arXiv:2606.04411v1 Announce Type: new
Abstract: We present Pepper, a high-bandwidth anonymous broadcast protocol that provides cryptographic sender anonymity against global adversaries. Pepper builds on a two-server DC-net architecture but introduces three key innovations: a self-contained anonymous registration subprotocol using verifiable distributed point functions, support for batch messaging via distributed multi-point functions, and a lightweight access control mechanism based on secret-shared proofs. Unlike prior systems, Pepper eliminates the need for external dialing services and allows each broadcaster to send multiple messages per epoch with a single audit, significantly improving throughput for large data transfers. Our implementation demonstrates that Pepper achieves millisecond-level registration audits, scales efficiently to thousands of channels, and delivers 1.2--20$\times$ higher effective messaging rates than state-of-the-art alternatives. Furthermore, Pepper is designed for practical deployment, with natural compatibility for co-deployment alongside Tor and federated social networks.
Science Journals
arXiv:2606.04414v1 Announce Type: new
Abstract: Multi-view cardiac magnetic resonance (CMR) imaging provides complementary anatomical information and is widely used for noninvasive disease assessment. Recent transformer-based models have demonstrated strong representation learning capabilities for CMR analysis; however, they typically learn unified latent embeddings that entangle view-specific anatomical variations with disease-related features. Such entanglement biases classifiers toward structural attributes rather than view-invariant pathological patterns. This issue is exacerbated in low-data regimes, particularly for underrepresented cardiac conditions, where limited samples increase the susceptibility to shortcut learning and view-dependent decision boundaries. To address this, we propose a Motion-Guided View--Disease Disentanglement framework MoViD built upon a ViT-MAE backbone. The model explicitly factorizes latent representations into view-specific and disease-discriminative components using dual-branch supervised contrastive objectives and a gradient-reversal adversarial constraint that minimizes disease leakage into the view embedding. Additionally, an annotation-free temporal motion feature, derived from inter-frame difference maps, is introduced to localize the beating heart region and suppress background artifacts. A focal reweighting mechanism is incorporated into the contrastive loss to mitigate class imbalance. We evaluate the framework on a private clinical venous thrombosis dataset and two public benchmarks (M&Ms, M&Ms2). Across disease classification and cardiac segmentation tasks, our approach consistently outperforms standard transformer baselines and demonstrates competitive performance against large-scale pretrained foundation models, validating the efficacy of structural disentanglement in medical image analysis.
arXiv:2605.19852v2 Announce Type: replace
Abstract: Tool-augmented reasoning has emerged as a promising direction for enhancing the reasoning capabilities of multimodal large language models (MLLMs). However, existing studies mainly focus on enabling models to perform tool invocation, while neglecting the necessity of invoking tools. We argue that tool usage is not always beneficial, as redundant or inappropriate invocations largely increase reasoning overhead and even mislead model predictions. To address this issue, we introduce AutoTool, a model that adaptively decides whether to invoke tools according to the characteristics of each query. Within a reinforcement learning framework, we design an explicit dual-mode reasoning strategy with mode-specific reward functions to guide the model toward producing accurate responses. Moreover, to prevent premature bias toward a single reasoning mode, AutoTool jointly explores and balances tool-assisted and text-centric reasoning throughout training, and promotes free exploration in later stages. Extensive experiments demonstrate that AutoTool exhibits outstanding performance and high efficiency, yielding a 21.8\% accuracy gain on V* benchmark compared to the base model, and a 44.9\% improvement in efficiency over existing tool-augmented methods on POPE benchmark. Code is available at https://github.com/MQinghe/AutoTool.
arXiv:2605.19398v3 Announce Type: replace
Abstract: Image-to-video models often generate videos that remain overly static, compared to text-to-video models. While prior approaches mitigate this issue by weakening or modifying the image-conditioning signal, they often require additional training or sacrifice fidelity to the reference image. In this work, we identify reference-frame dominance as a key mechanism behind motion suppression. We observe that non-reference frames in I2V models allocate excessive self-attention to reference-frame key tokens, causing reference information to be over-propagated across time and suppressing inter-frame dynamics. Based on this finding, we propose DyMoS (Dynamic Motion Slider), a training-free and model-agnostic method that rebalances the attention pathway from generated frames to the reference frame during initial denoising steps. DyMoS leaves both the input image and model weights unchanged and introduces a single scalar parameter for continuous control over motion strength. Experiments across multiple state-of-the-art I2V backbones demonstrate that DyMoS consistently improves motion dynamics while maintaining visual quality and fidelity to the reference image.
