arXiv:2603.06984v2 Announce Type: replace-cross Abstract: Many regulatory and analytic problems require that a prohibited variable influence a decision only through a designated allowable channel -- a conditional-independence requirement that arises in path-specific fairness, the handling of classified information, and the regulation of trading on non-public information, among other settings. Such requirements may be enforced either stratum-by-stratum or, more commonly (and more efficiently), through a single averaged constraint on the conditional effect. We study the resulting enforcement problem from two perspectives. From the regulator's side, we formulate causal masking as a linear program and show that averaged-constraint optimization almost surely produces policies that violate the stratum-wise requirement while satisfying the averaged one exactly. The gains from masking grow with confounding and outcome heterogeneity, and detection requires precisely the conditional-independence tests that average constraints aim to avoid. From the optimizer's side, the same construction shows that masked policies recover most of the reward of unconstrained exploitation while being far harder to detect, making them attractive in any setting where the basis of decisions is itself sensitive. Together, these results argue that regulating direct dependence through averaged statistics on observed decisions is structurally limited, and that meaningful enforcement must operate at the level of the decision rule itself.
Science Journals
arXiv:2506.05442v2 Announce Type: replace Abstract: Vision-Language Models (VLMs) offer a promising approach to end-to-end autonomous driving due to their human-like reasoning capabilities. However, troublesome gaps remains between current VLMs and real-world autonomous driving applications. One major limitation is that existing datasets with loosely formatted language descriptions are not machine-friendly and may introduce redundancy. Additionally, high computational cost and massive scale of VLMs hinder the inference speed and real-world deployment. To bridge the gap, this paper introduces a structured and concise benchmark dataset, NuScenes-S, which is derived from the NuScenes dataset and contains machine-friendly structured representations. Moreover, we present FastDrive, a compact VLM baseline with 0.9B parameters. In contrast to existing VLMs with over 7B parameters and unstructured language processing(e.g., LLaVA-1.5), FastDrive understands structured and concise descriptions and generates machine-friendly driving decisions with high efficiency. Extensive experiments show that FastDrive achieves competitive performance on structured dataset, with approximately 20% accuracy improvement on decision-making tasks, while surpassing massive parameter baseline in inference speed with over 10x speedup. Additionally, ablation studies further focus on the impact of scene annotations (e.g., weather, time of day) on decision-making tasks, demonstrating their importance on decision-making tasks in autonomous driving.
arXiv:2509.21092v2 Announce Type: replace Abstract: This study investigates how the majority group influences individual judgment formation and expression in anonymous, spontaneous online conversations. Drawing on theories of social conformity and anti-conformity, we analyze everyday dilemmas discussed on social media. First, using digital traces to operationalize judgments, we measure the conversations' disagreement and apply Bayesian regression to capture shifts of judgments formation before and after the group's exposure. Then we analyze changes in judgment expression with a linguistic analysis of the motivations associated with each judgment. Results show anti-conformity behaviors: individuals preserve the majority's positive or negative orientation of judgments but diverge from its stance, with persuasive language increasing post-disclosure. Our findings highlight how online environments reshape social influence compared to offline contexts.
arXiv:2605.18696v1 Announce Type: new Abstract: Tabular foundation models (TFMs) now match or beat tuned gradient-boosted trees on a growing fraction of tabular tasks, but no single TFM wins on every dataset. Ensembling is the go to fix here, and it works less well than expected. Six modern TFMs form a near-redundant pool: their mean pairwise Q-statistic is $0.961$, close enough to $1$ that any convex combination is bounded above. We benchmark six ensemble strategies over six TFMs on 153 OpenML classification tasks. The best ensemble, two-level cascade stacking, buys $+0.18\%$ accuracy over the strongest single TFM at $253\times$ the compute. A Friedman and Nemenyi analysis places three ensembles and the best base TFM in a single equivalence group; three other ensembles are significantly \emph{worse} than the best base. Stacking with a logistic-regression meta-learner is the most striking case: competitive accuracy and ROC-AUC, the worst log-loss rank among the ensembles. The meta-learner improves accuracy by sharpening class boundaries, which destroys calibration. We recommend greedy selection as the practical default.
