arXiv:2508.05301v2 Announce Type: replace
Abstract: The real-time data collection and automation capabilities offered by the Internet of Things (IoT) are revolutionizing and transforming Business Processes (BPs) into IoT-enhanced BPs, showing high potential for improving sustainability. Although already studied in Business Process Management (BPM), sustainability research has primarily focused on environmental concerns. However, achieving a holistic and lasting impact requires a systematic approach to address sustainability beyond the environmental dimension. This work proposes a conceptual model and a structured methodology with the goal of analyzing the potential of IoT to measure and improve the sustainability of BPs. The conceptual model formally represents key sustainability concepts, linking BPM and IoT by highlighting how IoT devices support and contribute to sustainability. The methodology guides the systematic analysis of existing BPs, identifies opportunities, and implements sustainability-aware, IoT-enhanced BPs. The approach is illustrated through a running example from the tourism domain and a controlled case study in healthcare.
Science Journals
arXiv:2606.03453v1 Announce Type: new
Abstract: Vulnerability disclosure volumes now far exceed organizational assessment capacity, yet three adjacent research communities (proof-of-concept generation, vulnerability prioritization, and detection rule engineering) operate largely in isolation. Existing automated exploit generation systems report binary pass/fail outcomes, discarding partial progress and producing no signal for the other two communities. This paper presents FORGE, a multi-agent system that bridges these three silos through graduated exploitation depth. Five specialized agents (Intel, Generator, Planner, Exploit, and Detector) execute in a fixed pipeline that (1) generates targeted vulnerable applications from CVE metadata, (2) conducts coached, multi-turn exploitation assessed by an LLM-primary oracle on a four-level taxonomy (L0: no evidence through L3: full compromise), and (3) produces Sigma and Snort detection rules grounded in OpenTelemetry exploitation traces. Graduated depth is the bridging mechanism: deeper exploitation yields richer behavioral traces for detection engineering, while depth data across scoring bands provides ground truth for prioritization validation. A tiered knowledge architecture accumulates intelligence across assessments, transferring build and exploitation experience to subsequent CVEs. Evaluation on 603 CVEs from the CVE-GENIE dataset achieves 67.8% end-to-end L1+ exploitation at USD 1.50 per CVE across eight languages and 187 CWE types. Exploitation rates remain near 68% regardless of EPSS or CVSS band, indicating that pattern-level reachability is orthogonal to metadata-based prioritization. Detection rules from L2+ exploitation achieve significantly higher span-normalized grounding than L1-derived rules (p=0.035), and 93.4% of generated Snort rules produce zero false positives against a synthetic benign corpus.
arXiv:2606.03387v1 Announce Type: new
Abstract: Smart contract vulnerabilities in Decentralized Finance (DeFi) protocols resulted in over 1.49 billion USD in confirmed losses in 2024 alone, across 192 incidents [1]. As LLM-based vulnerability detection emerges as a promising approach to address these threats, the quality of evaluation datasets has become a critical bottleneck. Existing datasets suffer from three fundamental problems: they are built on outdated Solidity versions (e.g., v0.4) that no longer reflect modern DeFi contracts [5][6][7]; they rely on automated or LLM-generated annotations that introduce hallucination-driven label noise [9][10]; and they apply coarse single-layer labeling that fails to capture the semantic complexity of real-world business logic vulnerabilities [6][7][11][12]. We present Bastet, an expert-labeled DeFi smart contract vulnerability dataset that addresses all three problems through real-world audit findings (2021-2024), human expert annotation with discussion-based consensus, and a two-layer taxonomy of 46 Tags and 77 Subtags. Bastet comprises 4,402 findings collected from 394 Code4rena competitive audit reports spanning April 2021 to November 2024, of which 849 findings are fully annotated by white-hat security researchers from the DeFiHackLabs community. All annotations are produced through a two-annotator consensus workflow, ensuring label accuracy grounded in real-world vulnerability root causes.
