arXiv:2605.15984v1 Announce Type: new Abstract: Toxic speech detection has become a crucial challenge in maintaining safe online communication environments. However, existing approaches to toxic speech detection often neglect the contribution of paralinguistic cues, such as emotion, intonation, and speech rate, which are key to detecting speech toxicity. Moreover, current toxic speech datasets are predominantly text-based, limiting the development of models that can capture paralinguistic cues.To address these challenges, we present ToxiAlert-Bench, a large-scale audio dataset comprising over 30,000 audio clips annotated with seven major toxic categories and twenty fine-grained toxic labels. Uniquely, our dataset annotates toxicity sources -- distinguishing between textual content and paralinguistic origins -- for comprehensive toxic speech analysis.Furthermore, we propose a dual-head neural network with a multi-stage training strategy tailored for toxic speech detection. This architecture features two task-specific classification headers: one for identifying the source of sensitivity (textual or paralinguistic), and the other for categorizing the specific toxic type. The training process involves independent head training followed by joint fine-tuning to reduce task interference. To mitigate data class imbalance, we incorporate class-balanced sampling and weighted loss functions.Our experimental results show that leveraging paralinguistic features significantly improves detection performance. Our method consistently outperforms existing baselines across multiple evaluation metrics, with a 21.1% relative improvement in Macro-F1 score and a 13.0% relative gain in accuracy over the strongest baseline, highlighting its enhanced effectiveness and practical applicability.
Science Journals
arXiv:2512.19634v3 Announce Type: replace-cross Abstract: We investigate the influence of magnetic order on the proximity-induced superconducting state in the Mn layers of a Mn-Nb(110) heterostructure by using a first-principles method. For this study, we use the recently developed Bogoliubov-de Gennes (BdG) solver for superconducting heterostructures [Csire et al., Phys. Rev. B 97, 024514 (2018)] within the first-principles calculations based on multiple scattering theory and the screened Korringa-Kohn-Rostoker (SKKR) Green's function method. In our calculations, we first study the normal-state density of states (DOS) in the single- and double-Mn-layer heterostructures, and calculate the induced magnetic moments in the Nb layers. Next, we compute the momentum-resolved spectral functions in the superconducting state for the heterostructure with a single Mn layer, and find bands crossing the Fermi level within the superconducting (SC) gap. We also study the SC state DOS in the single- and double-Mn-layer heterostructures and compare some of our results with experimental findings, revealing secondary gaps, plateau-like regions, and central V-shaped in-gap states within the bulk SC Nb gap that are magnetic-order-dependent. Finally, we compute the singlet and internally antisymmetric triplet (IAT) order parameters for each layer for both heterostructures, and find an order of magnitude difference in the induced singlet part of the SC order parameter in the Mn layer/s between the FM and AFM cases in favor of the AFM pairing with the maximum still being only 4.44% of the bulk Nb singlet order parameter value. We also find a negligible induced triplet part, yet comparable to the induced singlet values, indicating some singlet-triplet mixing in the Mn layer/s.
arXiv:2605.15350v1 Announce Type: cross Abstract: Stochastic compositional optimization minimizes objectives of the form $\min_{\bm{x} \in \mathcal{X}} F(\bm{f}(\bm{x}), \bm{x})$, where $\bm{f}$ is accessible only through noisy stochastic queries. Existing methods for this problem assume that the outer function $F$ is continuously differentiable, which excludes many practically important applications such as robust max-of-losses, Conditional Value-at-Risk, and norm regularizers. We propose the Hybrid Momentum Stochastic Frank--Wolfe algorithm, which drops the smoothness assumption on $F$. By combining a momentum-based Jacobian tracker with a Taylor-corrected function tracker, the algorithm feeds an entire stochastic linearization -- rather than a single gradient -- into a generalized linear minimization oracle. We establish an $\mathcal{O}(K^{-1/4})$ convergence rate in the generalized Frank--Wolfe gap for non-convex objectives with $L_F$-Lipschitz outer functions, matching the optimal complexity for projection-free single-sample stochastic methods under expected smoothness. The analysis extends to heavy-tailed noise oracles with bounded $r$-th moments for $r \in (1, 2]$ and recovers the deterministic rates of Vladarean et al (2023) as the noise vanishes.