arXiv:2605.18031v1 Announce Type: cross Abstract: We propose a quantum sidecar architecture family for future hybrid AI training and inference. The central idea is not to store an entire Transformer in a small quantum memory, nor to claim one-shot collapse into a fully trained model or an optimal answer. Instead, we identify two physically distinct operating modes for quantum co-processors attached to classical large-model pipelines. The first is a stateful protected-register mode, in which a protected register stores a reusable quantum resource while an ancilla or temporary register performs QND-style readout. The second is a stateless reset-and-reprepare mode, in which each query prepares a task-conditioned quantum circuit, evolves over bounded training or inference control variables, measures candidate signals, resets the qubits, and repeats. We simulate the stateful mode using 2/4/6/8 protected-qubit density-matrix QND-style parity readout with one ancilla and a Qiskit cross-check. For the stateless mode, we include both an abstract candidate-update sampler and a circuit-level QAOA-style statevector sampler over structured candidate landscapes, followed by reset-overhead sensitivity analysis. The resulting framework positions quantum sidecars as bounded signal generators for optimizer-side sampling, adapter or expert selection, retrieval, routing, and reasoning-path proposal. As a speculative outlook, we introduce quantum weight-state sidecars: restricted quantum representations over model-control variables, not direct encodings of complete classical weight tensors.
arXiv:2603.08462v2 Announce Type: replace Abstract: \ac{CoT} prompting improves LLM accuracy on complex tasks but often increases token usage and inference cost. Existing ``Budget Forcing'' methods reduce cost via fine-tuning with heuristic length penalties, suppressing both essential reasoning and redundant filler. We recast efficient reasoning as a lossy compression problem under the \ac{IB} principle, and identify a key theoretical gap when applying naive \ac{IB} to transformers: attention violates the Markov property between prompt, reasoning trace, and response. To resolve this issue, we model \ac{CoT} generation under the \ac{CIB} principle, where the reasoning trace $Z$ acts as a computational bridge that contains only the information about the response $Y$ that is not directly accessible from the prompt $X$. This yields a general Reinforcement Learning objective: maximize task reward while compressing completions under a prior over reasoning traces, subsuming common heuristics (e.g., length penalties) as special cases (e.g., uniform priors). In contrast to naive token-counting approaches, we introduce a semantic prior that measures token cost by surprisal under a language model. Crucially, the prior is queried only for token-level log-probabilities, adding negligible overhead to the training loop. Empirically, our \ac{CIB} objective prunes reasoning redundancy while preserving fluency and logic, improving accuracy at moderate compression and enabling aggressive compression with minimal accuracy drop. These gains generalize across model families and task domains, confirming \ac{CIB} as a domain-agnostic CoT compression framework.
arXiv:2603.20562v3 Announce Type: replace Abstract: Large language models (LLMs) are now widely used as judges, yet their decisions can change under presentation choices that should be irrelevant. We study one such source of instability: candidate-order sensitivity in listwise factuality evaluation, where several answers can look similarly polished while differing substantially in hallucination risk. We introduce PCFJudge, an inference-time method that reruns the same factuality-first listwise prompt over multiple orderings of the same candidate set and aggregates the resulting scores, ranks, and uncertainty signals into a single consensus decision. On RewardBench 2 Factuality, the final seven-permutation aggregate (K=7) improves top-1 selection accuracy from 86.00% to 91.33% with GPT-5.4 and from 86.33% to 89.67% with Claude Sonnet 4.6. These results suggest that candidate order can be a meaningful source of factuality-judging error and that marginalizing over this nuisance variation can improve the reliability of LLM evaluation.