arXiv:2511.04421v2 Announce Type: replace
Abstract: Action chunking is a widely adopted approach in Learning from Demonstration (LfD). By modeling multi-step action chunks rather than single-step actions, action chunking significantly enhances modeling capabilities for human expert policies. However, because action chunking makes a single decision only after a complete action block has been executed, the resulting reduction in decision frequency restricts the utilization of real-time observations, impairing reactivity in dynamic or noisy environments. Existing efforts to address this issue have primarily resorted to trading off reactivity against decision consistency, without achieving both. To address this limitation, we propose a novel algorithm, Temporal Action Selection (TAS), which caches predicted action chunks from multiple timesteps and dynamically selects the optimal action through a lightweight selector network. TAS achieves balanced optimization across both reactivity and decision consistency. Experiments across multiple tasks with diverse base policy architectures show that TAS significantly improves success rates. Furthermore, integrating TAS as a base policy with residual reinforcement learning (RL) improves both training efficiency and the performance ceiling. Experiments in both simulation and physical robots confirm the method's efficacy.
arXiv:2605.31067v2 Announce Type: replace
Abstract: Open-set task execution can significantly benefit from seamlessly switching between coarse and fine scene representations depending on the context and the evolving information as the robot explores the environment. For example, it is often sufficient to start with a coarse scene representation initially and only employ a finer, more granular scene representation when the robot encounters regions which are likely to contain the task relevant objects. Hence, in this work, we propose BiMoSG, a bimodal 3D scene graph generation approach for open-set tasks. BiMoSG employs a "fast" mode by default to efficiently generate a coarse 3D scene graph and can switch to a "slow" mode for generating a finer open vocabulary 3D scene graph of task relevant objects. We demonstrate that our proposed 3D scene graph generation approach is significantly faster than the open-source state-of-the-art approaches. This allows us to integrate the scene graph generation process with task execution for real-time deployment.
arXiv:2606.03361v1 Announce Type: new
Abstract: Rubric-based rewards are increasingly used for open-ended language model post-training, but criterion-level scores are often aggregated as independent utilities. This flat scalarization ignores rubric-specified prerequisite and activation relations among criteria, allowing reward or penalty to be counted even when the condition that licenses it is absent. We call this structural reward-aggregation failure \textbf{False Credit Propagation} (FCP). To address this limitation, we propose \ourname (\textbf{G}raphical \textbf{E}vent \textbf{A}ggregation for \textbf{R}ubric rewards), a probabilistic graphical framework for dependency-aware rubric aggregation. \ourname models each criterion outcome as a latent Bernoulli event in a typed rubric graph, propagates soft suppression from unsupported parent events to their children, and aggregates the resulting event probabilities into a normalized expected signed utility. This yields a linear-time reward computation that can be plugged into standard rubric-based RL pipelines without changing the outer optimization algorithm. Experiments on HealthBench, WritingBench, and PLawBench with two policy backbones show that \ourname consistently improves over flat aggregation and deterministic gating, achieving relative gains of up to 15.5\% over flat aggregation. FCP diagnostics further show that \ourname reduces leakage by 96.5\% relative to flat aggregation while preserving more licensed downstream utility than deterministic gating. Our code is publicly available at https://github.com/LvCan926/GEAR.
arXiv:2606.03359v1 Announce Type: new
Abstract: Speech emotion recognition is an important component of modern human-computer interaction systems. However, many state-of-the-art approaches rely on large pretrained models with high computational and memory requirements, limiting their applicability. This paper proposes ResLSTM-SA, a lightweight architecture that integrates residual connections with soft attention within an LSTM-based framework. Evaluated on the RAVDESS dataset under strict speaker-independent partitioning, the proposed model outperforms conventional attention-based LSTM baselines and several previously reported CNN- and hybrid CNN-LSTM architectures in terms of unweighted average recall (UAR). The best-performing variant (ResLSTM-SA-h64) achieves a maximum UAR of 0.6517 with only 46.8k trainable parameters, delivering competitive accuracy with three orders of magnitude fewer parameters than large-scale self-supervised alternatives, thereby enabling efficient deployment on edge devices and real-time voice assistants. The source code is available at https://github.com/Mak-Sim/ResLSTM-SER.