arXiv:2605.18936v2 Announce Type: replace
Abstract: Social media text data are often used to train Machine Learning (ML) models to identify users exhibiting high-risk mental health behaviors. However, sharing this sensitive data poses privacy risks and limits the growth of benchmark datasets. We comprehensively evaluate whether privacy-preserving ML techniques can enable safer data sharing while preserving performance. Specifically, we apply federated learning (FL) and Differentially Private FL for two widely-studied mental health prediction tasks: depression detection on X (Twitter) and suicide crisis detection on Reddit. We simulate realistic data-sharing scenarios by treating each user as a client in a non-IID setting, evaluating across different client fractions, aggregation strategies, and privacy budgets. While FL achieves comparable performance to centralized training (centralized F1 = 85.63; best FL model F1 = 83.16) on depression identification, we find that Differentially Private FL has a large performance-privacy trade-off (up to F1 = 27.01 drop) even with low levels of noise (epsilon = 50). This is due to the distortion of highly informative yet sparse mental health linguistic markers related to mental health, like health topics and emotion words. This research empirically demonstrates the potential and limitations of current privacy preservation techniques for mental health inference tasks.
arXiv:2606.04605v1 Announce Type: cross
Abstract: We use lubricated-flow discrete-element-method (LF-DEM) simulations to examine how contact-network organization shapes particle motion in dense shear-thickening suspensions. The primary system studied is a two-dimensional bidisperse monolayer where rigid clusters are identified by the $(3,3)$ pebble game; three-dimensional simulations are shown to have qualitatively similar rotational velocity statistics. Across the stress--solid-fraction state diagram, frictional contact number, $k\ge 3$ percolation, and rigid-cluster fluctuations all strengthen in the same region where translational velocity correlations grow, consistent with rigid clusters translating coherently while the surrounding non-rigid particles accommodate a disproportionate share of the local velocity gradient. Rotational motion provides a complementary view: non-affine angular-velocity distributions broaden, near-contact rotations become increasingly anti-correlated, and rigid and non-rigid particles carry distinct statistics. Connectivity, rigidity, and velocity correlations are related but distinct signatures of the constrained collective motion that accompanies shear-thickening and the approach to shear jamming.
arXiv:2606.04576v1 Announce Type: cross
Abstract: Learning Value-at-Risk (VaR) and Expected Shortfall (ES) is important for managing financial risks effectively. Existing approaches with limited parameters are vulnerable to model misspecification in the era of big data. To address this limitation, we propose a large tail risk model, the retrieval-enhanced self-grouping autoencoder (ReSGA), which is designed with millions of parameters to exploit the rich cross-sectional dependence and long-term temporal dynamics of assets using their characteristics. Applied to monthly US equity returns from 1926 to 2023 with 153 firm characteristics, ReSGA outperforms twelve econometric and machine learning competitors in terms of out-of-sample loss and statistical backtesting. In addition, its forecast advantages can translate into significant economic gains from long-short decile portfolios that are constructed by a new size-enhanced left-side momentum strategy. To clarify the role of complexity, we further conduct a systematic scaling analysis and demonstrate that improvements in joint VaR-ES forecasting are primarily driven by data complexity rather than model complexity. Finally, our analyses of group-importance and transfer-learning exhibit the interpretability and cross-market generalizability of ReSGA.
arXiv:2606.05115v1 Announce Type: new
Abstract: Children learn the meanings of words from a continuous, temporally structured stream of egocentric experience. Recent work shows that neural networks can also learn word-referent mappings from a child's egocentric video recordings, but they cycle through the shuffled data for hundreds of epochs, contrasting with how children actually encounter their environment. We introduce BabyCL, a continual multimodal learning framework that processes the SAYCam dataset in a single chronological pass, combining streaming visual representation learning with an image-text contrastive objective. BabyCL combines a multi-stage temporal segmentation of the stream with a dual replay buffer that independently manages visual and multimodal histories, and it is jointly trained with three contrastive losses on a shared backbone. Under a matched optimization budget, BabyCL outperforms streaming learning baselines on the SAYCam Labeled-S 4AFC benchmark, substantially narrowing the gap to an upper bound of offline training. Ablations show that the gains are robust to the length of the online temporal segmentation window and the eviction rule of the replay buffer. Together, these results show that meaningful word-referent mappings can emerge under training conditions much closer to a child's actual experience.
