arXiv:2603.24167v2 Announce Type: replace Abstract: WebAssembly's (Wasm) monolithic linear memory turns a single memory-corruption bug into a bidirectional threat: a compromised module can attack its embedding host, and a malicious host can tamper with a trusted module's state. Existing defenses require custom runtimes or source changes, and none can verify runtime integrity under an adversarial host. We present Walma, a framework for WebAssembly linear memory attestation that renders snapshots of linear memory as grayscale images and classifies them with a convolutional neural network. The network reads the row-aligned structure that compiled code imposes on memory and detects corruption that byte and texture statistics miss. On structured targets this extends to tampering that no program input triggers, such as direct memory writes by a malicious host. On real-world CVE-affected applications, hiding corruption from Walma's verdict requires overwriting a large fraction of memory, hundreds of kilobytes to megabytes on our targets. A Shannon-entropy analysis of benign memory bounds the class of out-of-band edits that the learned prior can detect. Across the 53 binaries of the WABench suite, attestation costs 1.07x--1.69x at the host boundary, making continuous, state-centric memory attestation for Wasm practical.
Science Journals
arXiv:2605.05207v2 Announce Type: replace Abstract: Dense 3D reconstruction and tracking of dynamic scenes from monocular video remains an important open challenge in computer vision. Progress in this area has been constrained by the scarcity of high-quality datasets with dense, complete, and accurate geometric annotations. To address this limitation, we introduce Syn4D, a multiview synthetic dataset of dynamic scenes that includes ground-truth camera motion, depth maps, dense tracking, and parametric human pose annotations. A key feature of Syn4D is the ability to unproject any pixel into 3D to any time and to any camera. We conduct extensive evaluations across multiple downstream tasks to demonstrate the utility and effectiveness of the proposed dataset, including 4D scene reconstruction, 3D point tracking, geometry-aware camera retargeting, and human pose estimation. The experimental results highlight Syn4D's potential to facilitate research in dynamic scene understanding and spatiotemporal modeling.
arXiv:2106.15277v5 Announce Type: replace Abstract: We study multi-sensor fusion for 3D semantic segmentation that is important to scene understanding for many applications, such as autonomous driving and robotics. Existing fusion-based methods, however, may not achieve promising performance due to the vast difference between the two modalities. In this work, we investigate a collaborative fusion scheme called perception-aware multi-sensor fusion (PMF) to effectively exploit perceptual information from two modalities, namely, appearance information from RGB images and spatio-depth information from point clouds. To this end, we project point clouds to the camera coordinate using perspective projection, and process both inputs from LiDAR and cameras in 2D space while preventing the information loss of RGB images. Then, we propose a two-stream network to extract features from the two modalities, separately. The extracted features are fused by effective residual-based fusion modules. Moreover, we introduce additional perception-aware losses to measure the perceptual difference between the two modalities. Last, we propose an improved version of PMF, i.e., EPMF, which is more efficient and effective by optimizing data pre-processing and network architecture under perspective projection. Specifically, we propose cross-modal alignment and cropping to obtain tight inputs and reduce unnecessary computational costs. We then explore more efficient contextual modules under perspective projection and fuse the LiDAR features into the camera stream to boost the performance of the two-stream network. Extensive experiments on benchmark data sets show the superiority of our method. For example, on nuScenes test set, our EPMF outperforms the state-of-the-art method, i.e., RangeFormer, by 0.9% in mIoU. Our source code is available at https://github.com/ICEORY/PMF.
arXiv:2110.00067v3 Announce Type: replace Abstract: The total variation diminishing (TVD) property is an important tool for ensuring nonlinear stability and convergence of numerical solutions of one-dimensional scalar conservation laws. However, it proved to be challenging to extend this approach to two-dimensional problems. Using the anisotropic definition for discrete total variation (TV), it was shown in \cite{Goodman} that TVD solutions of two-dimensional hyperbolic equations are at most first order accurate. We propose to use an alternative definition resulting from a full discretization of the semi-discrete Raviart-Thomas TV. We demonstrate numerically using the second order discontinuous Galerkin method that limited solutions of two-dimensional hyperbolic equations are TVD in means when total variation is computed using the new definition.
