arXiv:2607.14386v1 Announce Type: new
Abstract: Data science tasks span from closed-ended information extraction to open-ended analysis, presenting significant challenges for automation. Recent AI agents powered by language models show promise for handling such complex tasks. However, existing agents typically rely on a single initial state that conditions the entire agent's execution, making them vulnerable to cascading errors initiated by a suboptimal initial state. To mitigate this, we present CIPHER, an automated data science agent that leverages test-time scaling through the generation and selection of multiple initial states for concurrent execution. Unlike existing works on test-time scaling of AI agents, CIPHER explicitly decouples the generation of candidate initial states from their strategic selection for parallel execution. Through extensive evaluation on two benchmarks (closed-form and open-form tasks), we demonstrate that CIPHER exceeds state-of-the-art performance in matched-model comparisons, and remains competitive against larger-model baselines despite relying on a substantially smaller base LM. Our empirical study characterizes the design space of the Decoupled Exploration-Selection (DES) framework: we quantify how generation strategy, selection strategy, and aggregator model capacity contribute to overall performance, and derive actionable design recommendations for practitioners.
Science Journals
arXiv:2607.14709v1 Announce Type: new
Abstract: Financial question answering over hybrid tabular and textual data may require multi-source reasoning and precise numerical computation. While large language models (LLMs) can generate intermediate reasoning steps, natural-language rationales remain prone to arithmetic errors, making them an unreliable supervision source for distillation. Building on programmatic distillation, we develop an approach that transfers reliable numerical reasoning from a large teacher model to a compact student using execution-verified Python programs instead of free-form textual rationales. It leverages gold derivations to guide teacher-side program synthesis and retains only programs that execute correctly and produce the gold answer, ensuring high-quality supervision. We further introduce an iterative recovery stage that revisits teacher-failed examples, enabling the student to recover and incorporate newly verified programs into training. Experiments on TAT-QA show that our framework is highly effective for hybrid financial reasoning. Our best 7B student achieves 87.00 EM / 87.18 F1 on the test set, substantially outperforming the 72B teacher (78.46 EM) as well as traditional and strong LLM-based baselines, including TAGOP and TAT-LLM. These results demonstrate that execution-verified programmatic distillation provides an effective and extensible framework for training smaller models to perform reliable numerical reasoning.
arXiv:2607.14448v1 Announce Type: new
Abstract: We consider the problem of group testing, in which one seeks to identify a subset of defective items of size $k$ from a larger set of $n$ items based on pooled tests. We introduce a selectable threshold model, in which each test has an associated threshold that can be chosen, such that the test outcome is 1 if and only if the number of defectives in the test is no smaller than that threshold. In settings with a large or unbounded maximum threshold, we establish conditions under which high-probability recovery can be attained with a rate (i.e., the asymptotic ratio of $\log_2{n \choose k}$ to the number of tests) approaching its maximum possible value of 1. Moreover, in the case of a fixed maximum threshold, we establish an achievable number of tests using simple and computationally efficient decoding methods, and a converse that holds under suitable regularity conditions on the test design, with the two coinciding in the dense limit (i.e., $\theta$ approaching one in the scaling $k = \Theta(n^{\theta})$).
arXiv:2607.14471v1 Announce Type: new
Abstract: We investigated the impact of the anode's work-function on the performance of an all inorganic PeLED device which was hailed for stability and long lifetime with drift-diffusion and Poisson equations embedded as a computation package in Setfos, a commercially professional software to emulate LED. According to the simulation results, a conclusion that a thin high work-function metal layer inserted between indium tin oxide (ITO) and hole transport layer of nickel oxide (NiOx) substantially facilitated hole injection by decrementing the hole injection barrier. 2 nm thick gold (Au) had been checked and dramatically enhanced luminance intensity by more than 18 fold times had been attained. The situation went for other high work function metals like Pd and Pt as well. Our high work-function thin metal layers decorating anodes in PeLED in facilitating hole injection unfolds a new trajectory towards efficiently boosting PeLED luminance with a technique that is compatible with large scale fabrication process.