arXiv:2605.13073v2 Announce Type: replace Abstract: In-the-wild 3D Gaussian Splatting remains challenging due to transient distractors and illumination-induced cross-view appearance inconsistencies. Existing methods mainly rely on image-level masking to suppress unreliable supervision, but masking alone cannot fully eliminate residual occlusions or resolve illumination-induced inconsistencies, both of which can introduce conflicting cross-view gradients. These unresolved conflicts may destabilize Gaussian optimization and lead to visible reconstruction artifacts. We propose a conflict-aware 3DGS framework that addresses this problem from both image-space supervision and gradient-level optimization. Semantic Consistency-Guided Masking learns pixel-wise consistency scores to adaptively refine prior masks and suppress unreliable supervision before gradient formation. A dual-view Conflict-Aware Gradient Harmonization strategy further reconciles view-specific gradients by mutually rotating them into an orthogonal configuration, reducing negative directional interference across views. We also introduce conflict-aware densification and pruning to stabilize Gaussian growth and remove persistently conflicting primitives. Extensive experiments on standard in-the-wild benchmarks demonstrate that our method achieves state-of-the-art rendering quality under complex transient distractors and cross-view inconsistencies.
arXiv:2512.21651v3 Announce Type: replace Abstract: Large Language Models (LLMs) deliver strong performance across a wide range of NLP tasks, but their massive sizes hinder deployment on resource-constrained devices. To reduce their computational and memory burden, various compression techniques have been proposed, including quantization, pruning, and knowledge distillation. Among these, post-training quantization (PTQ) is widely adopted for its efficiency, as it requires no retraining and only a small dataset for calibration, enabling low-cost deployment. Recent advances for post-training quantization have demonstrated that even near 4-bit methods can maintain most of the original model performance. However, 1-bit quantization remains particularly challenging. A common strategy in 1-bit quantization is to determine binary weights by matching full-precision parameters, following a weight-driven criterion. However, this objective is not directly aligned with the quantized model's objective, which is to preserve the model's output behavior under the impact of quantization. A natural alternative is to adopt output-driven criteria that minimize discrepancies in model outputs using calibration data. Surprisingly, naive output-driven approaches often perform even worse in the 1-bit regime. In this paper, we show that this failure arises from two fundamental issues: error accumulation across layers and, more critically, \emph{anisotropic distortion} of the representation space. Based on these insights, we propose a novel PTQ method for 1-bit LLMs that explicitly addresses these issues while maintaining computational efficiency. Extensive experiments demonstrate that our approach consistently outperforms existing 1-bit PTQ methods.
arXiv:2605.16045v1 Announce Type: new Abstract: Memory systems often organize user-agent interactions as retrievable external memory and are crucial for long-running agents by overcoming the limited context windows of LLMs. However, existing memory systems invoke LLMs to process every incoming interaction for memory extraction, and such an eager memory consolidation scheme leads to substantial token consumption. To tackle this problem, we propose RecMem by rethinking when memory consolidation should be conducted. RecMem stores incoming interactions in a subconscious memory layer and encode them using lightweight embedding models for retrieval. LLMs are only invoked to extract episodic and semantic memory when sustained recurrence are observed for semantically similar interactions. Such recurrence-based consolidation works because these interactions correspond to a semantic cluster with rich information and thus are worth extraction and summarization. To improve accuracy, RecMem also incorporates a semantic refinement mechanism that recovers the fine-grained facts omitted by memory extraction. Experiments show that RecMem reduces the memory construction token cost of three SOTA memory systems by up to 87% while exceeding their accuracy.