arXiv:2512.22230v2 Announce Type: replace Abstract: Understanding the intricate interplay between soot dynamics and chemical reactions within catalytic diesel particulate filters (CDPF) is crucial for enhancing both filtration efficiency and regeneration performance. In this paper, we establish a unified pore-scale multiphysics model based on the Eulerian-Lagrangian framework to comprehensively resolve the transport, deposition, and oxidation of soot. Distinguishing itself from conventional empirical correlations and stochastic-based approximations, the system models soot deposition through fundamental physical principles, integrating elastic deformation and surface adhesion mechanics at the particle-wall interface. Simultaneously, it incorporates a robust oxidation model that accounts for the competitive kinetics of both $\textrm{O}_2$ and $\textrm{NO}_2$ pathways, enabling comprehensive coverage of all CDPF operating regimes. Validated against three classical benchmark cases, the model demonstrates superior accuracy in capturing interfacial mass transfer and particle-wall interactions. Simulation under a typical CDPF low-temperature operating condition emphasizes the pivotal role of $\textrm{NO}_2$ and catalyst in promoting regeneration and reveals complex synergistic and competitive effects between distinct reaction pathways. Notably, the reaction rate of direct $\textrm{O}_2$ pathway is accelerated by a factor of 87 in the presence of the catalyst. For ultra-fine soot particles ($50~\mathrm{nm}$), the Brownian motion and thermophoretic forces directly dictate the deposition efficiency. Their strong thermal sensitivity also underscores the necessity for an integrated soot transport-deposition-oxidation framework. To support further research, the model implementation can be accessed at https://github.com/zhangyujing2001.
arXiv:2605.18729v1 Announce Type: new Abstract: The ability to navigate and interact with complex environments is central to real-world embodied agents, yet navigation in unseen environments remains challenging due to "experiential amnesia," where existing trajectory-driven or reactive policies fail to synthesize generalizable strategies from past interactions. We propose Robo-Cortex, a self-evolving framework that enables robots to autonomously induce navigation heuristics and refine cognitive strategies through a continuous reflection-adaptation loop. By abstracting success patterns and failure pitfalls into natural-language heuristics, Robo-Cortex enables a transition from passive execution to active strategy evolution. Our core innovation is an Autonomous Knowledge Induction (AKI) mechanism that distills multimodal trajectories into a structured Navigation Heuristic Library for knowledge generalization. The architecture further incorporates a Dual-Grain Cognitive Memory system, comprising a Short-term Reflective Memory (SRM) for real-time local progress analysis, and a Long-term Principle Memory (LPM) that abstracts past trajectories into reusable guiding and cautionary principles. To ensure robust decision-making, we introduce a multimodal Imagine-then-Verify loop, where a world model simulates potential outcomes and a VLM-based evaluator validates action plans. Extensive evaluations on IGNav, AR, and AEQA show that Robo-Cortex consistently outperforms strong baselines in both task success and exploration efficiency, with gains of up to +4.16% SPL over the strongest prior method and up to +15.30% SPL under heuristic transfer to unseen environments. Preliminary real-world robotic experiments further support the effectiveness of Robo-Cortex in physical settings.
arXiv:2605.18707v1 Announce Type: new Abstract: Population protocols are a model of distributed computing where $n$ agents, each a simple finite-state machine, interact in pairs to solve a common task against a (adversarial) interaction scheduler. This model was intensively studied in recent years; in particular, the problem of relative majority received much attention: Each agent starts with an input opinion (or color) out of $k$ possibilities, and the goal is for each agent to eventually output the color with the largest support in the population. Before our work, the state complexity (the minimum number of states required per agent) was only known to be between $\Omega(k^2)$ and $O(k^{7})$. Our main contribution is a population protocol that solves the relative majority problem with $k^3$ states. We achieve this result with a new protocol called CIRCLES. While prior approaches in the literature relied on duels of agents to find the majority color -- an approach that proved effective for the case with two colors -- CIRCLES partitions the agents into circular linked lists of decreasing sizes, with the property that no two agents with the same initial color lie in the same circle. We show that CIRCLES always correctly computes the desired structure against the most adversarial of schedulers (weakly fair). We then show that a trivial extension of CIRCLES solves the relative majority problem. We extend our protocol to handle various tie-breaking mechanisms or to support the case where the agents do not share a prior ordering of the colors. Finally, we show that a modification of CIRCLES solves the ranking problem with $2 \cdot k^4$ states, where each agent must output the rank of its initial color in the population.