arXiv:2607.14890v1 Announce Type: new
Abstract: Autonomous coding agents increasingly execute multi-step software work, but lifecycle states such as reviewed, tested, DONE, and ready-to-merge remain claims unless supported by current evidence. We present Proof-or-Stop Lifecycle Control, a method that permits lifecycle transitions only when fresh, tracked-source-state-bound, mechanically verifiable evidence satisfies the relevant gate. The method treats agent outputs as claims rather than lifecycle state, and uses proof operationally to mean gate-admissible evidence under a stated trust model, not semantic program correctness.
We evaluate an open-source implementation through mechanism tests, a powered control-policy ablation, and operated self-application evidence. The unattended-loop engine passed 10 of 10 scenarios with zero false-DONE, and local-key receipt bundles rejected 18 tamper classes with zero false accepts. In a 9,240-cell ablation, the pre-registered A4 versus A2-prime comparison reduced visible-pass/hidden-fail amplification from 31 of 1,800 injected cells under a compute-budgeted naive loop to 2 of 1,800 under the gated loop, a 1.6 percentage-point improvement in not-amplified rate with a 95 percent confidence interval of [0.8, 2.5]. A near-compute A3 versus A4 comparison, 14 of 1,800 versus 2 of 1,800, indicates that the gain is associated with enforcing review as a lifecycle gate rather than merely adding a reviewer. The self-application corpus contains 565 stories and 1,007 review findings, with 94.8 percent resolved, plus a 68-row high/critical cross-vendor exhibit. These results support Proof-or-Stop as a model-agnostic, host-neutral control layer for deciding which autonomous-agent claims a lifecycle may act on. The evaluation is limited to one model family, 24 ablation tasks, and a self-hosted corpus.
arXiv:2607.14765v1 Announce Type: new
Abstract: Rare concept generation focuses on synthesizing customized images conditioned on text prompts that describe objects with unusual attributes. Previous works failed to align the generated images with rare concepts, resulting in incorrect attribute rendering or inconsistent composition of concepts. Such failures, as we observed, stem from the inherent common knowledge bias in the training stage of diffusion models, where objects are strongly associated with their common attributes, making it difficult to break these associations when generating rare concepts. To address such challenges, in this paper, we propose a novel Counterfactual Inference-based Diffusion approach, dubbed CI-Diff. CI-Diff blocks the interference of the model's inherent common knowledge bias and utilizes the Natural Direct Effect to capture the independent influence of the text prompt of rare concepts on image generation so that decoupling the unusual attributes from the rare concepts. To this end, we reformulate the classifier-free guidance mechanism to highlight the atypical attributes. To the best of our knowledge, we are the first to introduce causal inference into the rare concept generation task. Extensive experiments on the RareBench benchmark validate the superiority of CI-Diff over state-of-the-art diffusion models. Our code can be accessed from https://github.com/200204jzy/CI-Diff.
arXiv:2607.14506v1 Announce Type: new
Abstract: While reinforcement learning with verifiable rewards (RLVR) is widely used to improve the reasoning capabilities of large language models (LLMs), the generalizability of the resulting models remains poorly understood. In this work, we establish the first non-vacuous generalization bounds for parameter-efficient RLVR fine-tuning at the billion-parameter scale. Our approach adapts PAC-Bayes compression bounds to this setting, and addresses the inherent stochasticity of token generation by applying the Gumbel-max reparameterization trick. To operationalize these bounds, we propose the Progressive RLVR framework, which integrates RLVR with on-policy distillation, TinyLoRA, and model quantization. Progressive RLVR empirically retains 84-97% performance of standard LoRA fine-tuning while producing models that are 14,796x more compressible. We show that this framework yields non-vacuous generalization bounds in four domains: mathematical problem-solving, programming, general-knowledge reasoning, and Text-to-SQL. Our bounds exceed the accuracy of the base model by 9-51% and lie within 6-11% of the accuracy of the fine-tuned models.