arXiv:2511.02417v2 Announce Type: replace
Abstract: The adoption of agroecological practices in modern agriculture requires robotic systems capable of operating in highly diverse and complex field environments. Developing and evaluating such systems relies heavily on simulation, yet generating realistic and configurable 3D environments representative of agroecological diversity remains a major challenge. This paper presents CropCraft, an open-source procedural world generator built on Blender and Python, designed to produce 3D simulation environments tailored to agricultural robotics. CropCraft generates crop fields from a simple YAML configuration file, supporting a wide range of scenarios including intercropping, vineyards, and weed-infested fields. The tool includes a library of 3D plant models (crops, grasses, and weeds) at multiple growth stages, and uses stochastic placement algorithms to realistically reproduce the spatial variability observed in real fields. Generated worlds are directly importable into the Gazebo simulator and include ground-truth annotations for all placed elements, supporting both perception and navigation algorithm development. To demonstrate the practical utility of CropCraft, we apply it to the task of crop-weed semantic segmentation using deep learning. A dataset of 10,000 synthetic images of maize fields with varying weed densities, growth stages, and lighting conditions was generated and used to train several segmentation architectures. Models trained exclusively on synthetic data achieve a sim-to-real gap of approximately 10% mean Intersection over Union (mIoU) on real field images, outperforming previous state-of-the-art synthetic generation approaches. We further show that combining even a few real images with synthetic data improves generalization across domains, providing new insights into the effective use of synthetic data for agricultural perception tasks.
arXiv:2606.03289v1 Announce Type: new
Abstract: Cloud permission governance implicitly treats permission equivalence as a static relation. We show that for non-human identities (NHIs), equivalence has two irreducible components: structural equivalence, capturing identical permission profiles at a snapshot via graph fibration, and temporal equivalence, capturing recurring permission states via strongly connected components (SCCs) in a fiber transition graph. We call the equivalence classes under temporal equivalence privilege circuits.
We formalize a three-layer framework: (1) a spatial quotient of the permission graph via fibration, (2) a lineage partition organizing stable transition compartments, (3) windowed SCC analysis as a temporal quotient within lineages.
Empirical evaluation on a large Azure tenant supports the framework. Backtesting demonstrates that early observation of ratchet-type privilege circuits predicts long-term structural stability.
arXiv:2606.02644v1 Announce Type: new
Abstract: Agentic scaffolds have dramatically improved LLM performance on complex, long-horizon tasks, yielding both broad benefits and amplified risks in domains like cybersecurity. Existing benchmarks for AI agents in cybersecurity focus mainly on measuring proficiency--how effectively agents can complete offensive security tasks--but neglect a critical question: when and how should agents refuse harmful requests? We present the first framework for establishing refusal boundaries in offensive security contexts. Our framework defines (1) principled criteria for when tasks should be refused, (2) categories of tasks that warrant refusal, and (3) evaluation methodology for measuring agent robustness under both benign and adversarial conditions. We apply this framework to assess how current LLM-powered agents adhere to appropriate refusal boundaries across a range of web-based offensive security scenarios, finding that 6 of 8 frontier models tested show near-zero refusal rates, with only 2 models (GPT-5.2 and GPT-5.1 Codex) demonstrating any meaningful refusal behavior.
arXiv:2604.25835v2 Announce Type: replace
Abstract: The Jiangmen Underground Neutrino Observatory (JUNO) is a 20-kton liquid scintillator-based, low-radioactivity, multi-purpose neutrino detector located 693 meters (1800 m.w.e.) underground in the Guangdong province, China. To detect scintillation light produced in the target, the detector is equipped with 17,612 20-inch photomultipliers (PMTs), forming the Large PMT system (LPMT). In addition, 25,600 3-inch photomultipliers (the Small Photomultiplier System or SPMT) are deployed in the gaps between the LPMTs.