arXiv:2605.18681v1 Announce Type: new Abstract: Explainable AI (XAI) techniques are increasingly important for the validation and responsible use of modern deep learning models, but are difficult to evaluate due to the lack of good ground-truth to compare against. We propose a framework that serves as a quantifiable metric for the quality of XAI methods, based on continuous input perturbation. Our metric formally considers the sufficiency and necessity of the attributed information to the model's decision-making, and we illustrate a range of cases where it aligns better with human intuitions of explanation quality than do existing metrics. To exploit the properties of this metric, we also propose a novel XAI method, considering the case where we fine-tune a model using a differentiable approximation of the metric as a supervision signal. The result is an adapter module that can be trained on top of any black-box model to output causal explanations of the model's decision process, without degrading model performance. We show that the explanations generated by this method outperform those of competing XAI techniques according to a number of quantifiable metrics.
arXiv:2605.18563v1 Announce Type: new Abstract: A key question in psycholinguistics is how inferences about the meaning of linguistic input unfold incrementally a comprehender's mind. In this work, we study reading dynamics for ``noisy-channel garden-path'' sentences, which temporarily appear well-formed but feature late-appearing violations of expectation that can be resolved not by inferring an alternative syntactic structure, but by inferring the presence of an error. We find evidence for targeted regressions -- eye movements towards regions that are promising loci of possible errors in light of later-arriving information, showing patterns consistent with the posterior inferences of a model of noisy-channel processing with reanalysis. We discuss the implications of these findings for theories of noisy-channel language comprehension and information-theoretic explanations of reading dynamics.
arXiv:2605.17597v1 Announce Type: new Abstract: Distributed controller synthesis offers scalable and privacy-preserving control design, but typical state-of-the-art approaches either assume white-box models or resort to centralized synthesis. In this paper, we combine partially known model knowledge and an input-state dataset within a distributed gray-box scheme to design \(\mathcal{H}_2\) controllers. Our method can handle unknown dynamics and offers scalable synthesis. Each agent communicates with a set of neighbors determined by the physical coupling topology of the system such that we can apply the Alternating Direction Method of Multipliers (ADMM) to solve the problem iteratively in a fully distributed fashion (i.e., without a central server). The effectiveness and flexibility of the proposed approach is demonstrated in simulations of the IEEE 39-bus power system test case.
arXiv:2605.18672v1 Announce Type: new Abstract: This position paper argues that enforcing LLM agent safety within a single abstraction layer is not merely suboptimal but categorically insufficient for deployed LLM agents -- a structural consequence of how agent execution works, not a contingent limitation of current systems. The three dimensions that jointly constitute safe operation -- semantic intent and policy compliance, environmental validity, and dynamical feasibility -- each depend on a strictly distinct set of information that becomes available at different stages of execution. No single guardrail can certify all three. We argue that the community must respond with a contract-based architecture in which each safety dimension is enforced by an independently certified layer whose probabilistic guarantee satisfies the next layer's assumption. We sketch such an architecture and derive the compositional system-level safety bounds it admits via the chain rule of probability. Three open problems stand between this and a deployable standard: bound estimation from non-i.i.d.\ traces, graceful degradation of contracts under deployment drift, and extension to multi-agent settings -- the most important unfinished business in LLM agent runtime assurance.
arXiv:2605.18668v1 Announce Type: new Abstract: Interactions of relativistically intense laser pulses with structured targets have long been explored for controlling energy absorption and particle acceleration. However, at upcoming multi-petawatt laser facilities, the survivability of such nanostructures under realistic temporal contrast conditions remains a key concern. We report an experimental and simulation study of nanowire targets irradiated by the ELI-NP 1-PW laser without a plasma mirror. At the built in, readily available contrast of $10^{-10}$, the nanowires survive the laser pre-pulse and produce a robust enhancement in relativistic electron flux, energy, and directional emission compared to flat targets indicating that at better contrasts they can show similar enhancement at the 10 PW level. These results establish nanowire targets as resilient and reliable tools for relativistic electron manipulation at state of the art facilities.