arXiv:2607.14555v1 Announce Type: new
Abstract: We study the variance of Q-function estimators in finite-horizon, finite-state Markov decision process (MDP) tree search. We show that the variance decomposes into three components attributed to the immediate reward collected, probabilistic state transitions, and uncertainty in future state value function estimates. Using this decomposition, we show that the sample variance estimator based on the assumption of i.i.d. paths is biased, underestimating the true variance, and the bias does not vanish in the limit. We then propose a recursive variance estimator that is consistent. To enable efficient storage and computation, we derive an equivalent implementation of the recursive estimator using only node-local statistics that can be iteratively updated. This consistent variance estimator is integrated into two Monte Carlo Tree Search (MCTS) sampling procedures for finite-horizon MDPs. In numerical examples from inventory control and kidney paired donation matching, the new estimator improves the performance of the MCTS algorithm relative to a baseline that uses the i.i.d.-based sample variance estimator.
arXiv:2607.14563v1 Announce Type: new
Abstract: Estimating spatially heterogeneous elastic properties from low-resolution displacement measurements is a severely ill-posed inverse elasticity problem because low resolution obscures spatial details needed to distinguish heterogeneous property variations, and small measurement perturbations or fitting errors are amplified through inverse estimation. Existing inverse methods often rely on high-fidelity observations and manually prespecified loss weights, limiting their adaptability and making them sensitive to noise and resolution degradation. We propose a Probabilistic Inverse Elasticity Physics-Informed Neural Network (PIE-PINN) framework for robust estimation of Young's modulus and Poisson's ratio from noisy, low-resolution displacement data. PIE-PINN models displacement observation, strain-discrepancy, and equilibrium residuals using Laplace distributions within a unified probabilistic model. To improve robustness, the framework combines a B-spline-guided displacement network with a hierarchical half-Cauchy model for displacement residual scales. The B-spline provides a smooth global representation of the displacement field, while the neural network correction captures local variations. The hierarchical scale model adaptively downweights severe displacement fitting errors, enabling more robust recovery of the latent mean displacement field. An alternating maximum-likelihood training strategy updates the mean through weighted residual minimization and updates the scales to adjust the loss weights. Systematic case studies across varying noise levels and observation resolutions demonstrate the robustness of PIE-PINN.
arXiv:2607.14744v1 Announce Type: new
Abstract: Surface electromyogram (sEMG) signals are widely used in human-machine interfaces for gesture recognition and user identification, but existing models often struggle to generalize across individuals due to subject-specific neuromuscular characteristics. This study introduces a disentanglement model that separates task-specific and subject-specific components from sEMG signals, thereby improving the generalization and interpretability of gesture recognition and user identification systems. Experimental results demonstrate that the disentangled components significantly improve the accuracy of both gesture classification and user identification across subjects and days, outperforming conventional methods under the same experimental conditions. Further analysis reveals that the task-specific components capture consistent activation patterns associated with the same gestures across individuals. In contrast, the subject-specific components reflect unique neuromuscular characteristics that can be used for user identification. Notably, the subject-specific components show lower similarity across days than the task-specific components, contributing to a greater decrease in user identification accuracy than in gesture recognition accuracy. These findings suggest that the disentanglement approach not only improves classification performance but also provides deeper insights into the physiological mechanisms underlying sEMG signals. The model's ability to isolate and interpret different neuromuscular components holds promise for enhancing the robustness of sEMG-based applications in real-world settings, including rehabilitation and user authentication. Our code is available at https://github.com/Open-EXG/HandDisentanglement.