This paper presents the design and performance of the underwater front-end electronics developed for the SPMT system. It details the individual electronics boards and their key components, the inter-board interfaces, the system-level design, and the firmware architecture that supports data acquisition and control. It also outlines mechanical and thermal integration, board validation procedures, and system performance metrics. The readout chain includes digitization of 128 PMT channels per unit, synchronized time-stamping, charge measurement, event packaging, and bandwidth management. Comprehensive validation confirms the system's readiness to meet JUNO's stringent physics goals. The underwater electronics achieve noise levels as low as 0.04 photoelectrons with minimal crosstalk (below 0.4%) and a bandwidth of 57 MB/s, ensuring reliable single photo-electron detection and operation under high-rate conditions. The SPMT system has now been fully integrated and installed in JUNO. Its commissioning and physics performance will be reported in a future publication.
arXiv:2511.19208v3 Announce Type: replace
Abstract: A Locally Checkable Labeling (LCL) is a distributed constraint satisfaction problem defined on a bounded-degree graph that relates a finite set of input labels to a finite set of output labels through a finite set of locally checkable constraints. In this work we define labels and local constraints that encode solutions to two classical problems: leader election and spanning tree construction. It is known that leader election cannot be expressed as an LCL in arbitrary graphs using constant-size labels. In fact, it is known that there does not exist a finite set of labels and local constraints for leader election even for the class of rings. On the other hand, there exists a finite set of labels and local constraints characterizing leader election on trees. In this work, we prove that there exists a finite set of labels and local constraints for leader election also in the much larger class of dismantlable graphs. Our labels need one bit per edge or equivalently $O(\Delta)$ bits per node (where $\Delta$ is the maximum degree in the graph) and are checkable within the graph induced by the 1-neighborhood of each node. To the best of our knowledge, these are the first local labeling results tailored to dismantlable graphs, potentially highlighting structural properties useful for designing labels and constraints for additional LCL problems. Finally, we present a generic transformation that converts any finite set of labels and local constraints into a silent self-stabilizing algorithm by adding only one extra state, assuming a Gouda fair scheduler. This transformation may be of independent interest.
arXiv:2606.03326v1 Announce Type: new
Abstract: Compliance pipelines detect violations as transient query results and do not keep the violation itself as a persistent graph object with review state, affected entities, or audit history. The Violation Situation Pattern (VSP) closes this gap. Building on the Situation pattern of Gangemi and Mika, VSP reifies each detected violation as a graph node with a rule identifier, a temporal validity interval, a lifecycle state, and evidence links to the entities involved. Lifecycle transitions are stored as immutable, PROV-O-aligned events, so audit history is a graph traversal. We instantiate VSP in a legal entity and contract lifecycle property graph and operationalize four deontic rules (V1 unauthorized signature, V2 expired mandate, V3 missing confidentiality clause, V4 missing breach-notification clause) through an FCL->Cypher->MERGE pipeline. We check V1 and V2 against BODACC corporate-officer publications, evaluate V4 on 73 GDPRhub enforcement decisions, and run a SHACL cross-formalism check on V3 and V4. The central finding is rule-body independence: extending V4 from clause-presence to deadline checking raises F1 from 0.312 to 0.602, while the pattern's identity, lifecycle, and evidence semantics stay the same. This separates a pattern contribution from a detector contribution, so detection logic can evolve without invalidating accumulated audit history.
arXiv:2506.02018v2 Announce Type: replace
Abstract: Paraphrasing re-expresses meaning to enhance applications like text simplification, machine translation, and question-answering. Specific paraphrase types facilitate accurate semantic analysis and robust language models. However, existing paraphrase-type generation methods often misalign with human preferences due to reliance on automated metrics and limited human-annotated training data, obscuring crucial aspects of semantic fidelity and linguistic transformations.
This study addresses this gap by leveraging a human-ranked paraphrase-type dataset and integrating Direct Preference Optimization (DPO) to align model outputs directly with human judgments. DPO-based training increases paraphrase-type generation accuracy by 3 percentage points over a supervised baseline and raises human preference ratings by 7 percentage points. A newly created human-annotated dataset supports more rigorous future evaluations. Additionally, a paraphrase-type detection model achieves F1 scores of 0.91 for addition/deletion, 0.78 for same polarity substitution, and 0.70 for punctuation changes.