arXiv:2605.18667v1 Announce Type: new Abstract: Earth embedding models transform Earth observation data into embeddings uniquely tied to locations on the Earth's surface. These models are typically evaluated in isolation, comparing the downstream task performance across different Earth embeddings. However, spatially aligned embeddings can naturally be fused, providing richer information per location, a capability that isolated evaluations fail to capture. We therefore propose assessing Earth embeddings by their complementarity: the performance gain of fused embeddings over the best single-model baseline. To operationalise this, we introduce an embedding complementarity index applicable to any embedding and task, and evaluate four Earth embedding models (AlphaEarth, Tessera, GeoCLIP, SatCLIP) in isolation, in all pairs, and jointly across six downstream tasks. Fused embeddings outperform the best single model in four out of six tasks, confirming that single-embedding evaluations often underestimate Earth embedding capabilities. Complementarity proves both task- and location-dependent. Further, for a land cover regression task, we find that complementarity is partially determined by the spatial scale of land cover classes. Complementarity reframes Earth embeddings: the greatest future gains may come not from any single Earth embedding model, but from combinations that are better together.
arXiv:2510.21523v2 Announce Type: replace Abstract: Sequential generative models conditioned on uncertain rewards are central to AI-driven scientific discovery, yet the epistemic uncertainty they inherit from imperfect reward estimates remains unquantified. We propagate this uncertainty through generative flow networks (GFlowNets) by fitting polynomial chaos expansions (PCEs) to small ensembles of trained models. The PCE coefficients yield analytical Sobol sensitivity indices, providing the first interpretable decomposition of which reward components drive which generative decisions, a capability unavailable from deep ensembles, Bayesian neural networks, or Monte Carlo dropout. Convergence guarantees are established theoretically and four of five are formally verified in the Lean 4 proof assistant. Across three real-world tasks the framework reveals actionable structure invisible to ensembles alone. On the Doyle-Dreher Buchwald-Hartwig dataset catalyst selection is robust ($D_{\mathrm{catalyst}}\approx 71$) while additive selection is fragile ($D_{\mathrm{additive}}\approx 179$, $2.5\times$ higher). In fragment-based molecular design the linker position is the most sensitive ($D_{\mathrm{linker}}\approx 28$) while decoration positions are the most robust ($D\approx 14$-$18$), reversing the conventional scaffold-robust / decoration-fragile assumption. On the Sachs protein signalling network, MAPK-cascade edges and PKA/PKC hub edges separate into distinct sensitivity regimes, providing a targeted map for perturbation experiments. Calibration coverage at the 95% level reaches 0.97-1.00 across the dominant steps, and the surrogate evaluates 10{,}000 policy samples in milliseconds - $10^{3}$-$10^{4}\times$ faster than exhaustive retraining.
arXiv:2605.17687v1 Announce Type: new Abstract: We introduce a stochastic framework based on complex absorbing potentials (CAPs) to investigate exciton transport in large molecular aggregates. Within this approach, CAPs act as non-Hermitian reservoirs and sinks that enable effective measurement of transport efficiency. We apply this framework to cyanine dye aggregates and examine how vacancy defects and system size influence exciton dynamics in two-dimensional sheets and quasi-one-dimensional tubes. We also introduce a CAPs-based classification scheme that links molecular packing in 2D aggregates to transport behavior. Our results demonstrate how aggregate topology and structural disorder govern exciton dynamics and provide guidance for designing materials with enhanced energy transport.
arXiv:2602.17159v5 Announce Type: replace Abstract: The paper presents a geometric duality between the spherical squared-Hellinger distance and a hyperbolic isometric invariant of the Poincare disc under the action of the general Mobius group. Motivated by the geometric connection, we propose the usage of the L2-embedded hyperbolic isometric invariant as an alternative way to quantify divergence between Gaussian measures as a contribution to information theory.