arXiv:2601.18349v2 Announce Type: replace Abstract: The Orbit Problem asks whether the orbit of a point under a matrix reaches a given target set. When the target is a single point, the problem was shown to be decidable in polynomial time by Kannan and Lipton. This decidability result was later extended by Chonev et al. to targets of dimension 3 (in arbitrary ambient dimension), but decidability remains open for subspaces of dimension 4. At the other extreme, the special case of the Orbit Problem in which the target set is a hyperplane of codimension 1 is equivalent to the Skolem Problem for linear recurrence sequences, whose decidability has been open for many decades. In this paper, we show that the Orbit Problem is decidable if the target subspace has dimension logarithmic in the dimension of the orbit. Over rationals, we moreover obtain a complexity bound NP^RP in this case. On the other hand, we show that the version of the Orbit Problem where the dimension of the target subspace is linear in the dimension of the orbit is as hard as the Skolem Problem.
arXiv:2505.03307v3 Announce Type: replace-cross Abstract: Simulating Clifford and near-Clifford circuits using the extended stabilizer formalism has become increasingly popular, particularly in quantum error correction. Compared to the state-vector approach, the extended stabilizer formalism can solve the same problems with fewer computational resources, as it operates on stabilizers rather than full state vectors. Most existing studies on near-Clifford circuits focus on balancing the trade-off between the number of ancilla qubits and simulation accuracy, often overlooking performance considerations. Furthermore, in the presence of high-rank stabilizers, performance is limited by the sequential property of the stabilizer formalism. In this work, we introduce a parallelized version of the extended stabilizer formalism, enabling efficient execution on multi-core devices such as GPU. Experimental results demonstrate that, in certain scenarios, our Python-based implementation outperforms state-of-the-art simulators such as Qiskit and Pennylane.
arXiv:2601.20477v4 Announce Type: replace Abstract: We study the training dynamics of neural classifiers through the lens of binary hypothesis testing. We re-formalize classification as a collection of binary tests between class-conditional distributions induced by learned representations and show empirically that, along training trajectories, well-generalizing networks progressively approach Neyman-Pearson optimal decision rules, as measured by monotonic growth in the KL divergence retained by learned representations. We provide sufficient conditions for exact optimality, discuss its implications for training regularization, and define an informational plane, (so-called Evidence-Error plane) where convergence can be assessed methodically across network architecture.
arXiv:2601.03707v2 Announce Type: replace Abstract: Existing UAV vision-and-language navigation (VLN) benchmarks rarely provide realistic aerial scenes, natural process-level instructions, and sufficient scale simultaneously, making it difficult to systematically train and evaluate UAV VLN agents under realistic settings. To address this, we propose \textbf{AirNav}, a large-scale benchmark built on real urban aerial data, comprising 137K navigation samples with natural and diverse instructions generated via a human--LLM collaborative pipeline with 10 user personas. We conduct a systematic evaluation of representative approaches on AirNav, ranging from traditional models to multimodal large language models (MLLMs), under unified metrics with open-source implementations. We further propose \textbf{AirVLN-R1}, trained via supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT), achieving state-of-the-art performance with a 51.82\% success rate on the test-unseen split. Real-world experiments on a physical UAV platform provide preliminary evidence of sim-to-real transferability, and our dataset and code are publicly available.
arXiv:2605.15983v1 Announce Type: new Abstract: We formulate the Resource-Constrained Project Scheduling Problem (RCPSP) as optimal search over the reachability graph of a Timed Transition Petri Net with Resources, using relative-delay tokens so that scheduling decisions correspond to transition firings in the induced state space. We solve the resulting problem with $A^*$ guided by a heuristic that combines Critical Path and resource-based lower bounds, and prove that it is consistent under our token-based time semantics. Experiments on the PSPLIB benchmarks show that the approach outperforms strong exact Mixed-Integer Linear Programming (MIP) baselines (SCIP, CBC) in both success rate and solve time. Per-instance analysis shows that heuristic search and MIP degrade along independent axes, resource tightness for $A^*$ and formulation size for MIP, with resource strength mediating which solver benefits from scale.
