arXiv:2506.08359v3 Announce Type: replace
Abstract: Inference-time steering aims to alter a large language model's (LLM's) responses without changing its parameters, but a central challenge is identifying the internal modules that most strongly govern the target behavior. Existing approaches often rely on simplistic cues or ad hoc heuristics, leading to suboptimal or unintended effects. We introduce REAL, a framework for identifying behavior-relevant modules (attention heads or layers) in Transformer models. For each module, REAL trains a vector-quantized autoencoder (VQ-AE) on its hidden activations and uses a shared, learnable codebook to partition the latent space into behavior-relevant and behavior-irrelevant subspaces. REAL quantifies a module's behavioral relevance by how well its VQ-AE encodings discriminate behavior-aligned from behavior-violating responses via a binary classification metric; this score guides both module selection and steering strength. We evaluate REAL across eight LLMs from the Llama and Qwen families and nine datasets spanning truthfulness enhancement, open-domain QA under knowledge conflicts, and general alignment tasks. REAL enables more effective inference-time interventions, achieving an average relative improvement of 20% (up to 81.5%) over the ITI method on truthfulness steering. In addition, the modules selected by REAL exhibit strong zero-shot generalization in cross-domain truthfulness-steering scenarios.
Science Journals
arXiv:2607.11444v1 Announce Type: new
Abstract: Mixture-of-Experts (MoE) models scale capacity without proportional compute cost and have become a key architecture for frontier large language models (LLMs). Yet domain-specific post-training inherits an expert pool shaped by mixed-domain pre-training: a substantial subset of experts contributes little on the target domain, and standard supervised fine-tuning (SFT) leaves the composition of this pool unchanged. We propose a simple, budget-preserving pipeline that realigns the expert pool to the target domain before fine-tuning. Given a target domain, we (1) prune the experts with lowest domain-aligned saliency, (2) regrow the expert pool to its original size through perturbation-based expert expansion, and (3) apply standard SFT. The resulting model preserves the original expert count, parameter count, and inference cost. With a single frozen recipe and no per-domain hyperparameter tuning, UMoE consistently improves over direct sft across two MoE architectures (Qwen3-30B-A3B and Qwen3.5-35B-A3B), five domains (math, code, science, tool-use, and agentic coding), and 12 benchmarks. Representative improvements are 3.4 points in math average accuracy, 6.0 points on SWE-bench Verified. On a strong in-house math corpus, direct sft already surpasses Qwen3-30B-A3B-Thinking (82.81 vs.\ 81.06), yet UMoE further raises the average to 84.17, an additional 1.36 points, demonstrating robustness to a substantially stronger SFT regime. Data-scaling experiments further show that the gain persists as training data grows. Analysis reveals that the direct-SFT model allocates substantial routed-expert compute to a low-saliency subset that can be removed post hoc with little average degradation; UMoE turns this redundant capacity into useful domain capacity and achieves lower training loss, with gains spanning all difficulty levels in downstream evaluation.
arXiv:2607.10395v1 Announce Type: new
Abstract: High-quality alpha mattes are notoriously expensive to annotate, creating a fundamental data bottleneck for deep image matting. While prior work attempts to reduce annotation cost using coarser labels like trimaps or masks, they remain reliant on costly per-pixel supervision, limiting scalability and generalization. In this work, we push the boundary further and ask: can we train an automatic matting model using only RGB images, with no manual annotation at all? We answer this by presenting SSMatte, a self-supervised framework that for the first time achieves performance on par with fully-supervised automatic matting. Our key insight is to decompose the problem into semantic anchoring and detail matting. SSMatte first generates a semantic matting prompt from frozen self-supervised ViT features by propagating class-token seeds via a novel, training-efficient semantic anchoring loss based on a generalized Rayleigh quotient. This prompt then anchors a detail matting network, which is optimized via a fixed-point-based loss that enforces alpha-RGB consistency. Extensive experiments show SSMatte outperforms prior weakly-supervised methods, matches the performance of fully-supervised models on portrait benchmarks, and demonstrates favorable scaling and generalization behaviors with additional data. Our work pushes automatic matting to an fresh, fully annotation-free paradigm. Code will be available.