arXiv:2606.04444v1 Announce Type: cross
Abstract: Existing datasets cannot support large-scale learning in multi-agent, multi-sensor, or multi-domain autonomy, where diversity and coordination are essential. We present a modular dataset generation pipeline that creates terabyte-scale, ground-truth-labeled data for ground, aerial, and infrastructure-based systems using the AVstack framework and CARLA simulator. Supporting single- and multi-agent configurations with flexible sensor suites, the pipeline enables controllable experimentation across challenging conditions. Representative perception and fusion studies show how generated data can support application-specific training and collaborative autonomy.
arXiv:2606.04429v1 Announce Type: cross
Abstract: A common heuristic used to explain the generalization of first-order gradient methods on non-convex neural networks is that "flat interpolators generalize well" (Hochreiter and Schmidhuber, 1994; Keskar et al., 2017), where flatness can be measured by the trace of the Hessian of the empirical loss. However, Dinh et al. 2017) showed that, using symmetry of the network that can change flatness while keeping the population and empirical losses unchanged, any interpolator can be made sharper or flatter. This result makes the earlier heuristic statement vacuous. In this paper, we show that for learning an unknown multi-index model with $2$-layer non-convex homogeneous neural networks, there is a connection between flatness and generalization, despite the existence of symmetries. This connection pertains to the "flattest" interpolators, i.e., the interpolators that have orderwise minimum flatness among all interpolators. First, we show that there exists a natural class of non-generalizing interpolators whose flatness cannot be made closer to the flattest possible, even using symmetries. Second, we show that for data generated by a sum of single-index models, if the approximation error and label noise are low, any flattest interpolator achieves small population loss, i.e., the flattest interpolators always generalize. This establishes a direct link between flatness and generalization which applies to a large class of activations and realistic data distributions.
arXiv:2605.18879v3 Announce Type: replace
Abstract: Large language models inevitably retain sensitive information, defined as inputs that may induce harmful generations, due to training on massive web corpora, raising concerns for privacy and safety. Existing machine unlearning methods primarily rely on retraining or aggressive fine-tuning, which are either computationally expensive or prone to degrading related knowledge and overall model utility. In this work, we reformulate machine unlearning as a precise knowledge re-mapping problem via model editing. We propose ZeroUnlearn, a few-shot unlearning framework. It overwrites sensitive inputs by mapping them to a neutral target state and removing their original representations. ZeroUnlearn enforces representational orthogonality through a multiplicative parameter update with a closed-form solution, enabling efficient and targeted unlearning. We further extend ZeroUnlearn to a gradient-based variant for multi-sample unlearning. Experiments demonstrate that our approach outperforms existing baselines while preserving general model utility. Our code is available at the github: https://github.com/XMUDeepLIT/ZeroUnlearn.
arXiv:2605.18102v3 Announce Type: replace
Abstract: Monocular video human mesh recovery is essential for digital humans, avatar animation, and embodied simulation, where both temporal stability and expressive whole-body motion are required. Existing video HMR methods produce coherent body motion but often overlook detailed hand articulation, while image-based whole-body methods recover SMPL-X meshes independently per frame, often leading to jittery and inaccurate hand motion. We present a temporally coherent whole-body HMR framework for challenging in-the-wild monocular videos. Our model unifies body context and part-specific hand observations through residual body-hand fusion, enabling stable body motion and detailed hand recovery within a single temporal architecture. We further introduce close-up-aware augmentation to improve robustness under upper-body framing. Experiments on whole-body and body-only benchmarks demonstrate improved hand reconstruction and competitive body accuracy. Our method also produces temporally stable and 2D-consistent SMPL-X motion in challenging real-world videos.