arXiv:2605.15786v1 Announce Type: new Abstract: We present a new strategic voting model where we use uncertainty representation to model preferences. Specifically, we use probability sets as uncertainty representations, together with lower and upper expected utility gains to take strategic decisions. Focusing on belief functions in particular, we demonstrate that this very expressive model includes in one sweep many existing models based on probabilities, sets or incomplete preferences. Additionally, we generalize several well-known convergence results from the literature to this broader representational setting. Furthermore, we illustrate how this model can capture more realistic scenarios for practical applications but also raises theoretical challenges.
arXiv:2605.15409v1 Announce Type: new Abstract: In this paper, we introduce a model for a buoyancy-driven, air-to-air heat exchanger. This model, derived from first principles, features a conservative boundary condition at inflow based on the compressible Bernoulli equation, and a dissipative boundary condition at outflow based on pressure continuity. We solve for the steady-state behavior numerically and asymptotically, with excellent agreement between the two, and we study the tradeoff between the efficiency and air flow predicted by the model.
arXiv:2605.16117v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated strong capabilities across diverse NLP applications, such as translation, text generation, and question answering. Nevertheless, they remain limited in complex settings that demand deep reasoning and logical inference. Since these models are trained on large-scale text corpora, their generation process may still introduce irrelevant, noisy, or factually inconsistent content. To mitigate this problem, we introduce SGR, a stepwise framework that enhances LLM reasoning through external subgraph generation. SGR builds query-specific subgraphs from external knowledge bases and uses their semantic structure to support multi-step inference. By grounding intermediate reasoning steps in structured external knowledge, the framework helps the model concentrate on relevant entities, relations, and supporting evidence. In particular, SGR first constructs a subgraph tailored to the input question. It then guides the model to reason progressively over the generated structure and combines multiple reasoning trajectories to obtain the final prediction. Experimental results across several benchmark datasets show that SGR achieves consistent improvements over competitive baselines, highlighting its value for improving both reasoning accuracy and factual reliability.
arXiv:2605.15240v1 Announce Type: cross Abstract: This paper investigates the critical role of eigenalignments between the kernel matrix and learning targets in achieving robust generalization in learning problems. We establish a direct connection between generalization performance in kernel methods and the estimation of eigenvectors and eigenvalues of matrices, offering a more intuitive understanding compared to prior work with minimal assumptions. We also show that, since the prediction task in KRR is essentially the weighted sum of eigenvectors/singular vectors, by analyzing how much error can be caused by perturbations to the kernel matrix, we can then derive a bound on this generalization error using the estimation stability of matrix eigenvalues and eigenvectors. Compared with previous work, our analysis concentrates on finite-sample settings and on the generalization error arising from having a suboptimal finite training set. Our findings reveal that in kernel methods, as long as the kernel is of high rank, the near-zero reconstruction error can be trivially obtained, implying that the reconstruction error will have limited predictive power for generalization. Finally, we establish a generalization bound from an eigenvalues/eigenvectors estimation perspective, showing that strong generalization requires increasing eigenvector alignment, eigenvalue magnitude, or gaps between consecutive eigenvalues.
