVEs x no upstream fuzzer x no academic study" intersection structurally unmeasured. We report per-axis orderings, per-product portraits, and a threat-model qualification matrix; no overall ranking is proposed. Companion repository (code, Apache-2.0): https://github.com/orbitalab/RnD-ai-sandboxes-sec-study-part-1. License: CC BY 4.0.
arXiv:2606.08439v1 Announce Type: new Abstract: Radio map (RM) estimation is a key enabler for environment-aware optimization in 6G wireless networks. In practice, RM construction increasingly relies on crowdsourced received signal strength (RSS) feedback that is inherently sparse and noisy. A further and often overlooked challenge is location drift, whereby privacy constraints and user mobility cause reported sampling coordinates to deviate from the true measurement locations. Unlike additive measurement noise, location drift perturbs the sensing operator itself, since each RSS sample effectively queries the underlying RM at an incorrect spatial coordinate. This operator uncertainty, compounded with sparse noisy sensing, renders the inverse problem severely ill-posed and limits conventional estimators that rely on analytically specified priors. This paper proposes RadioDiff-Inv2, a differentiable diffusion inversion framework that estimates RMs from sparse noisy measurements under location drift. A Gaussian resampling scheme is introduced to construct a differentiable, drift-aware measurement operator on grid-based maps, and the probability-flow ordinary differential equation (ODE) is exploited to cast the diffusion sampler as a deterministic, differentiable mapping from an initial noise code to the estimated RM. By optimizing the noise code via backpropagation against a drift-marginalized data-fidelity objective, RadioDiff-Inv2 produces reconstructions that are both prior-plausible and measurement-consistent without costly posterior sampling. Extensive experiments show that RadioDiff-Inv2 outperforms the best competing baseline by 4 to 14 dB in PSNR across varying sparsity and drift levels. The advantage is most pronounced in low-SNR regimes, where the learned diffusion prior maintains near-constant reconstruction fidelity while conventional methods degrade severely.
Influence of Aspect ratio in the Convection in Rotating Annulus In the Presence of Localized Heating
arXiv:2606.04151v2 Announce Type: replace Abstract: Two-dimensional (2D) axisymmetric simulations are conducted to investigate convection in a rotating cylindrical annulus with localized heating at the outer bottom edge and uniform cooling at the inner cylindrical wall. The resulting radial and vertical temperature gradients generate buoyancy-driven motion and produce a stratification pattern relevant to atmospheric circulation. The effects of aspect ratio (\(\Gamma\)), Rayleigh number (\(Ra = 2.4 \times 10^{7}\)--\(1.2 \times 10^{9}\)), and Taylor number (\(Ta = 1.6 \times 10^{7}\)--\(1.2 \times 10^{9}\)), including the non-rotating limit (\(Ta=0\)), are examined. Convection is largely confined to thin boundary layers, while the fluid interior remains diffusion dominated. Without rotation, the temperature field exhibits nearly horizontal isotherms. Rotation establishes quasi-hydrostatic and geostrophic balances that redistribute heat and promote deeper penetration of isotherms into the interior. Heat transfer, quantified by the Nusselt number (\(Nu\)), depends strongly on \(Ra\), \(Ta\), and \(\Gamma\). For moderate and high \(Ra\), \(Nu\) follows the scaling \(Nu \sim Ra^{1/4}\) and is only weakly influenced by rotation. At low \(Ra\) and high \(Ta\), rotational suppression of buoyancy reduces \(Nu\) significantly. Increasing \(\Gamma\) enhances heat transfer, although the growth rate diminishes for \(\Gamma > 1\). The relative thermal and Ekman boundary-layer thicknesses govern the sensitivity of heat transfer to rotation.
arXiv:2606.04430v3 Announce Type: replace Abstract: We experimentally demonstrate atom interferometry using the transverse phase profile of an optical mode. As proof-of-principle, we use the helical phase windings of Hypergeometric Gaussian beams for Ramsey interferometry with ensembles of ballistically-expanding cold Rb87 atoms, and we show that the interferometer can measure rotations induced by a motor with a sensitivity that scales linearly with orbital angular momentum and interferometer time. We characterize the thermal decoherence of the interferometer, deriving and experimentally confirming a closed-form expression for the spatially-varying interferometer visibility arising near the singularity of the helical phase winding, motivating the use of condensed atoms in ring-shaped traps.
arXiv:2606.04627v2 Announce Type: replace Abstract: Mobile agents are increasingly expected to operate everyday applications from screenshots and language goals, where reliable control requires reasoning over screen affordances, multi-step navigation, and future state changes. However, many agents externalize this computation as long textual chains of thought, which slows interaction, increases supervision cost, and complicates deployment. We introduce MIRAGE, a framework that learns continuous latent reasoning representations from visible textual reasoning traces. MIRAGE transfers explicit reasoning into compact hidden states, enabling the agent to reason internally without decoding long rationales. It also incorporates a generative world-model objective: latent reasoning vectors are aligned with future screenshots, encouraging the agent to anticipate upcoming interface states before acting. This turns hidden computation into both a compressed thought representation and a forward-looking model of environment dynamics. At inference time, MIRAGE reasons in continuous latent space, reducing token generation while improving execution efficiency. On AndroidWorld, MIRAGE matches explicit chain-of-thought supervised fine-tuning in the 4B ablation with a 3-5x lower decoded-token budget and improves a comparable instruction-tuned baseline by 10.2 points; on AndroidControl, it improves action grounding while generating over 75% fewer tokens.