arXiv:2607.04529v1 Announce Type: new Abstract: In this paper, we examine unsupervised, content-based collaboration recommendations using publication text in scholarly settings. We compare three families of methods: a TF-IDF baseline, topic-based models (LDA and BERTopic, including clone variants), and embedding-based retrieval using SciBERT with Faiss. To evaluate model behavior beyond simple lexical matching, we introduce a constrained setting where publication overlap between researchers is partially removed while still using historical co-authorship as proxy ground truth for post-hoc evaluation. Results show clear differences across methods. TF-IDF performs best under full information but drops significantly as overlap is reduced. In contrast, topic-based and embedding-based approaches show more stable performance, suggesting they capture broader distributional similarities, rather than relying only on direct lexical overlap. We also examine explainability through two perspectives: intrinsic topic-based explanations and post-hoc, retrieval-based explanations generated using language models. These provide complementary trade-offs between transparency and human readability.
arXiv:2507.11878v5 Announce Type: replace Abstract: LLMs are trained to refuse harmful instructions, but do they truly understand harmfulness beyond just refusing? Prior work has shown that LLMs' refusal behaviors can be mediated by a one-dimensional subspace, i.e., a refusal direction. In this work, we identify a new dimension to analyze safety mechanisms in LLMs, i.e., harmfulness, which is encoded internally as a separate concept from refusal. There exists a harmfulness direction that is distinct from the refusal direction. As causal evidence, steering along the harmfulness direction can lead LLMs to interpret harmless instructions as harmful, but steering along the refusal direction tends to elicit refusal responses directly without reversing the model's judgment on harmfulness. Furthermore, using our identified harmfulness concept, we find that certain jailbreak methods work by reducing the refusal signals without reversing the model's internal belief of harmfulness. We also find that adversarially finetuning models to accept harmful instructions has minimal impact on the model's internal belief of harmfulness. These insights lead to a practical safety application: The model's latent harmfulness representation can serve as an intrinsic safeguard (Latent Guard) for detecting unsafe inputs and reducing over-refusals that is robust to finetuning attacks. For instance, our Latent Guard achieves performance comparable to or better than Llama Guard 3 8B, a dedicated finetuned safeguard model, across different jailbreak methods. Our findings suggest that LLMs' internal understanding of harmfulness is more robust than their refusal decision to diverse input instructions, offering a new perspective to study AI safety.
arXiv:2607.04112v1 Announce Type: new Abstract: Multimodal LLMs struggle to systematically model the temporal evolution of visual scenes in videos or multi-image sequences. Such inputs require models to predict or simulate multiple levels of dynamic constituents, such as actions taken in the visual sequence, and the associated changes to the visual environment that result. To address this challenge, we propose a dynamic schema-guided world model, DynaVieW, optimized for visual dynamic prediction and simulation. DynaVieW achieves an in-depth understanding of visual dynamics by learning interleaved state-transition sequences, where states cover broad visual scenes from video keyframes, and transitions capture comprehensive dynamic constituents within a hierarchical schema. DynaVieW jointly models transition prediction and state simulation under a mixture-of-experts architecture, with a cross-expert selective attention and a schema token re-weighted loss, to ensure effective and robust learning. DynaVieW's understanding of visual dynamics boosts its downstream performance in visual narrative creation and world simulation, showing improved consistency, controllability, and instruction-following.
arXiv:2603.03955v3 Announce Type: replace Abstract: Post-training with reinforcement learning (RL) has recently shown strong promise for advancing multimodal agents beyond supervised imitation. However, RL remains limited by poor data efficiency, particularly in settings where interaction data are scarce and quickly become outdated. To address this challenge, GIPO (Gaussian Importance sampling Policy Optimization) is proposed as a policy optimization objective based on truncated importance sampling, replacing hard clipping with a log-ratio-based Gaussian trust weight to softly damp extreme importance ratios while maintaining non-zero gradients. Theoretical analysis shows that GIPO introduces an implicit, tunable constraint on the update magnitude, while concentration bounds guarantee robustness and stability under finite-sample estimation. Experimental results show that GIPO achieves state-of-the-art performance among clipping-based baselines across a wide range of replay buffer sizes, from near on-policy to highly stale data, while exhibiting superior bias--variance trade-off, high training stability and improved sample efficiency. Code is available at https://github.com/distanceLu/GIPO.