arXiv:2605.15699v1 Announce Type: new Abstract: A defining feature of the present-day global overturning circulation (GOC) is the absence of deep water formation in the Pacific, in contrast to the Atlantic. This asymmetry, associated with higher surface salinities in the North Atlantic, is reflected in the Atlantic Meridional Overturning Circulation (AMOC) and the lack of a Pacific overturning (PMOC). A commonly cited explanation is the asymmetry in surface freshwater fluxes, with the Pacific receiving more freshwater per unit area than the Atlantic. Here, we develop a two-basin conceptual ocean model, consisting of a wide and a narrow basin. The model admits three states: sinking confined to the narrow basin, sinking confined to the wide basin, and sinking in both basins. We analyze the (co-)existence of these states as a function of freshwater asymmetry and hydrological cycle strength, defined as the longitudinally symmetric freshwater flux. For a weak hydrological cycle, representative of warm Pliocene-like climate conditions, sinking occurs in both basins, with symmetry breaking only when one basin is sufficiently more evaporative. For intermediate conditions, representative of the present-day climate, the basin with slightly stronger evaporation tends to host sinking, with a stronger preference in the narrow basin. For a strong hydrological cycle, single-basin sinking states are preferred, although a large interbasin freshwater asymmetry is required to uniquely localize sinking. These results provide insight into GOC sinking configurations under past, present, and potential future climates, and show good agreement with a three-dimensional global circulation model.
arXiv:2605.16181v1 Announce Type: new Abstract: Training data attribution (TDA) for music generation must answer two questions that copyright analysis requires, namely which training songs influence a generated output and along which musical aspects the influence operates. Existing methods reduce influence to a single scalar, without revealing which musical aspects are dominant in that influence. We propose ARIA, a framework that decomposes attribution along musical aspects (five for symbolic music, three for audio) and pairs the decomposition with reliability diagnostics computed from the segment-level score matrix. It measures within-group similarity among the top-K attributed tracks against random reference groups drawn from the training pool, and diagnoses the score matrix through its singular value decomposition and column statistics. On a symbolic-music model where attribution ground truth is available through counterfactual retraining, the reliability diagnostics rank four attribution methods identically to that ground truth. On an audio music generation model, ARIA reveals attribution behaviors that vary substantially across TDA methods, flags score matrices whose retrieved tracks are nearly identical across queries rather than reflecting per-query attribution, and characterizes embedding-similarity retrieval baselines by the musical aspect each encoder surfaces. Together, ARIA produces per-aspect attribution evidence aligned with the musical aspects considered under the idea-expression distinction in copyright analysis.
arXiv:2605.15904v1 Announce Type: new Abstract: Cybersecurity Knowledge Graphs (CKGs) unify diverse Cyber Threat Intelligence (CTI) sources into structured, queryable formats, offering scalable solutions for automating proactive and real-time security responses. Their increasing adoption has significantly enhanced the workflow and decision-making efficiency of security professionals. However, constructing CKGs requires extracting entity-relation triples from unstructured CTI reports, a task hindered by complex report structure, domain-specific language, and semantic ambiguity. As a result, existing pipeline-based approaches often suffer from error propagation, reducing extraction accuracy and limiting generalizability. This paper introduces the Context-aware Threat Intelligence Knowledge Graph (CTiKG) framework, a pipeline architecture designed to accurately extract and classify threat entities and their relationships from CTI reports. CTiKG incorporates hybrid NLP models that leverage SecureBERT+ contextual embeddings and expert knowledge from a domain ontology to reduce misclassifications and mitigate cascading errors. Experiments on the DNRTI-AUG-STIX2 dataset, which comprises 21 entity types aligned with STIX 2.1, demonstrate significant improvements over state-of-the-art baselines, yielding 3-4% gains in NER and up to 8% in RE performance, based on precision, recall, and F1-score. Additional validation on DNRTI and STUCCO benchmarks confirms the framework's robustness and practical applicability. All datasets, including the curated DNRTI-AUG-STIX2, are released on GitHub to foster reproducibility and further research.