These findings demonstrate that preference data and DPO training produce more reliable, semantically accurate paraphrases, enabling downstream applications such as improved summarization and more robust question-answering. The PTD model surpasses automated metrics and provides a more reliable framework for evaluating paraphrase quality, advancing paraphrase-type research toward richer, user-aligned language generation and establishing a stronger foundation for future evaluations grounded in human-centric criteria.
arXiv:2606.02600v1 Announce Type: cross
Abstract: We study autoencoder and variational-autoencoder latent spaces through the lens of spin-glass theory. The paper has two components. First, we formalize a latent-space spin-glass dictionary: for a fixed decoder, the reconstruction term together with a hyperspherical coordinates prior induces a Hamiltonian on the latent sphere, where latent coordinates play the role of continuous spins and the prior acts as an external magnetic field. This allows us to import operational spin-glass diagnostics -- overlap distributions, susceptibility, and block-spin coarse-graining -- to detect ordered, disordered, and edge-of-stability phases in trained latent representations. Second, we show that deliberately driving the latent system toward the edge-of-stability of the topological trivialization regime has concrete downstream consequences. In generation, hyperspherical compression improves the reconstruction-generation trade-off on CIFAR-10 and CelebA64, yielding lower self-FID while preserving or improving reconstruction. In anomaly detection, the same semi-ordered latent geometry improves both fully unsupervised and conditional OOD detection, including real-world Mars Rover and Galaxy Zoo datasets, as well as CIFAR-10/100 and Imagenette-based OOD benchmarks. We therefore advocate a phase-aware evaluation paradigm for AEs/VAEs, in which spin-glass observables complement standard ML metrics and expose the latent regimes that underlie downstream success or failure in many cases.
arXiv:2604.07366v2 Announce Type: replace
Abstract: Partial differential equations (PDEs) govern nearly every physical process in science and engineering, but solving them at scale remains prohibitively expensive. Generative AI has transformed language, vision, and protein science, but learned PDE solvers have not undergone a comparable shift. Existing paradigms each capture part of the problem. Physics-informed neural networks embed residual structure, although they are often difficult to optimize in stiff, multiscale, or large-domain regimes. Neural operators amortize across instances, although they commonly inherit a snapshot-prediction view of solving and can degrade over long rollouts. Diffusion-based solvers model uncertainty, although they are often built on a solver template that still centers on state regression. We argue that the core issue is the abstraction used to train learned solvers. Many models are asked to predict states, while many scientific settings require modeling how uncertainty moves through constrained dynamics. The relevant object is transport over physically admissible futures. This motivates flow learners: models that parameterize transport vector fields and generate trajectories through integration, echoing the continuous dynamics that define PDE evolution. This physics-to-physics alignment supports continuous-time prediction, native uncertainty quantification, and new opportunities for physics-aware solver design. We explain why transport-based learning offers a stronger organizing principle for learned PDE solving and outline the research agenda that follows from this shift.
arXiv:2603.29123v3 Announce Type: replace
Abstract: The next-token prediction (NTP) objective trains language models to predict a single token at each step, even though many continuations can express the same meaning. For example, in the sentence ``this sticker can be placed here'', positioned, attached, or put are all plausible alternatives. While standard NTP training treats these alternatives as mutually exclusive targets, we explore a self-supervised framework that encourages models to predict concepts, approximated as sets of semantically equivalent tokens. Models trained with this concept supervision align better with human similarity judgments, improve classification, clustering, and reranking performance, and achieve comparable or stronger downstream reasoning. These gains come with lower perplexity on semantically meaningful words (Section 3.2) and only minimal increases in global perplexity, suggesting that concepts enhance semantic alignment while preserving language modeling quality. Our code is available at https://anonymous.4open.science/r/learning-concepts-9025 .
arXiv:2603.27455v2 Announce Type: replace
Abstract: In this paper, we introduce NAS3R, a self-supervised feed-forward framework that jointly learns explicit 3D geometry and camera parameters with no ground-truth annotations and no pretrained priors. During training, NAS3R reconstructs 3D Gaussians from uncalibrated and unposed context views and renders target views using its self-predicted camera parameters, enabling self-supervised training from 2D photometric supervision. To ensure stable convergence, NAS3R integrates reconstruction and camera prediction within a shared transformer backbone regulated by masked attention, and adopts a depth-based Gaussian formulation that facilitates well-conditioned optimization. The framework is compatible with state-of-the-art supervised 3D reconstruction architectures and can incorporate pretrained priors or intrinsic information when available. Extensive experiments show that NAS3R achieves superior results to other self-supervised methods, establishing a scalable and geometry-aware paradigm for 3D reconstruction from unconstrained data. Code and models are publicly available at https://ranrhuang.github.io/nas3r/.