arXiv:2607.10401v1 Announce Type: new
Abstract: In the last decade, data-driven computational mechanics (DDCM) has emerged as a novel paradigm in computational mechanics, enabling the direct use of constitutive data - such as stress-strain pairs obtained from experiments, without relying on ad-hoc material models and thereby avoiding information loss. In this work, we extend our data-driven solving strategy GO-ADM, which combines a greedy optimization algorithm with the alternating direction method (ADM), to the structural analysis of geometrically exact beams formulated using director-based kinematics. We discuss a data initialization strategy for nonlinear systems based on a conventional finite element analysis of the same structure using a prescribed constitutive model. The resulting discrete stress and strain fields, possibly obtained under multiple loading scenarios, may also be employed as artificial datasets for the subsequent data-driven computations. Furthermore, we investigate the thermomechanical consistency of both the dataset and the discrete solution, and propose a weak enforcement of this consistency in the latter via a penalty approach. Numerical examples involving single- and multi-member structures demonstrate that the proposed penalty term leads to thermomechanically consistent discrete stress and strain fields. Moreover, for the studied examples, the solving strategy GO-ADM yields a generally improved approximation of the globally optimal solution compared to the standard ADM-based direct solver.
arXiv:2607.09748v1 Announce Type: new
Abstract: In distributed systems, the classical State Machine Replication (SMR) model assumes that correct replicas execute deterministic transitions to yield identical bitwise states. However, the rise of agentic distributed systems -- where autonomous, stochastic, and model-driven agents orchestrate infrastructure -- presents scenarios where deterministic, bitwise replication is insufficient. Replicas operating with generative models may exhibit divergent reasoning paths, summaries, and token boundaries, yet reach semantically equivalent and correct operational decisions. Forcing bitwise agreement across these stochastic participants degrades execution flexibility, induces context amnesia, and limits performance.
We argue that in such settings replicas should agree on belief, not bits. We propose Epistemic State Replication (ESR), a belief-replication layer for agentic distributed systems that shifts the replication boundary from data visibility to knowledge visibility. We formalize the epistemic node state as a pair K = (L, B) separating the deterministic, immutable evidence log (L) from the stochastic, evolving belief lineage (B). To govern execution safety, we define Semantic Linearizability, which requires operations to reflect the latest committed operational meaning within a verifier-bounded semantic compatibility metric, and Bounded Eventual Coherence, which bounds expected semantic divergence under fair delivery, monotonic evidence, bounded verifier disturbance, and a contractive graft operator. We outline protocols for propagating derived insights using structured epistemic deltas, and formalize Verifiable Semantic Rollbacks to prune faulty premises from belief lineages without inducing context amnesia. We prototype ESR and report preliminary simulation results that show feasibility under the stated assumptions and illustrate reductions in secondary cognitive faults.
arXiv:2607.10243v1 Announce Type: new
Abstract: At the limit of handling, a stabilizing MPC depends on the yaw-rate reference it tracks and the stable-handling envelope it enforces, both operating-point-dependent and unknown a priori, so fixed or worst-case settings are either too conservative or unsafe. We learn this uncertainty with a conditional diffusion residual model and apply it to the controller's reference and constraints rather than its control law. Conditioned on the steering command, the model returns the residual's mean and a predictive spread: the mean re-sizes the tracked yaw reference, while the spread, propagated over the prediction horizon, tightens the stable-handling envelope through a one-sided chance back-off. Together these form the proposed diffusion-residual MPC (D-res), so caution is anticipated ahead of the tracking error rather than corrected after it by a high-gain loop. Because only two moments per command are needed, the generator is tabulated offline and the online controller adds a single table lookup to the baseline MPC, with no in-loop diffusion; it runs within the 100 Hz budget on an NVIDIA Jetson AGX Xavier (worst-case 4.08 ms per step). Across a 7-DOF model and high-fidelity CarMaker co-simulation spanning vehicle, tire, road, and maneuver diversity, D-res reduces peak side-slip where the fixed bicycle model is least accurate and restores directional stability on low-friction maneuvers, where the fixed reference over-commands the available grip.