arXiv:2605.17273v3 Announce Type: replace
Abstract: State-of-the-Art (SOTA) claims pervade Artificial Intelligence (AI) and Machine Learning (ML) research. These claims rest on benchmark evaluations, where models are ranked by aggregate scores across tasks. Public benchmarks or leaderboards are the most visible instance, but the same structure appears in paper tables throughout the literature. However, such minimal evidence often cannot support these strong claims. We identify a widespread claim-evidence gap in AI benchmarking. Claiming SOTA carries implicit assumptions beyond mean score superiority, suggesting that a model meaningfully outperforms alternatives across most tasks. However, a marginal improvement in the mean score merely indicates a top average rank rather than true superiority. Analyzing ten cross-domain benchmarks from public leaderboards, we found that in more than half of top-model comparisons, at least one commonly assumed property of superiority does not hold. These properties include meaningful effect size, consistency across tasks, or robustness to dataset removal. Instead, aggregate gains are frequently driven by outlier datasets. This fragility persists even in benchmarks with many tasks. We argue that claim language should reflect the strength of the underlying evidence. This requires no additional experiments, only honest reporting of what results actually show, enabling more precise and interpretable comparisons across models.
arXiv:2605.15949v2 Announce Type: replace
Abstract: This paper presents a reproducible and physically feasible dynamic parameter identification framework for CRANE-X7, a low-cost robot arm driven by modular smart actuators. To improve practical identifiability, products of inertia are removed according to approximate link symmetry, reducing the rigid-body model from 65 to 39 base parameters. Identification motions are hand-designed from structured single-joint and adjacent-joint primitives under practical joint-range limits. The proposed pipeline combines preprocessing, inverse-dynamics-regressor-based ordinary least squares (OLS), conditional semidefinite-programming (SDP) projection for feasibility recovery, and closed-loop input error (CLIE) refinement. Candidate solutions from 40 structured trajectories are analyzed in a common principal component analysis (PCA) space to select a statistically central representative model. Because statistical centrality alone does not ensure physical acceptability, the selected model is finally screened by an all-pose positive-definiteness audit of the inertia matrix and, when necessary, corrected by a localized post-CLIE SDP rescue step. Experiments show that the parameter cloud becomes progressively more concentrated from OLS to SDP and CLIE, while the final accepted model preserves high predictive accuracy on held-out validation motions. These results demonstrate a practical route to statistically coherent and physically feasible dynamic models for low-cost robot platforms.
arXiv:2302.05590v3 Announce Type: replace-cross
Abstract: A central tenet in mechanism design is the ability to irrevocably commit to a mechanism. Commitment is achieved by public declaration, letting players verify incentive properties in advance and the outcome in retrospect. However, public declaration can reveal superfluous information that is private to the mechanism designer, such as her target function or costs. We propose a new approach to commitment, and show how to commit to, and run, any given mechanism without disclosing it, while enabling the verification of incentive properties and the outcome -- all without any mediators. Our framework leverages zero-knowledge proofs -- a cornerstone of modern cryptographic theory.
arXiv:2606.04047v1 Announce Type: new
Abstract: In recent years, the detection of anti-patterns in microservice architecture has gained traction, particularly to identify instances of Microservice Architectural Degradation. In such tasks, the microservice architecture is often modeled as a network of microservice dependencies. Recent works have explored how to assess the evolution of such architectural networks by considering the architecture of consecutive releases of the project. Particular anti-patterns related to the structure of the service network include Wrong cuts and Knot services.
Community detection is a way to identify groups of services in a network that strongly depend on each other. If such groups cannot be mapped to business processes in the system, or if the same service belongs to multiple communities, this could indicate architectural degradation due to an inappropriate division of responsibilities or unoptimized communication. Temporal community detection methods have been proposed to analyze community structure that evolves in time.
We performed temporal community detection within the microservice architecture of six releases of the train-ticket benchmark and analyzed the composition of the discovered communities and their activities over time. We observed a stable architecture with a clear separation of services into two communities, which we could identify with two business processes performed by the system. We found services belonging to several communities, as well as services within the same community with both incoming and outgoing connections. The membership strength metric provided by the leveraged algorithm enables fine-grained assessment of the microservice communities.
arXiv:2606.04656v1 Announce Type: new
Abstract: Object detection is a safety-critical component of autonomous driving. It is essential to quantify the uncertainty in bounding-box predictions for safety assurance. Post hoc uncertainty quantification without retraining aligns with real-world deployment requirements; therefore, we employ the Laplace approximation. Because instance-level uncertainty is needed, linearized inference methods that require multiple backpropagations are not time-efficient, and sampling-based methods are not fully post hoc. We propose Monte-Carlo generalized linearized model (MC-GLM), which provides instance-level and approximately post hoc uncertainty quantification. The number of samples required in the Monte Carlo step is constant and independent of the number of output instances, so it can be parallelized. Experiments on the nuScenes dataset with the CenterPoint detector validate the effectiveness of our method, and the resulting uncertainties exhibit good quality.