arXiv:2605.18642v1 Announce Type: new Abstract: We present a controlled benchmark evaluating three LLMs -- Claude Sonnet 4.5, Gemini 2.5 Pro, and GPT-3.5 Turbo -- across four prompt formats (from concise narrative to structured JSON with explicit iteration trace) on Gauss--Seidel AC power flow computation for a three-bus system. Against 50 test cases with reference solutions computed numerically, Gemini 2.5 Pro with the simplest narrative prompt achieves the lowest mean absolute error (MAE = 0.257 MW/MVar, 54\% of cases within 5\% relative error), while the same model with a JSON-structured prompt raises MAE to 0.789 -- a 3.1$\times$ increase. Adding a worked example degrades accuracy for Gemini but provides a marginal gain for Claude. GPT-3.5 Turbo fails on at least 90\% of cases under all prompt formats. An independent 100-case replication with related prompt-format families confirms the qualitative ordering (Gemini $>$ Claude $>$ GPT-3.5): the best 100-case configuration (Gemini with explicit iteration trace) achieves MAE = 0.402 and 53\% within 5\%, while Claude Sonnet 4.5's near-flat accuracy profile ($\approx$38\% within 5\% across formats) and GPT-3.5's near total ineffectiveness (92--97\% above 20\% error) both replicate. In neither evaluation does any configuration achieve sufficient reliability for use as a direct numerical solver. These findings offer a diagnostic baseline for practitioners and researchers evaluating LLMs for smart-grid decision-support assistance.
arXiv:2605.15641v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used as general planners in embodied intelligence, enabling high level coordination and low level task planning for both single robot and multi-robot collaboration. This increasing reliance on embodied LLM planners also raises critical security concerns, since misaligned or manipulated instructions can be translated into physical actions. Prior work has studied such threats in single robot settings, while security risks in LLM controlled multi-robot collaboration, especially those propagated through inter robot communication, remain largely unexplored. To bridge this gap, we propose a novel attack paradigm for multi-robot system in which the adversary interacts with only a single entry robot. The compromised robot then propagates malicious intent through peer communication, leading to coordinated unsafe actions across the system. Our evaluation, covering high risk dimensions of dereliction of duty, privacy compromise, and public safety hazards, reveals a persistent safety alignment gap in multi-robot planners. We quantify this process with three metrics, obedience, infectiousness, and stealthiness. Experiments demonstrate both persistent attacker control and rapid propagation: obedience reaches 1.00 in the strongest cases, and infectiousness rises to 0.90. Notably, the attack is highly efficient, requiring as few as 3.0 rounds to compromise all the robots while maintaining a stealthiness score of 0.81. Such risks are amplified when robots must resolve trade offs in critical situations, such as emergencies or conflicts of rights, because the coordination mechanism can unintentionally allow adversarial instructions to override safety requirements. The code is available at https://github.com/TheFatInsect/InfectBot.
arXiv:2605.17356v1 Announce Type: new Abstract: Existing works typically focus on presentation generation under isolated input settings, whereas real-world use cases span diverse scenarios, including vague user prompts, long documents, multimodal materials, and multiple heterogeneous sources. Moreover, current evaluations are often insufficiently scenario-specific. They mainly rely on generic presentation-quality criteria, such as visual appeal, layout quality, and overall coherence, but fail to assess the core capabilities required by different input settings, including grounded compression, visual-text alignment, and cross-source synthesis. Consequently, the field lacks a unified benchmark and a scenario-aware evaluation framework for faithfully diagnosing presentation-generation systems across diverse real-world settings. We present UniPPTBench, a unified benchmark for presentation generation across four representative input settings: vague-prompt, long-document, multimodal-document, and multi-source generation. We further introduce UniPPTEval, a scenario-aware evaluation protocol that combines shared metrics for cross-setting comparison with scenario-specific metrics tailored to the core requirements of each setting. We also provide transparent reference baselines to support reproducible comparison. Experiments on UniPPTBench reveal substantial performance variation across settings and recurring failure modes in content grounding, multimodal integration, and cross-source synthesis. In particular, strong performance on generic presentation-quality metrics does not necessarily imply strong task fulfillment in grounded scenarios. Together, UniPPTBench and UniPPTEval provide a faithful and diagnostic foundation for evaluating presentation generation across diverse real-world scenarios. Code and data will be publicly available.