arXiv:2607.09835v1 Announce Type: new
Abstract: The recently introduced YOLO26 architecture incorporates NMS-free end-to-end inference and is optimized for deployment on resource-constrained CPU-based devices, making it well-suited for edge-based aquaculture applications. However, its performance, operational efficiency, and deployment suitability have not been systematically validated in aquaculture-specific scenarios. This study presents a comprehensive benchmark of YOLO26 against three Ultralytics predecessors (YOLOv5u, YOLOv8, and YOLO11) across nano, small, and medium model scales for fish mortality detection, a critical indicator of fish population health and welfare. Twelve model variants were evaluated for detection accuracy, training efficiency across seven dataset sizes, and inference performance on high-performance NVIDIA A100 GPUs and a CPU-only Raspberry Pi 5 edge platform. All models achieved comparable performance on the full dataset, with mAP50 differing by only 1.04 percentage points, indicating that architectural generation has little influence on final detection accuracy when sufficient training data are available. However, clear trade-offs emerged in data efficiency and deployment performance. YOLOv8 achieved 90% mAP50 with only 400 training images, whereas the YOLO26 nano and small variants required 1,000 images to reach comparable accuracy. Conversely, YOLO26n achieved the highest inference speed on the Raspberry Pi 5 (7.51 FPS), while YOLOv5mu outperformed all contemporary medium-scale architectures on CPU-based hardware. These results show that architectural novelty alone is insufficient for model selection and that training data availability, target hardware, and inference requirements should be considered jointly when selecting object detection models for practical edge AI deployment in aquaculture.
arXiv:2607.10244v1 Announce Type: new
Abstract: State Space Models (SSMs) have emerged as a powerful paradigm for efficient long-sequence modeling, offering parallel training and fast linear-time recurrent inference. However, like other recurrent architectures, SSMs must compress an unbounded history into a fixed-size state, which limits context retention and makes precise retrieval over long-range context inherently difficult. To overcome this limitation, we propose Delay State Space Models (DSSMs), a delay differential equation (DDE)-inspired extension of diagonal SSMs that augments discrete SSM recurrences with explicit delayed-state feedback. Making explicit delayed feedback practical requires new stability parameterization, history management, and FFT-training tools. We address these challenges with a practical discretization and parameterization grounded in a simple delay-independent stability condition. To bypass direct time-domain kernel construction, we derive the DSSM transfer function and compute kernels in the frequency domain, using a kernel contour shift to suppress aliasing and recover accurate FFT training. Empirically, DSSMs substantially improve targeted delayed-retrieval tasks while outperforming S4D on most standard sequence metrics and remaining close on the others.
arXiv:2607.10415v1 Announce Type: new
Abstract: The Aharonov-Bohm (AB) effect remains a cornerstone of fundamental and applied physics. In this work, we utilize the AB caging effect originated from an effective magnetic field induced by multi-orbital interactions, creating an all flat band (FB) lattice system. Normally, FB states are known for being compact in space and having a zero tail; therefore, their mobility in a linear environment is generally understood as impossible. We propose a Floquet driving protocol in an all-FB photonic system to fully control the dynamics of localized photonic states. The modulation of the Hamiltonian along the propagation coordinate allows the translation of compact states in the direction of constructive interference, resulting in an effective stroboscopic quantum walk-like effect. We find that the traveling states exist in chiral pairs and have a related topological invariant (winding number) equal to +1 or -1, with the sign determining the propagation direction. We experimentally implement the Floquet driven protocol using femtosecond laser written photonic waveguides and demonstrate directional control of the propagation, determined by the relative phase of the input condition.