arXiv:2605.13169v2 Announce Type: replace Abstract: Multimodal large laboratory models (MLLMs) still struggle with spatial understanding under the dominant perspective-image paradigm, which inherits the narrow field of view of human-like perception. For navigation, robotic search, and 3D scene understanding, 360-degree panoramic sensing offers a form of supersensing by capturing the entire surrounding environment at once. However, existing MLLM pipelines typically decompose panoramas into multiple perspective views, leaving the spherical structure of equirectangular projection (ERP) largely implicit. In this paper, we study pano-native understanding, which requires an MLLM to reason over an ERP panorama as a continuous, observer-centered space. To this end, we first define the key abilities for pano-native understanding, including semantic anchoring, spherical localization, reference-frame transformation, and depth-aware 3D spatial reasoning. We then build a large-scale metadata construction pipeline that converts mixed-source ERP panoramas into geometry-aware, language-grounded, and depth-aware supervision, and instantiate these signals as capability-aligned instruction tuning data. On the model side, we introduce PanoWorld with Spherical Spatial Cross-Attention, which injects spherical geometry into the visual stream. We further construct PanoSpace-Bench, a diagnostic benchmark for evaluating ERP-native spatial reasoning. Experiments show that PanoWorld substantially outperforms both proprietary and open-source baselines on PanoSpace-Bench, H* Bench, and R2R-CE Val-Unseen benchmarks. These results demonstrate that robust panoramic reasoning requires dedicated pano-native supervision and geometry-aware model adaptation. All source code and proposed data will be publicly released.
arXiv:2605.15645v1 Announce Type: new Abstract: Irregular memory accesses pose challenges for effective and efficient data prefetching. While temporal prefetchers have recently shown promise for irregular memory access patterns, their effectiveness fundamentally depends on temporal address recurrence and large metadata storage. When memory addresses exhibit weak or no recurrence, as in indirect memory accesses, temporal prefetchers achieve limited performance gains while incurring substantial storage overhead. This paper proposes Instruction-Correlation Prefetching (ICP), a new hardware prefetching mechanism that exploits instruction-level correlations rather than memory-address correlations to handle irregular memory accesses. ICP observes that although memory addresses may not repeat, the instructions generating them often recur with stable data-dependency relationships. By learning these persistent instruction correlations, ICP speculatively computes and prefetches future irregular accesses using the execution results of their correlated predecessors. Across irregular SPEC CPU and GAP benchmarks, ICP outperforms the state-of-the-art temporal prefetcher Triangel by 14.0% and the indirect prefetcher DMP by 6.0%, while requiring only 2.1 KB of hardware storage, over three orders of magnitude smaller than temporal prefetchers.
arXiv:2604.22393v2 Announce Type: replace-cross Abstract: Hexagonal diamond (HD), an exotic carbon allotrope recently synthesized in bulk form, exhibits superior mechanical properties compared to cubic diamond (CD) and holds promise for advanced industrial and quantum applications. Using first-principles calcu-lations, we systematically investigate intrinsic defects, extrinsic dopants, and defect complexes in HD. Our study shows that VC dominates intrinsic conductivity, while Ci is unstable. Among extrinsic dopants, boron acts as a benign acceptor enhancing p-type conductivity, whereas nitrogen and phosphorus serve as effective donors for n-type conductivity. Group II and Group IV dopants, however, introduce high formation energies or neutral charge states with limited impact. Furthermore, VC, MgC and XV defect com-plexes display multiple spin and charge states within the HD band gap, highlighting their potential as color centers for hosting qubits. These results not only clarify the defect physics of HD but also demonstrate its broader implications for conductivity engineering and quantum technologies.