arXiv:2202.14019v3 Announce Type: replace Abstract: Maintaining proper form while exercising is important for preventing injuries and maximizing muscle mass gains. Detecting errors in workout form naturally requires estimating human's body pose. However, off-the-shelf pose estimators struggle to perform well on the videos recorded in gym scenarios due to factors such as camera angles, occlusion from gym equipment, illumination, and clothing. To aggravate the problem, the errors to be detected in the workouts are very subtle. To that end, we propose to learn exercise-oriented image and video representations from unlabeled samples such that a small dataset annotated by experts suffices for supervised error detection. In particular, our domain knowledge-informed self-supervised approaches (pose contrastive learning and motion disentangling) exploit the harmonic motion of the exercise actions, and capitalize on the large variances in camera angles, clothes, and illumination to learn powerful representations. To facilitate our self-supervised pretraining, and supervised finetuning, we curated a new exercise dataset, Fitness-AQA (https://github.com/ParitoshParmar/Fitness-AQA), comprising of three exercises: BackSquat, BarbellRow, and OverheadPress. It has been annotated by expert trainers for multiple crucial and typically occurring exercise errors. Experimental results show that our self-supervised representations outperform off-the-shelf 2D- and 3D-pose estimators and several other baselines. We also show that our approaches can be applied to other domains/tasks such as pose estimation and dive quality assessment.
arXiv:2507.12721v3 Announce Type: replace Abstract: Human-AI interfaces play a pivotal role in integrating clinicians' expertise with artificial intelligence to enhance both healthcare practice and research. However, designing effective interfaces in this domain remains a significant challenge. The inherent complexity of medical data, the influence of domain-specific conventions, and the diverse needs of clinical users compound the challenge of developing practical and usable solutions. In this study, we review existing solutions and synthesize a set of design patterns - recurring approaches that support the design of human-AI interfaces in clinical settings. We conducted a comprehensive literature review of human-AI interaction designs in clinical contexts, through which we identified 15 information entities commonly presented to users and 12 design patterns used to organize and communicate this information effectively. For each design pattern, we summarize the underlying design problem, the proposed solution, and the rationale for when the pattern should or should not be applied, based on insights from both the literature and semi-structured interviews with 12 healthcare professionals. We evaluated the proposed design patterns through an online workshop involving 14 experienced UI designers. During the workshop, participants were asked to create interface sketches for healthcare-related scenarios drawn from their own professional experience, using our design patterns as guidance. Our findings show that the proposed design patterns helped participants ground their designs in user needs, generate a wider range of design alternatives, and simplify complex interface structures. We further analyzed and summarized the participants' usage strategies and feedback regarding the applicability and usefulness of the design patterns.
arXiv:2607.05339v1 Announce Type: new Abstract: Group Relative Policy Optimization (GRPO) is effective when the current policy already samples useful reasoning trajectories, but it stalls on hard prompts whose correct solution modes lie outside the student's on-policy support. We propose TREK (Teacher-Routed Exploration via Forward KL), a simple staged procedure that uses distillation not for imitation but for exploration support expansion. A key advantage of TREK is its generality: because it only consumes verified output trajectories, it can use an external black-box teacher, a white-box teacher, or the same model given additional inference-time context, and it can efficiently identify which hard-prompt samples are most worth consolidating even when teacher internals are unavailable. TREK first identifies prompts where the unaided student has very low pass rate, queries a proposal source to produce verified candidate solutions, keeps the top-$r$ proposals ranked by current student likelihood, applies a short forward-KL phase to pull those verified modes into the student's support, and then returns to standard on-policy GRPO refinement. On mathematical reasoning, TREK with DeepSeek-V4 proposals improves Qwen3 models across all tested scales on AIME 2024 and AIME 2025; for Qwen3-8B, it improves AIME 2025 from 36.9 to 40.3 and AIME 2024 from 47.9 to 51.1 (avg@16), while the self-context variant reaches 38.5 and 49.6 without an external teacher. On agentic tasks, TREK raises ALFWorld success rate from 75.8 to 82.8 and ScienceWorld success rate from 12.5 to 26.7; notably, on the hardest task types, TREK achieves high success rates early in training while unaided GRPO requires substantially more optimization steps to reach comparable levels.
arXiv:2607.04549v1 Announce Type: new Abstract: Many of the timescales of life have speed limits set by quantum-mechanical constraints along with non-fundamental quantities, such as the temperature of the environment, which are however bounded by anthropic considerations. Here, some of such speed limits are examined, including those for the rates of elementary chemical reactions and biomolecular folding. Limitations of simple back-of-envelope estimates are also discussed.