arXiv:2502.14017v4 Announce Type: replace
Abstract: The convergence of Operational Technology (OT) and Information Technology (IT) under Industry~4.0 has widened the cyber-attack surface of critical infrastructure across manufacturing, energy, transportation, water, and healthcare. This survey synthesizes OT/IT cybersecurity along four axes. First, we taxonomize \emph{attack vectors} that traverse the IT--OT boundary, separating IT-side initial access (phishing, exploits, supply-chain compromise, exposed remote access) from OT-side propagation and impact (insecure protocols, weak authentication, firmware tampering, control-logic manipulation). Second, we review \emph{defensive technologies} -- signature-based intrusion detection, AI/ML anomaly detection, Zero Trust Architecture, blockchain-based event logging, digital twins, and OT-aware Security Operations Centers -- and identify remaining \emph{gaps}: OT-specific patch management, dataset scarcity for ML, IoMT segmentation, and the absence of consistent resilience metrics. Third, we compile a cross-validated \emph{historical record} of 69 high-impact incidents spanning 2010--2025, from Stuxnet to Jaguar Land Rover, and quantify their \emph{commercial effects} sector by sector using figures sourced from SEC filings, government post-incident reviews, and primary regulatory disclosures. Fourth, we map the \emph{regulatory landscape}: NIST Cybersecurity Framework2.0 and SP~800-82~Rev.~3, IEC~62443, the EU NIS2 Directive, DORA, the Cyber Resilience Act, NERC~CIP, and healthcare-specific regimes (IEC-80001-1, FDA, NIST-SP-1800-8). A sectoral deep-dive on healthcare illustrates the IT--OT convergence threat model under high-consequence conditions. The result is a single, source-traceable reference on where OT/IT cybersecurity stands, what the historical record costs defenders who lag, and where investment yields the highest marginal return on resilience.
arXiv:2506.14411v2 Announce Type: replace
Abstract: In standard reinforcement learning (RL) settings, the interaction between the agent and the environment is typically modeled as a Markov decision process (MDP), which assumes that the agent observes the system state instantaneously, selects an action without delay, and executes it immediately. In real-world dynamic environments, such as cyber-physical systems, this assumption often breaks down due to delays in the interaction between the agent and the system. These delays can vary stochastically over time and are typically unobservable when deciding on an action. Existing methods deal with this uncertainty conservatively by assuming a known fixed upper bound on the delay, even if the delay is often much lower. In this work, we introduce the interaction layer, a general framework that enables agents to adaptively handle unobservable and time-varying delays. Specifically, the agent generates a matrix of possible future actions, anticipating a horizon of potential delays, to handle both unpredictable delays and lost action packets sent over networks. Building on this framework, we develop a model-based algorithm, Actor-Critic with Delay Adaptation (ACDA), which dynamically adjusts to delay patterns. Our method significantly outperforms state-of-the-art approaches across a wide range of locomotion benchmark environments, including real-world measured delays.
arXiv:2506.16112v4 Announce Type: replace
Abstract: Inspired by text prompts in large language models, visual prompts have been explored to enhance the perceptual capabilities of large vision-language models (LVLMs). However, performance tends to saturate under single visual prompt designs, making further prompt engineering increasingly ineffective. To address this limitation, we shift from prompt engineering to prompt retrieval and propose AutoV, a lightweight framework for instance-adaptive visual prompt identification. Given an input image and a textual query, AutoV automatically locates the most suitable visual prompt from a diverse candidate pool. Training such a retrieval framework requires prompt-level supervision, yet prompt quality is inherently ambiguous and difficult to assess reliably, even for humans. To enable automatic supervision, we evaluate visual prompts using a pre-trained LVLM and label them according to their prediction losses. Using the loss-oriented ranking as a robust training signal, AutoV learns to retrieve the query-aware optimal prompt for each instance without manual annotation. Experiments indicate that AutoV enhances the performance of various LVLMs on image understanding, captioning, grounding, and classification tasks. For example, AutoV improves LLaVA-OV by $\textbf{10.2}\%$ on VizWiz and boosts Qwen2.5-VL by $\textbf{3.8}\%$ on MMMU, respectively.