arXiv:2607.14780v1 Announce Type: new
Abstract: Bio-nanoparticles are pivotal to next generation nanotherapeutics, but providing single-particle biomolecular characterization remains a crucial challenge. Herein we present ultra-violet direct absorption microscopy (UV-DAM) to tackle this challenge. UV-DAM is based on a tailored illumination scheme for absorption-only imaging, combined with a custom deep UV light source. Combined, they provide biomolecular specificity with single particle sensitivity. As such, UV-DAM provides rapid, label-free, high resolution, biochemical imaging. Enabled by these capabilities, we implement the classic nucleic acid:protein absorption assay at the single virus level where we demonstrate UV-DAM's ability to distinguish empty from DNA-loaded viral capsids based on bimolecularly specific absorption fingerprints. UV-DAM presents the translation from bulk UV-visible spectrometry to single-particle assessment, a crucial advancement for nanomedicine characterization where particle loading efficiencies are often heterogenous. Beyond this, UV-DAM is applicable to a wide range of nanomaterials or investigation of biological process with high spatio-temporal resolution and intrinsic molecular contrast.
arXiv:2607.15258v1 Announce Type: new
Abstract: The growing use of Bitcoin as a decentralized digital asset and investment tool has sparked strong interest in understanding its market behavior. This study presents a new approach to analyze Bitcoin market sentiment by combining on-chain and financial data with social media posts. Unlike models that aim to predict prices, this work focuses on explaining market sentiment using blockchain transactions, historical price data of Bitcoin, and daily Twitter sentiment classifications. The method merges sentiment trends with on-chain and financial metrics, normalized into a dataset for detailed market analysis. Multiple machine learning models were tested using cross-validation, with Gradient Boosting (XGBoost) emerging as the most reliable model for classifying sentiment, achieving an average F1-score of about 0.84. SHAP (SHapley Additive exPlanations), a game theory-based method for model interpretability, was used to quantify the contribution of on-chain features to the model's predictions, improving transparency. The results indicate that this data combination yields meaningful predictive signals and insights, supporting data-driven cryptocurrency analysis and future improvements with deep learning.
arXiv:2607.14915v1 Announce Type: new
Abstract: The p-type point-contact germanium detectors have emerged as the ideal detection technology for rare-event experiments such as direct dark matter searches and neutrinoless double beta decay, and have been verified to be capable of single-site spatial position resolution. Accurately characterizing the position-dependent pulse shape responses of the detector is a crucial prerequisite for deepening background understanding and achieving background reduction. Relying on an optimized cross-scanning localization method and a full-chain physical framework, this study extracted the pulse shape responses in critical regions of the CDEX detector, quantitatively evaluated its intrinsic spatial position resolution for the first time, and ultimately achieved the position tracing of real environmental backgrounds using the constructed pulse shape database. This study completely establishes a physical analysis closed-loop for spatial position resolution, providing critical theoretical and technical support for background analysis in future ton-scale arrays.
arXiv:2607.14932v1 Announce Type: new
Abstract: Synthetic face datasets have become effective enough to train face recognition models with accuracy rivaling that of models trained on real photographs. This progress sidesteps the ethical and legal burdens of collecting real biometric data, yet evaluation has not kept pace. Even studies that train entirely on synthetic images still rely on real-face benchmarks to measure performance, leaving the privacy problem only half solved. We ask whether synthetic datasets can replace real benchmarks for face recognition evaluation. We test 12 synthetic datasets against 7 established real benchmarks using 24 pre-trained models that span both convolutional and transformer architectures. Our evaluation covers biometric verification metrics, similarity score distributions, cross-model ranking consistency, and the underlying distributional properties of each dataset. Benchmarking fidelity varies widely across the synthetic candidates, but the two strongest, MorphFace and Vec2Face, reproduce the relative behavior of real benchmarks and reach agreement levels that fall within the natural disagreement already observed among the real benchmarks themselves. These results establish that well-constructed synthetic datasets can support reliable comparative evaluation for face recognition, moving the field closer to a fully synthetic and privacy-preserving pipeline for both training and benchmarking.
arXiv:2607.14654v1 Announce Type: new
Abstract: We give a new dual fitting algorithm which gives improved approximation ratios of $3+\ln 2 + \epsilon\ (\approx 3.694)$ and $4.9+\epsilon$ for $k$-Means in (high-dimensional) Euclidean and general metrics respectively, improving upon the previously known ratios of $4+\epsilon$ [Charikar, Cohen-Addad, Gao, Grandoni, Lee, and van Wijland STOC'26] and $5+\epsilon$ [Byrka, Guo, Hu, Li, Wan, Wang FOCS'26], resp. In particular, our result for Euclidean $k$-Means breaks the hardness barrier of $1+8/e\approx 3.94$ for Metric $k$-Means. Prior to our work, no such separation between general and Euclidean metrics was known for $k$-Median, $k$-Means, or Facility Location in terms of their approximability.