arXiv:2606.05155v1 Announce Type: cross
Abstract: Formulating stable numerical methods for hyperbolic systems in curvilinear coordinate with singularities, e.g. spherical coordinates, is complicated by the presence of these singularities. We present a method for constructing high-order accurate, energy-stable finite difference operators satisfying the Summation-by-Parts (SBP) property on spherical domains, extending ideas presented by [C. Gundlach, J. M. Mart\'in-Garc\'ia, and D. Garfinkle, CQG 30, 145003 (2013)]. We define discrete gradient and divergence operators that mirror the continuous integration-by-parts principle, even though there is a $1/r^p$ coordinate singularity present at the origin. We explicitly construct such operators up to order six. Our operators place a grid point directly on the origin. We also review how to construct stable SBP operators that straddle the origin. We analyze the accuracy and spectral radii of these operators, and we show example evolutions of the scalar wave equation to demonstrate the advantages of such operators.
arXiv:2606.04419v1 Announce Type: cross
Abstract: MRI provides excellent soft-tissue contrast without ionizing radiation, but long acquisition times increase patient discomfort while also raising exam costs and limiting scanner throughput. A common approach to reduce scan time is to acquire fewer measurements, which yields an ill-posed linear inverse problem; recovering diagnostic-quality images therefore requires incorporating prior knowledge beyond the measured data. In follow-up exams, the most recent prior scan of a patient can provide a highly informative subject-specific context, but practical use is complicated by temporal changes (including pathology progression), misalignment between scans, and protocol drift across acquisitions. In this work, we introduce L-TGVN, a Longitudinal Trust-Guided Variational Network that leverages prior scans as side information to reconstruct the current scan from heavily undersampled measurements. Crucially, L-TGVN constrains the influence of prior scans to be consistent with the acquired measurements. Unlike many existing longitudinal reconstruction methods, it does not require explicit pre-registration between prior and current scans. It further accommodates differences in acquisition protocols across visits (e.g., changes in sequence parameters). We evaluate L-TGVN against matched-capacity baselines, including prior-guided methods and methods that do not use longitudinal priors, and observe consistent improvements in standard quantitative metrics together with better preservation of fine structures at challenging accelerations. Source code is available at github.com/sodicksonlab/L-TGVN.
arXiv:2606.04757v1 Announce Type: cross
Abstract: We study decentralized stochastic smooth convex optimization, where $M$ workers minimize an average objective using local stochastic gradients and neighbor-only communication over a fixed gossip network. A central question in this setting is to determine the largest number of workers that can be used under a total budget of $N$ gradient samples while still preserving the centralized $O(1/\sqrt N)$ statistical rate. We introduce an accelerated decentralized method that preserves this rate for up to $\smash{M\lesssim \sqrt{\rho}\,N^{3/4}}$ workers, where $\rho$ is the spectral gap of the gossip network, improving the best prior maximal scaling of $\smash{M\lesssim \rho\sqrt N}$. The method is based on a one-step-delayed stochastic acceleration scheme that enables workers to interleave minibatching with accelerated gossip while controlling residual disagreement, and its guarantee depends only logarithmically on the optimum-local heterogeneity. We also establish a matching lower bound for linear-span decentralized first-order methods, showing that the method is optimal up to logarithmic factors.
arXiv:2606.04027v1 Announce Type: new
Abstract: Diffusion large language models (dLLMs) generate text by iteratively denoising partially masked sequences under bidirectional context, exposing a safety surface distinct from autoregressive LLMs. Because mask tokens are native inputs and tokens are committed by confidence rather than position, harmful content can be induced through infilling and outside the monitored prefix. Existing jailbreaks either miss this native infill capability or rely on low-diversity mask-bearing templates applied uniformly across goals, with little structural adaptation or accumulated attack experience. We propose MaskForge, a fully black-box adaptive attack that casts dLLM red-teaming as optimized search over a growing library of structural patterns. MaskForge abstracts successful attempts into reusable schemas, selects goal-compatible patterns with a UCB bandit, and invokes a scorer-guided fallback when the current library fails. Successful attempts are distilled back into the pattern library, enabling experience to accumulate across goals. Across five public dLLMs and three benchmarks, MaskForge achieves an average attack success rate of 79.3%, a 17.6% relative improvement over the strongest competing dLLM baseline. The matured pattern library further transfers to AdvBench without any updates, achieving a 88.2% attack success rate and a 67% relative improvement over the strongest competing baseline.