arXiv:2607.10294v1 Announce Type: new
Abstract: Conventional colonoscopy remains limited by patient discomfort and procedural risks, motivating research into compliant robotic alternatives. Eversion robots, which advance via pressure-driven tip growth, eliminate sliding friction against the colon wall and offer a less intrusive approach to traversal. However, no existing design simultaneously satisfies all clinical requirements. This paper benchmarks four recent eversion robot architectures against the key anatomical and clinical constraints of colonoscopy, including colon length, minimum luminal diameter, bending angle, and working-channel needs. We identify the central trade-offs each design reveals, particularly the difficulty of integrating steering and payload delivery without sacrificing the soft, low-friction behaviour that makes eversion robots attractive. We then provide design guidance across material selection, steering strategy, and payload delivery, and highlight open problems for clinical translation.
arXiv:2607.10429v1 Announce Type: new
Abstract: Traditional computational models, such as cellular automata and phase-field methods, are effective for simulating microstructural evolution but often face computational bottlenecks, limiting their application in high-throughput and on-demand process optimization. Generative machine learning approaches, such as denoising diffusion models, have emerged as powerful tools for surrogate modeling of process-structure maps, specifically producing representative microstructures conditioned on process parameters. However, they often require large amounts of data for training, particularly when process conditions are continuous rather than discrete categorical variables. To address this, we present a continuous conditional denoising diffusion model for generating microstructures conditioned on processing parameters. Trained on a compact dataset of process-microstructure pairs, this framework first adds noise to microstructure images and then trains a neural network to progressively remove the noise, learning the underlying statistical patterns of the microstructure. To address data inefficiencies associated with continuously valued process conditions, we propose a vicinal-loss training strategy that associates process conditions in data-sparse regions with nearby conditions in the dataset. Combined with classifier-free guidance and denoising diffusion implicit sampling, this approach enables data-efficient continuous conditional generation of microstructures compared to classical denoising diffusion models. The model successfully generates representative microstructures for low-carbon steel conditioned on manganese composition, matching key physical features such as phase and grain morphology, grain size distribution, phase fraction, and interfacial area distribution. More generally, this approach opens avenues for efficient process design and optimization of materials and their microstructures.
arXiv:2607.10303v1 Announce Type: new
Abstract: We study nested wire media possessing $C_6$ and $C_4$ rotational symmetries whose plasma frequencies can be controlled through breathing deformation. Numerical simulations reveal tunability exceeding $80\%$ and $60\%$ for the considered $C_6$ and $C_4$ breathing geometries, respectively. To describe these structures, we develop an analytical model based on the local field approach within the thin-wire approximation, which confirms the high tunability of the studied wire structures. We also propose an approximation for dense wire media. The developed framework is applicable to the general case of nested wire structures -- wire media with an arbitrary parallelogram lattice and multiple wires per unit cell.
arXiv:2607.11264v1 Announce Type: new
Abstract: This work proposes a geometry-optimized complex-domain error-diffusion encoding method for Fourier single-pixel imaging. Instead of independently binarizing multiple grayscale phase-shifting patterns, the proposed method directly represents each complex-valued Fourier basis pattern using K (K >= 3) weighted binary patterns while diffusing the residual error in the complex domain. A geometric interpretation is further established, revealing that the encoding process can be viewed as approximating the Fourier-basis unit circle by a regular polygon in the complex plane. Based on this geometric interpretation, practical optimization strategies are developed for K = 3, K = 4, and K = 7. Both numerical simulations and real-object experiments demonstrate consistently superior reconstruction quality compared with conventional phase-shifting dithering.