Unlike prior dual fitting approaches for $k$-Means, our new dual fitting algorithm tightly accounts for dual payments while still facilitating an effective dual feasibility analysis. We introduce a new framework that uses spectral analysis for determining the approximation factor of our algorithm.
arXiv:2607.14990v1 Announce Type: new
Abstract: When a camera moves fast during exposure, blur destroys the intra-exposure motion a 3D model needs to recover the sharp scene, while event cameras capture exactly this signal at microsecond resolution. Turning them into reliable 3D supervision faces two obstacles. First, the two restoration priors fail in opposite ways: physics-based event-integration priors preserve edges but accumulate drift; learned networks recover texture but distort boundaries. Second, existing pipelines run in one direction only, so raw event noise or the biases of fixed 2D pseudo-labels pass uncorrected into the geometry. JADE-GS addresses both: a pixel-adaptive routing gate fuses the complementary priors, and the resulting 2D restorer is coupled to a 3D Gaussian Splatting student in a bidirectional loop, where detached, multi-view-consistent renders and a physics-based reblurring constraint regularize the restorer, turning a fixed preprocessor into a geometry-aware predictor. Across synthetic and real benchmarks, JADE-GS attains the best perceptual quality, leading LPIPS and CLIP-IQA on both benchmarks with competitive PSNR and SSIM, and trainsin about one hour under 5 GB on a single consumer GPU while preserving real-time rendering.
arXiv:2607.15260v1 Announce Type: new
Abstract: What problems can one solve on a tournament if only its score sequence is known?
Tournaments are oriented complete graphs that form an extensively-studied class of directed graphs (digraphs), both from combinatorial and algorithmic perspectives. Over the years, researchers have identified multiple classical digraph problems that can be solved on a tournament from only its score sequence (indegree sequence). These problems include acyclicity testing and topological sorting [Chakrabarti, Ghosh, McGregor, and Vorotnikova; SODA'20], $s,t$-reachability, strong connectivity, and decomposition into strongly connected components (SCC) [Ghosh and Kuchlous; ESA'24], and vertex-ordering problems such as cutwidth and optimal linear arrangement [Barbero, Paul, and Pilipczuk; ICALP'17]. These prior works showed the sufficiency of the score sequence by designing distinct algorithms for the individual problems. In this work, we give a simple unified framework that solves all these problems using only indegrees and, in fact, completely characterises the class of problems that is determined by the indegree information: problems whose answers are invariant under cycle reversals. This characterisation is a special case of a much more general result that we establish: for any arbitrary digraph, the knowledge of its skeleton (underlying undirected graph) and the vertex indegrees completely determines its properties that are invariant under cycle reversal.
As a byproduct of our results, we obtain algorithms for a variety of connectivity-based, cut-based, and vertex-ordering problems on tournaments and ``almost tournaments'' in the streaming, the two-player communication, and the cut-query models of computation. Some of these algorithms match existing optimal bounds and others provide bounds improving the state of the art.