arXiv:2606.04464v1 Announce Type: new
Abstract: As data becomes increasingly central across engineering and scientific disciplines, effective visualization is essential for interpreting complex, high-dimensional structures. Dimensionality reduction techniques project high-dimensional data into lower dimensions while aiming to preserve structural properties such as pairwise distances and local neighborhoods. In this paper, we focus on improving homological preservation, that is, the retention of topological features such as connected components and loops, which is critical for maintaining global shape and continuity. We first introduce AdaMapper, a Mapper-based algorithm that leverages persistence diagrams to guide both skeleton construction and landmark selection. AdaMapper incorporates an adaptive refinement strategy that automatically increases cover resolution in regions exhibiting topological loops. We then propose AdaHIsomap, which extends landmark Isomap by incorporating homology-informed landmark selection and augmenting it with random anchor points to better balance distance and homology preservation. We evaluate both methods on a diverse set of datasets, including high-dimensional point clouds, scientific simulations, networks, and image data, and benchmark their performance against state-of-the-art approaches.
arXiv:2606.04506v1 Announce Type: new
Abstract: The conversion of functions to quantics tensor trains is a well-established procedure and can either be done analytically or numerically. Numerical conversion schemes are based on singular value decompositions, where access to the full tensor is necessary, or on cross interpolations, which only depend on sampling a function. When dealing with large binary tensors, the first approach becomes prohibitively expensive while the second approach might fail to converge due to the non-smoothness of the data. In this work, we provide insight into how binary tensors, where the positions of the non-zero entries are defined by some Boolean function, can be converted into quantics tensor trains with a hybrid analytical-numerical approach utilizing the rank product. The proposed construction scheme nicely reproduces results from the literature and can be used for discrete convolutions, the construction of discrete wavelet transforms or slicing and assignment operations of multi-dimensional quantics tensor trains.
arXiv:2606.04173v1 Announce Type: cross
Abstract: It is well-known that any class of simple graphs, that is characterized by finitely many forbidden minors, also admits a characterization by finitely many forbidden topological minors; furthermore, the list of forbidden topological minors may be derived from the list of forbidden minors.
We prove a similar result in Matching Theory. Our Main Theorem states that any class of matching covered graphs, that is characterized by finitely many forbidden $S$-minors that are cubic, also admits a characterization by finitely many forbidden conformal minors that are cubic as well; once again, the list of forbidden conformal minors may be derived from the list of forbidden $S$-minors.
In order to establish the above, we first prove that every matching covered graph has one of two graphs as a conformal minor -- either $K_4$, or the $\Theta$ graph (that is, two vertices joined by three edges). (In fact, we need and prove a much stronger statement.) This is reminiscent of a theorem due to Lov\'asz: every nonbipartite matching covered graph has one of two graphs as a conformal minor -- either $K_4$, or the triangular prism $\overline{C_6}$.
As applications of our Main Theorem, we deduce known 'forbidden conformal minor characterizations' of pfaffian near-bipartite graphs, and of pfaffian solid graphs, using their respective known 'forbidden $S$-minor characterizations'.
arXiv:2606.04188v1 Announce Type: new
Abstract: Offline goal-conditioned reinforcement learning requires both long-horizon reachability estimates and local action comparisons. Dual goal representations provide value fields that capture global goal reachability, but they do not directly specify which action should be preferred at a given state. We propose Dual Advantage Fields, a policy-extraction method that turns a bilinear dual value model into a local advantage signal. Under bilinear dual parameterization, the goal embedding is the gradient of the value field with respect to the state representation. DAF learns an action-effect model that predicts the discounted feature displacement induced by an action and scores actions by the alignment between this displacement and the goal direction. In the realizable case, this score equals the goal-conditioned Bellman advantage, yielding a standard local policy-improvement guarantee. On OGBench locomotion, manipulation, and puzzle tasks, DAF improves aggregate RLiable metrics and performs strongly in settings where locally correct actions differ from direct movement toward the final goal.