arXiv:2607.10312v1 Announce Type: new
Abstract: The rapid proliferation of online polarization threatens social cohesion, necessitating robust automated detection systems that operate effectively across diverse linguistic contexts. This paper presents our system description for the POLAR Shared Task 2026, focusing on the detection and characterization of polarized discourse in English and Hausa. We propose a hybrid modeling strategy: for English binary detection, we leverage the monolingual strength of \textbf{DeBERTa}, while for Hausa and all fine-grained subtasks (Types and Manifestations), we utilize \textbf{AfroXLMR-Social}. This domain-adapted multilingual model proved critical for capturing the nuances of polarization in social media text. To further address computational constraints and data scarcity, we implement Low-Rank Adaptation (LoRA) and textual data augmentation via \texttt{nlpaug}. We report competitive results across all three subtasks, demonstrating that model selection tailored to specific subtask requirements yields the best balance of performance.
arXiv:2607.10316v1 Announce Type: new
Abstract: Retrieval-Augmented Generation (RAG) systems enhance large language models by retrieving relevant documents from external knowledge bases. Recent work by Sarthi et al. (2024) introduced RAPTOR, which organizes documents into hierarchical tree structures for efficient retrieval, but requires expensive LLM-based abstractive summarization at each internal node -- making large-scale deployment prohibitively costly.
We present SVD-RAG, the first method to apply Singular Value Decomposition (SVD) on dense sentence embedding matrices for extractive summarization in hierarchical RAG. Unlike classical LSA which operates on sparse TF-IDF matrices, SVD-RAG exploits the rich semantic representations of modern embedding models, identifying the most informative sentences through their energy contribution in the principal components. Our approach is (1) deterministic -- unlike LLM-based summarization, SVD produces identical results for the same input; (2) cost-efficient -- tree construction requires no additional API calls beyond the initial embedding, reducing token consumption by ~85%; and (3) content-adaptive -- the energy-ratio threshold tau automatically adjusts compression based on content complexity. In a controlled head-to-head comparison using identical corpora, clustering, and beam search, SVD-RAG achieves retrieval quality within 1-5% of RAPTOR with LLM summarization (MRR 0.867 vs. 0.875, Recall@1 0.483 vs. 0.458) while building the tree 317x faster (0.1s vs. 31.7s). On a scaled multi-topic benchmark with 205 chunks and 100 queries across 20 topic variations, SVD-RAG achieves a 4.2x improvement in Recall@1 and 3.1x improvement in MRR over flat embedding retrieval. We provide a detailed cost analysis and parameter sensitivity study. Our implementation is released as an open-source Python package.
arXiv:2607.11184v1 Announce Type: new
Abstract: SLAM methods based on 3D Gaussian Splatting (3DGS) have demonstrated impressive tracking and mapping performance, but typically require additional geometric information from external depth sensors. Meanwhile, recent SLAM systems that leverage geometric priors from pre-trained feed-forward models enable real-time dense reconstruction, yet often discard original RGB information during optimization, thus degrading overall reconstruction quality. We present GeoGS-SLAM, an online monocular dense reconstruction system that combines the 3DGS-based map representation with learned geometric priors. Given uncalibrated RGB input, we first employ a feed-forward visual geometry model to predict camera and scene priors. The Gaussian scene map is then expanded by directly sampling Gaussian primitives from both RGB input and geometric priors. Camera poses and the scene map are jointly optimized through a coarse-to-fine strategy that minimizes both photometric and geometric losses. To ensure global consistency, we further incorporate online loop closure detection and pose graph optimization. Extensive experiments across indoor and outdoor benchmarks demonstrate that GeoGS-SLAM achieves superior rendering quality and tracking accuracy compared to state-of-the-art methods while maintaining online real-time performance. Project page: https://rlgao.github.io/geogs_slam.