arXiv:2607.14105v1 Announce Type: new
Abstract: For Large Language Models to reliably answer user queries, users must clearly specify requirements, context, and constraints. In practice, however, user queries are often underspecified, forcing models to infer unstated assumptions that may misalign with the actual user intent. Existing prompt engineering guidelines aim to mitigate this issue, they are typically generic and task-agnostic, limiting their practical utility. Additionally, existing guidelines are formed manually and in a non-systematic way. To this end, we study prompt guideline optimization: the problem of automatically generating task-specific guidelines that help write better-specified prompts for a given task and model. Our key observation is that existing (completed) task examples (aka reference answers) often implicitly encode the missing information required to complete underspecified queries, including behavioral constraints, contextual assumptions, and evaluation criteria. We therefore propose AGOPS, an automatic approach that evolves task-specific guidelines via an optimization scheme that involves a prompt LLM writer, a solver LLM and prompt evolution, which maximize downstream effectiveness on a set of examples (user queries with reference answers). At inference time, our guidelines help users write well-specified prompts, boosting the effectiveness of LLMs. We show across mathematical reasoning, medical question answering, and coding tasks, that prompt underspecification leads to major drops (up to 95.3%) in downstream task performance (compared to well-specified prompts) and, perhaps more importantly, that this drop can hardly be recovered by existing prompt optimization techniques. Users following AGOPS guidelines can regain this loss (increasing performance between 15.5 to 81.7% on average) consistently across all benchmarks.
arXiv:2607.14681v1 Announce Type: new
Abstract: Recent diffusion-based video generation models have made significant progress in multi-reference image-conditioned video editing. However, existing methods still struggle to coordinate information from multiple visual sources accurately. We identify a critical deficiency in existing approaches. Existing editing instructions lack explicit reference relationships, and most multimodal large language models (MLLMs) cannot generate them reliably. To address this problem, we propose ReBind, a systematic framework that introduces semantic instructions with embedded reference tokens as the intermediate representation for multi-reference image-conditioned video editing. Our key insight is embedding reference tokens at semantic positions to eliminate ambiguity and establish precise bindings between visual attributes and their sources. We develop ReBind-Instruct, a specialized MLLM that learns to establish explicit bindings between visual attributes and their reference sources through a two-stage progressive scheme for precise reference relationships. We further develop ReBind-Edit, which enables lightweight adaptation of text-to-video models to coordinate multiple references by binding visual attributes to their designated sources. Extensive experiments demonstrate that ReBind substantially outperforms general-purpose MLLMs in instruction quality and achieves state-of-the-art performance among open-source methods on reference image conditioned video editing. Our project webpage: https://rebind-mrv2v.github.io/.
arXiv:2607.14201v1 Announce Type: new
Abstract: Amazonian earthworks, also called geoglyphs, are thousands of man-made earthen structures, mostly of geometrical shape, which progressively emerged from the tropical forest due to progressive deforestation. They were probably built between the fifth century BC and the end of the first millennium AD, but archaeological investigation on the culture of their builders is yet at the beginning. Nevertheless, a ceremonial rather than practical function seems likely, at least for those having a very regular shape. In the present paper, simple remote-sensing technique, combined with the methodological approach of modern Archaeoastronomy, are applied to study for the first time their orientation. The analysis takes in consideration virtually all known squared and rectangular structures for a total of 326 earthworks. The results show without doubts a non-random choice for their orientation and a clear interest of their builders for the annual cycle of the Sun.
arXiv:2607.14689v1 Announce Type: new
Abstract: We propose and analyze a novel numerical framework for the non-Markovian stochastic Schr\"odinger equation (NMSSE) based on a low-rank approximation of the bath correlation functions. By decomposing the memory kernel into a finite-dimensional representation, we derive a truncated system of hierarchical equations that effectively balances computational tractability with physical fidelity. A rigorous convergence analysis is established for the hierarchical framework under mild assumptions. We demonstrate that our formulation serves as a mathematical generalization of the Hierarchy of Pure States (HOPS), encompassing it as a special case while offering a more flexible representation of non-Markovian effects. Numerical experiments across several benchmark models are presented to illustrate the validity and efficacy of the proposed method.