arXiv:2504.01250v2 Announce Type: replace
Abstract: This paper presents the Robust Recurrent Deep Network (R2DN), a scalable parameterization of robust recurrent neural networks for machine learning and data-driven control. We construct R2DNs as the feedback interconnection of a linear time-invariant system and a 1-Lipschitz deep feedforward network, and directly parameterize the weights so that our models are stable (contracting) and robust to small input perturbations (Lipschitz) by design. Our parameterization uses a structure similar to the previously-proposed recurrent equilibrium network (REN), but without the requirement to iteratively solve an equilibrium layer at each time-step. This speeds up both model inference and backpropagation on GPUs, and makes it computationally feasible to scale up the network size, batch size, and input sequence length in comparison to RENs. We compare R2DNs to RENs on three representative problems in nonlinear system identification, observer design, and learning-based feedback control. We find that training and inference are both up to an order of magnitude faster with similar test set performance, and that they scale more favorably with respect to model expressivity.
arXiv:2606.03965v1 Announce Type: new
Abstract: Large language models improve final-answer accuracy through extended chain-of-thought reasoning, but often spend tokens inefficiently and offer little inference-time control. Existing efficient reasoning methods control thinking length by shortening, early-stopping, or compressing traces, leaving how the model thinks implicit. In this paper, we propose Agentic Chain-of-Thought Steering (ACTS), which formulates reasoning steering as a Markov decision process where a controller agent adaptively steers a frozen reasoner during inference. At each step, the controller observes the reasoning trace and remaining thinking budget, then issues a steering action consisting of a reasoning strategy and a steering phrase that initiates the next reasoner step. This enables budget-aware strategy control for efficient reasoning while preserving the reasoner's generation continuity. We initialize the controller agent from our constructed synthetic steering trajectories with multi-budget augmentation, and further optimize it via reinforcement learning with budget-conditioned reward shaping. Experiments across multiple benchmarks show that ACTS matches full-thinking performance with substantial token savings, and enables controllable accuracy-efficiency trade-offs across different reasoners and tasks. The code is available at https://github.com/Andree-9/ACTS.
arXiv:2606.03391v1 Announce Type: new
Abstract: Model merging has emerged as a cost-effective approach for consolidating the capabilities of multiple LLMs without retraining. However, existing merging techniques, largely based on linear parameter arithmetic or optimization, struggle when applied to Mixture-of-Experts (MoE) architectures. We identify a critical failure mode in MoE merging, termed routing breakdown, in which the merged router fails to dispatch tokens to suitable experts. Routing breakdown stems from the sensitivity of the non-linear softmax and discrete Top-k routing mechanisms to parameter perturbations from merging, a sensitivity further amplified by load-balancing constraints imposed during MoE pretraining. Because fine-tuned experts exhibit distinct specializations, even modest misrouting can cause severe performance degradation. To address this issue, we propose Hessian-Aware Router Calibration (HARC), a training-free framework that leverages second-order curvature information to realign the merged router. This approach admits a closed-form solution that can be efficiently solved using a matrix-free conjugate gradient method. Experiments on mathematical reasoning and code generation tasks show that HARC effectively mitigates routing breakdown across diverse MoE merging baselines and leads to substantial performance improvements. Our code is available at https://github.com/huangcb01/HARC.