arXiv:2607.10458v1 Announce Type: new
Abstract: ALFE2 is a front-end ASIC developed for the ATLAS Liquid Argon (LAr) Calorimeter upgrade during the High-Luminosity Large Hadron Collider (HL-LHC) phase. ALFE2 comprises four preamplifier/shaper channels, each providing two distinct gain outputs to cover a 16-bit dynamic range. A robotic system has been developed for the automatic quality control test of ALFE2, and over 10% of the 80,000 chips have been evaluated by September 2025. The evaluation has allowed us to establish grading criteria. Using these criteria, a yield of over 85% was achieved in the evaluation tests, and these criteria are now being applied to the ongoing full-production QC. Irradiation tests were also performed for the quality assurance of ALFE2. No significant performance degradation was observed during the total-ionizing-dose (TID) test. Based on the single-event effect (SEE) test results, an error rate of fewer than 4.6 single-event upsets (SEUs) per day is extrapolated for the entire ATLAS LAr Calorimeter during HL-LHC operation.
arXiv:2607.10588v1 Announce Type: new
Abstract: Tasks such as customs tariff classification, export control categorization, and standards-based equipment coding require assigning an input instance to a fine-grained class under an explicit regulatory hierarchy. Unlike standard text classification, the correct label in these tasks is not determined by semantic similarity alone, but by rule-defined boundaries, threshold conditions, exclusion clauses, definitions, and local exceptions. As a result, two highly similar inputs may require different labels, while a retrieved passage that appears relevant may still be inapplicable under the governing rules. Existing flat classifiers, hierarchical text classification methods, and retrieval-augmented LLM systems are not designed to jointly enforce hierarchical validity, rule consistency, and fine-grained boundary reasoning. In this paper, we formulate this setting as regulation-driven fine-grained hierarchical classification, where an external instance must be assigned to a fine-grained class through a valid path in a regulatory hierarchy and supported by auditable evidence. We construct four benchmark datasets from representative regulation-intensive scenarios and validate the annotations through an expert-in-the-loop process. We further propose a constraint-aware hierarchical search framework that converts regulatory documents into a searchable tree, retrieves only valid local candidate nodes, and uses structured regulatory fields with evidence snippets to guide each next-hop decision. Experiments show that our method achieves the best mean accuracy on all four datasets and provides interpretable decision paths, with the largest gains on cases involving fine-grained neighboring categories and rule-based boundary conditions.
arXiv:2607.11196v1 Announce Type: new
Abstract: Deploying object-centric models for real-world scene understanding typically requires complex pipelines to achieve both robust scene decomposition and high-fidelity generation. Recent diffusion-based approaches have improved visual quality, but they almost universally rely on heavy, pretrained generative priors (e.g., Stable Diffusion) and external VAE latent spaces. In this paper, we propose Slot-RAE, a much simpler, fully integrated framework that operates directly within the continuous semantic feature space of visual foundation models (e.g., DINOv3). Slot-RAE employs a feature-space diffusion process using a Diffusion Transformer (DiT) decoder and a Representation Alignment (REPA) head. Unlike existing diffusion-based objectcentric methods that rely heavily on subsidized text-toimage priors, the generative core of Slot-RAE (Slot Attention and the DiT) is trained from scratch within the frozen VFM feature space. This eliminates the need for VAE bottlenecks and task-agnostic generative pre-training. Experiments on the COCO dataset demonstrate that despite its architectural simplicity, Slot-RAE achieves state-of-the-art results. It delivers comparable unsupervised object discovery, higher-fidelity image reconstruction, and robust zero-shot compositionality, all while being significantly faster and more computationally efficient than existing object-centric latent diffusion models.