arXiv:2605.18078v1 Announce Type: new Abstract: Multi-agent policy-gradient methods have been shown to converge locally near stable Nash equilibria. Local convergence, however, does not determine which equilibrium is reached. We study this question through basin-entry probability with respect to a target set of equilibria selected by an external criterion, such as payoff dominance. For finite-unroll Meta-MAPG, we show that the update decomposes into ordinary policy gradient plus own-learning and peer-learning corrections, with controlled sampling noise and finite-unroll bias. We identify the peer-learning correction as the main equilibrium-selection mechanism: under a local alignment condition, the probability of entering the certified attraction region of the target stable-Nash set increases, relative to ordinary policy gradient. Because persistent correction may shift zero-update points of the original game, annealing the correction after entering the basin recovers ordinary policy-gradient dynamics and inherits local stable-Nash convergence guarantees. Experiments in Stag Hunt, iterated Prisoner's Dilemma, and preliminary neural-policy coordination environments support this basin-entry view, showing increased entry into cooperative basins under peer-aware updates.
arXiv:2605.17368v1 Announce Type: new Abstract: Anatomical structure labels for chest radiographs are essential for medical image segmentation and a broad range of downstream diagnostic tasks. However, annotating anatomy directly on 2D chest radiographs is labor-intensive and intrinsically ambiguous, as 3D anatomical structures are projected onto a single 2D plane where boundaries may overlap, be occluded, or appear only partially visible. Consequently, existing anatomy-labeled chest radiograph datasets remain limited in scale, anatomy coverage, and label reliability. To address these limitations, we introduce RadGenome-Anatomy, the largest anatomy-labeled chest radiograph dataset, containing over 10 million segmentation masks across 210 anatomical structures in 25,692 studies. It is constructed by projecting large-scale 3D anatomical masks from CT volumes into 2D radiographic space through canonical radiographic geometry. This shifts annotation from directly tracing uncertain 2D boundaries to defining anatomy in volumetric space, where structures that overlap or become partially invisible in radiographs remain spatially separable. As a result, each 2D mask represents the physically grounded projected footprint of a volumetrically defined structure. The scale and broad anatomical coverage of RadGenome-Anatomy, including structures that are overlapping, partially visible, or difficult to delineate directly, enable research on geometric measurements as explicit evidence for chest radiograph interpretation. We demonstrate this by training XAnatomy to predict structure-specific masks and derive clinically relevant measurements, achieving diagnostic accuracies of 96.4%, 95.6%, and 89.2% for cardiomegaly, kyphosis, and scoliosis, respectively.
arXiv:2605.18700v1 Announce Type: new Abstract: Prior work on fine-grained image recognition (FGIR) has established the importance of the backbone selection, but has neglected the accuracy-vs-cost trade-offs under different training and evaluation settings. In this work we conduct a large-scale study with over 2000 experiments across 6 training and evaluation settings, 9 pretrained backbones, and 17 datasets. Preliminary observations on the effectiveness of data augmentation for fine-grained training motivate us to extend Counterfactual Attention Learning (CAL), a state-of-the-art method based on data-aware cropping and masking augmentations, with cross-image discriminative region mixing augmentation. We also propose an efficient evaluation-only variant that maintains competitive accuracy while reducing inference costs by forfeiting the forward pass on discriminative crops that is normally used by CAL and similar FGIR methods. Our results show that data-aware augmentations during training only can enable a model to achieve excellent accuracy even without crops, significantly reducing inference costs. To support future research we share our code and checkpoints at: \url{https://github.com/arkel23/FGIR-Backbones}