arXiv:2605.15481v1 Announce Type: new Abstract: InGaP-on-insulator, with its intrinsically high $\chi^{(2)}$ optical nonlinearity, has emerged as an efficient and bright integrated photonic platform for frequency conversion and on-chip entanglement generation, but high waveguide propagation loss in the visible wavelength range has limited its overall performance. Here, we identify the dominant loss mechanism through mode-profile analysis and effectively mitigate the loss using a surface treatment method. Statistical analysis of the resonator quality factor and propagation loss reveals the optimal ring radius that maintains a strong nonlinear interaction while suppressing significant bending related loss, resulting in loss as low as 0.49 dB/cm (4.31 dB/cm) at 1560 nm (780 nm). The method provides a 3.5--4$\times$ linear performance enhancement, enabling a second-harmonic generation efficiency of $3.01\times10^{5}$ %/W and a photon-pair generation rate of $11.7,\mathrm{MHz}/\mu\mathrm{W}$ and coincidence-to-accidental ratio as high as 10,000. The quasi-phase matching condition is experimentally verified, and nonlinear conversion is systematically characterized across the entire parameter space. This work establishes a scalable pathway for classical and quantum photonics in a low-loss, highly nonlinear, and wafer-scale integration platform.
arXiv:2605.15933v1 Announce Type: new Abstract: We first show how the cohomology of some Bernstein-Gelfand-Gelfand (BGG) sequences that are important for the numerical analysis of partial differential equations, can be obtained through the construction of a long exact sequence connecting cohomology groups. Then we explain the extension of this result to the non-injective/surjective case through the systematic use of short exact sequences of complexes and their associated long exact sequences of cohomology groups. Finally an interpretation in terms of spectral sequences is given.
arXiv:2605.15726v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has emerged as a scalable paradigm for improving the reasoning capabilities of large language models. However, its effectiveness is fundamentally limited by exploration: the policy can only improve on trajectories it has already sampled. While increasing the number of rollouts alleviates this issue, such brute-force scaling is computationally expensive, and existing approaches that modify the optimization objective provide limited control over what is explored. In this work, we propose NudgeRL, a framework for structured and diversity-driven exploration in RLVR. Our approach introduces Strategy Nudging, which conditions each rollout on lightweight, strategy-level contexts to induce diverse reasoning trajectories without relying on expensive oracle supervision. To effectively learn from such structured exploration, we further propose a unified objective, which decomposes the reward signal into inter- and intra-context components and incorporates a distillation objective to transfer discovered behaviors back to the base policy. Empirically, NudgeRL outperforms standard GRPO with up to 8 times larger rollout budgets, while outperforming oracle-guided RL baseline on average across five challenging math benchmarks. These results demonstrate that structured, context-driven exploration can serve as an efficient and scalable alternative to both brute-force rollout scaling and feasibility-oriented methods based on privileged information. Our code is available at https://github.com/tally0818/NudgeRL.
arXiv:2605.16116v1 Announce Type: new Abstract: Developing and evaluating e-commerce web agents requires environments that preserve meaningful task structure while enabling controllable, reproducible, and scalable scientific comparison. Existing methodologies force a tradeoff: live storefronts provide realism but are non-stationary, difficult to inspect, and irreproducible, while hand-built sandbox benchmarks provide control but cover only a narrow range of layouts, catalogs, policies, and interaction patterns. We argue that the core bottleneck is methodological: the field lacks a scalable way to construct evaluation settings that are simultaneously realistic, diverse, controllable, inspectable, and reproducible. We introduce ShopGym, an integrated framework for realistic simulation and scalable benchmarking of e-commerce web agents. ShopGym is a framework for constructing e-commerce simulation environments and grounded benchmark tasks. Its simulation layer, ShopArena, converts live seed storefronts into self-contained sandbox shops through anonymized shop specifications and a staged, validated generation process. On top of these simulated storefronts, ShopGuru synthesizes benchmark tasks across seven skill categories, grounding each task in the shop's catalog, navigation structure, policies, and interaction affordances. Together, ShopArena and ShopGuru produce self-contained, resettable, inspectable, and stable evaluation artifacts that preserve structural properties and agent-evaluation signals relevant to shopping tasks. We validate the framework through graph-based structural analysis and agent-based behavioral evaluation with 224 generated tasks across six sandbox shops: three constructed with synthetic data and three with real data. Our results show that the synthetic shops preserve key structural properties of live storefronts, with agent performance on synthetic shops positively correlated with performance on live storefronts.