arXiv:2506.19806v3 Announce Type: replace
Abstract: This position paper argues that LLM-based social simulations require clear boundaries to make meaningful contributions to social science. While Large Language Models (LLMs) offer promising capabilities for simulating human behavior, their tendency to produce homogeneous outputs, acting as an "average persona", fundamentally limits their ability to capture the behavioral diversity essential for complex social dynamics. We examine why heterogeneity matters for social simulations and how current LLMs fall short, analyzing the relationship between mean alignment and variance in LLM-generated behaviors. Through a systematic review of representative studies, we find that validation practices often fail to match the heterogeneity requirements of research questions: while most papers include ground truth comparisons, fewer than half explicitly assess behavioral variance, and most that do report lower variance than human populations. We propose that researchers should: (1) match validation depth to the heterogeneity demands of their research questions, (2) explicitly report variance alongside mean alignment, and (3) constrain claims to collective-level qualitative patterns when variance is insufficient. Rather than dismissing LLM-based simulation, we advocate for a boundary-aware approach that ensures these methods contribute genuine insights to social science.
arXiv:2607.10394v1 Announce Type: new
Abstract: The evolution from 5G towards 6G reinforces interest in connected robotics, where mobile robots offload compute-intensive tasks to edge servers over ultra-reliable low-latency communication (URLLC) links. Simultaneous localization and mapping (SLAM), a fundamental yet demanding robotics function, is increasingly considered for edge deployment within mobile edge computing (MEC) frameworks. In parallel, integrated sensing and communications (ISAC) enables the use of radio channel information, such as channel state information (CSI), as an additional sensing modality in radio-based SLAM. In this paper, we design and implement a CSI-assisted Edge SLAM testbed integrating a custom unmanned ground vehicle (UGV), a ROS2-based SLAM framework, and a 5G Open Radio Access Network (O-RAN) system. The proposed architecture provides an end-to-end, cross-layer view of ROS2 sensor data streaming over 5G, explicitly enabling CSI exposure and integration into the SLAM pipeline. We analyze ROS2 DDS communication, RTPS packetization, and 5G user-plane transport, and discuss mechanisms for CSI extraction and delivery via O-RAN components. The platform enables realistic experimentation with communication-aware SLAM and reveals key challenges related to latency, data streaming, synchronization, and cross-system integration, providing insights for future 6G-enabled robotic platforms.
arXiv:2607.09805v1 Announce Type: cross
Abstract: Percutaneous Coronary Intervention (PCI) is a minimally invasive procedure used to restore coronary blood flow obstructed by atherosclerotic plaque. During PCI, repeated injections of iodine-based contrast agents are required to visualize the coronary arteries and guide interventional devices. However, frequent contrast injections increase radiation exposure and the risk of contrast-induced nephropathy, with acute kidney injury reported in up to 30% of patients with renal impairment. Dynamic Coronary Roadmapping (DRM) reduces these risks by overlaying a precomputed angiographic vessel map onto live fluoroscopy and continuously updating it throughout the procedure. Accurate DRM relies on precise cardiac phase matching between angiography and fluoroscopy, together with reliable catheter tip tracking for motion compensation. These tasks remain challenging in ECG-free settings and when only limited manual annotations are available. We present a unified DRM framework that simultaneously performs cardiac phase matching and catheter tip tracking for accurate real-time guidance. Our method employs a large-scale spatio-temporal encoder pretrained on 16 million X-ray frames to learn cardiac motion dynamics. To the best of our knowledge, this is the first application of large-scale spatio-temporal pretraining for motion compensation in DRM. We further introduce auxiliary tasks based on ECG R-peak detection and catheter tip tracking, improving optimization while eliminating the need for extensive catheter mask annotations. Finally, a majority-voting postprocessing strategy aggregates temporal predictions, improving robustness and providing a confidence score that correlates with phase-matching error. Comprehensive evaluation on clinical X-ray datasets demonstrates state-of-the-art performance, achieving low temporal misalignment and robust phase-matching accuracy suitable for real-time DRM.