arXiv:2607.15186v1 Announce Type: new
Abstract: Photonic integrated circuits (PICs) generate, route, and process light with high efficiency, scalability, and functional density on a single chip. Yet the tightly confined on-chip modes can not easily access or effectively interact with atomic vapors, fluids, gain media, and biological samples. Existing approaches require bringing the medium onto the chip or into a weak, tightly confined evanescent field, which restricts the interaction volume and the range of accessible media. Here, we demonstrate a re-entrant chip-free-space interface in which a thin-film lithium niobate circuit frequency-doubles telecom light, emits the 780~nm field through a Rubidium vapor cell, and recollects the reflected probe on the same chip. This emit-interact-recollect loop resolves the saturated absorption spectrum and stabilizes the telecom laser to within $\pm 280$~kHz over 2 hours. Our study paves an route to embed external media into PICs through the re-entrant photonic interface.
arXiv:2607.14725v1 Announce Type: new
Abstract: In robotic motion planning, equivariance to rigid body transformations is crucial for robust spatial generalization. However, current learning-based planners face a critical dilemma: they either lack inherent equivariance, treating transformed tasks as novel scenarios, or enforce it via computationally expensive specialized architectures that bottleneck real-time inference. To break this trade-off, we propose BridgeFlow, a fast and strictly SE(2)-equivariant generative motion planning framework. Rather than relying on heavy equivariant networks, BridgeFlow achieves exact spatial equivariance via a lightweight task-centric canonicalization module, enabling generalization using standard architectures. To further accelerate inference, we pair a Brownian bridge informative prior with context-aware mini-batch optimal transport. This constructs a straightened vector field that minimizes transport costs and stabilizes training. Furthermore, environmental awareness is explicitly embedded via Classifier-Free Guidance. Evaluations in dense 2D environments and on a 7-DoF Franka manipulator demonstrate that BridgeFlow achieves up to a 15x inference speedup and a 2x higher valid trajectory rate over state-of-the-art diffusion baselines, alongside robust generalization to entirely unseen environments and arbitrary spatial transformations.
arXiv:2607.14727v1 Announce Type: new
Abstract: Autonomous driving vision-language models (VLMs) struggle in roadwork zones, where familiar visual cues such as lane markings and permanent signs are altered or absent, and temporary devices such as cones and barriers redefine the drivable corridor. VLMs can detect these objects, but without explicit guidance they anchor their reasoning on familiar elements from pre-training and fail to connect work-zone observations to correct planning decisions. We propose WorkDrive, a framework that constructs perception-grounded causal reasoning for work zones and aligns it with trajectory prediction. An automated multitask perception pipeline extracts structured scene facts and injects them into a Chain-of-Causation (CoC) annotation pipeline, redirecting the annotator's attention to domain-specific elements. The resulting reasoning labels are used for supervised fine-tuning, followed by reinforcement learning with a single reward: consistency between lateral meta-actions and the predicted trajectory. On ROADWork, the largest public work-zone dataset, the proposed roadwork CoC reduces trajectory average displacement error (ADE) by 9.0\%, and consistency-based GRPO yields a further 3.0\%, achieving progressive improvement over the trajectory-only baseline. Code and data will be publicly released.
arXiv:2607.14728v1 Announce Type: new
Abstract: Tactile image generation significantly reduces the dependency on expensive and wear-prone sensors by synthesizing high-fidelity tactile data, offering an efficient solution for tactile information acquisition in robotic perception and human-machine interaction systems. However, existing methods depend on large-scale, diverse datasets from specific sensors and lack efficient data utilization and robust generalization capabilities, struggling in vision-limited environments. To address this, we introduce VQ-Touch, a tactile generation framework that supports both cross-sensor and multi-scenario applications. Specifically, to efficiently extract complex deformation and texture features from the data, we propose DM-VQGAN, an effective tactile representation learner. Furthermore, we introduce a discrete diffusion decoder with a unified conditioning interface, supporting multimodal generation tasks such as images and labels, and enhances the model's generalization capability through few-shot mixed training, thus achieving compatibility with current mainstream sensors and their variants. Experiments show that VQ-Touch surpasses state-of-the-art methods in multiple tasks.