arXiv:2606.02916v1 Announce Type: new
Abstract: Large-scale graph neural network (GNN) training often requires distributed clusters because graph structure and feature tensors no longer fit in a single node's memory. In sampling-based training, each mini-batch expands into a receptive field that spans partitions and triggers thousands of remote feature fetches per epoch. This wastes energy for two main reasons: each small RPC pays a fixed initiation and protocol cost, and GPUs continue drawing substantial baseline power while waiting for remote features. We present GreenGNN, an energy-aware distributed GNN training system that reduces communication energy by exploiting the bursty, short-lived temporal locality of neighbor sampling. GreenGNN groups training into windows of W consecutive mini-batches, stages each window's hot features in a local cache, and merges remote requests from each partition owner into a small number of bulk transfers. This amortizes RPC overhead across many features while preserving an on-demand path for cache misses. Because window size controls the trade-off between communication amortization and hot-set staleness, GreenGNN selects W offline using a discrete-event simulator that replays a deterministic one-epoch access trace with a hybrid energy model. We implement GreenGNN on DGL and evaluate it on a 4-node GPU cluster with benchmark datasets. Across datasets and batch sizes, GreenGNN reduces total system energy by 27--43% relative to baseline while improving end-to-end throughput by up to 3.9x. GPU energy drops by 36--71%, driven by fewer RPC initiations and lower GPU stall time.
arXiv:2411.15851v2 Announce Type: replace
Abstract: While vision-language models like CLIP have shown remarkable success in open-vocabulary tasks, their application is currently confined to image-level tasks, and they still struggle with dense predictions. Recent works often attribute such deficiency in dense predictions to the self-attention layers in the final block, and have achieved commendable results by modifying the original query-key attention to self-correlation attention, (e.g., query-query and key-key attention). However, these methods overlook the cross-correlation attention (query-key) properties, which capture the rich spatial correspondence. In this paper, we reveal that the cross-correlation of the self-attention in CLIP's non-final layers also exhibits localization properties. Therefore, we propose the Residual Cross-correlation Self-attention (RCS) module, which leverages the cross-correlation self-attention from intermediate layers to remold the attention in the final block. The RCS module effectively reorganizes spatial information, unleashing the localization potential within CLIP for dense vision-language inference. Furthermore, to enhance the focus on regions of the same categories and local consistency, we propose the Semantic Feedback Refinement (SFR) module, which utilizes semantic segmentation maps to further adjust the attention scores. By integrating these two strategies, our method, termed ResCLIP, can be easily incorporated into existing approaches as a plug-and-play module, significantly boosting their performance in dense vision-language inference. Extensive experiments across multiple standard benchmarks demonstrate that our method surpasses state-of-the-art training-free methods, validating the effectiveness of the proposed approach. Code is available at https://github.com/yvhangyang/ResCLIP.
arXiv:2606.03338v1 Announce Type: new
Abstract: Self-supervised learning (SSL) has emerged as a powerful paradigm for learning meaningful representations from unlabeled data. However, the standard protocol for evaluating these representations, linear probing, is computationally expensive, sensitive to hyperparameters, and provides limited insight into the geometric structure of the representation space. In this work, motivated by connections between neural network generalization and intrinsic dimension (ID) we propose IdEst, a method for estimating the ID of SSL representations via the Minimum Spanning Tree dimension estimator ($\mathrm{dim}_\mathrm{MST}$). Across diverse datasets, architectures, and SSL pretraining objectives, we show that IdEst strongly correlates with downstream linear probe performances. Furthermore, we demonstrate that IdEst enables efficient hyperparameter selection, significantly reducing the computational cost compared to supervised alternatives. Our results highlight intrinsic dimensionality as a principled geometric proxy for assessing SSL representations, complementing standard supervised probing protocols.
arXiv:2606.03335v1 Announce Type: new
Abstract: Large scale GPU-parallel reinforcement learning has changed what can be trained in robot simulation, yet most systems still optimize one specialist policy per task. We propose a construction methodology for turning structured manipulation task families into GPU-parallel multi-task RL benchmarks, and instantiate it as MT-Libero using LIBERO assets and task predicates in Isaac Lab. The resulting benchmark supports simultaneous reinforcement learning over heterogeneous task suites with parallel rendering, physics randomization, and state-input or visual-input policies. To make such training practical under sparse success signals and limited prior data, we further propose DGPO, an on-policy demonstration guided method that combines importance weighted PPO with adaptive behavior cloning on matched demonstration actions. DGPO enables a tunable preference toward demonstrated task distributions, outperforming both prior-free RL and existing demonstration-based methods while preserving the stability and online improvement benefits of on-policy PPO.