arXiv:2606.22908v3 Announce Type: replace Abstract: Purely functional programming makes immutability and non-strict evaluation its defaults; we push further, making interning and memoization the defaults of computation. In prior operational semantics for the pure $\lambda$-calculus, representing and transforming cyclic and infinite data requires an added recursion construct, a $\texttt{letrec}$, a $\mu$-binder, or a built-in $Y$ for graph reduction, and sharing the repeated work of a memoized or dynamic-programming function requires an impure cache. We show that no extension is needed. We apply tabling, the standard method for solving a least-fixpoint equation, to weak-head reduction; this defines a new operational semantics for the pure $\lambda$-calculus that keeps each term's standard lazy meaning. A term that reaches finitely many distinct states, each solved in finitely many steps, comes out as a finite graph, possibly cyclic; the calculus stays pure, and the graph is sound and independent of reduction order. We implemented this operational semantics as a $\lambda$-calculus interpreter. It does dynamic programming automatically, sharing repeated subproblems with no memoization table. It creates and transforms cyclic graphs with no added recursion construct. And it decides an unproductive loop, returning $\bot$ for $\Omega$ in finite time. What the evaluator returns is a graph, so the $\lambda$-calculus becomes a DSL for graph computation: the memo table of dynamic programming, the transposition table of game search, and the derived-fact table of Datalog reachability and points-to analysis are all tabling on state identity, and none of them is written by hand. Compilation is one more such problem: we write a bootstrap compiler that compiles its own source, all as a pure $\lambda$-term.
arXiv:2510.22298v2 Announce Type: replace-cross Abstract: Uncovering the causal mechanisms of complex real-world systems remains a significant challenge, as these systems often entail high data collection costs and involve unknown interventions. We introduce MetaCaDI, the first framework to cast the identification of unknown interventions as a meta-learning problem, explicitly leveraging a jointly learned causal graph. MetaCaDI is a Bayesian framework that learns a shared causal structure across multiple environments and is optimized to rapidly adapt to new, few-shot intervention target identification tasks. A key innovation is our model's analytical adaptation, which uses a closed-form solution to bypass expensive and potentially unstable gradient-based bilevel optimization. Extensive experiments on synthetic and complex gene expression data demonstrate that MetaCaDI significantly outperforms state-of-the-art methods. It excels at identifying intervention targets from as few as 3 samples - where existing methods collapse to random chance - while robustly recovering the shared causal graph, proving its effectiveness in data-scarce scenarios.
arXiv:2607.04614v1 Announce Type: new Abstract: Modern software development relies heavily on open-source components. Reusing components accelerates innovation but increases exposure to supply-chain attacks exploiting known vulnerabilities. Software Bills of Materials (SBOMs) improve software supply chain transparency by enumerating components, their versions, and their provenance. GitHub, the largest open-source development hosting platform, now automatically generates SBOMs for repositories, providing valuable metadata for risk assessment. Yet, it is unclear whether GitHub SBOMs can serve as a reliable source for vulnerability and license analysis, and how incomplete or inconsistent metadata may affect different programming ecosystems. To address this, we conduct a large-scale analysis of 10,000 GitHub repositories across ten programming language ecosystems, evaluating GitHub SBOMs against three other popular SBOM generators: Syft, Trivy, and the Microsoft SBOM Tool. Our study finds a lack of NTIA compliance in GitHub SBOMs, though core metadata is consistently present. We also find that component version and license information availability is highly dependent on the programming ecosystem. Compared with the other three tools, GitHub yields results similar to the Microsoft SBOM Tool and often outperforms Syft and Trivy in providing version and license information. Finally, we discuss potential shortcomings of the GitHub SBOM Tool, directly related to how each ecosystem manages its dependencies.
arXiv:2508.05049v1 Announce Type: cross Abstract: AI-powered medical devices have driven the need for real-time, on-device inference such as biomedical image classification. Deployment of deep learning models at the edge is now used for applications such as anomaly detection and classification in medical images. However, achieving this level of performance on edge devices remains challenging due to limitations in model size and computational capacity. To address this, we present MedMambaLite, a hardware-aware Mamba-based model optimized through knowledge distillation for medical image classification. We start with a powerful MedMamba model, integrating a Mamba structure for efficient feature extraction in medical imaging. We make the model lighter and faster in training and inference by modifying and reducing the redundancies in the architecture. We then distill its knowledge into a smaller student model by reducing the embedding dimensions. The optimized model achieves 94.5% overall accuracy on 10 MedMNIST datasets. It also reduces parameters 22.8x compared to MedMamba. Deployment on an NVIDIA Jetson Orin Nano achieves 35.6 GOPS/J energy per inference. This outperforms MedMamba by 63% improvement in energy per inference.
arXiv:2607.04619v1 Announce Type: new Abstract: Modern automated audio captioning systems pair a frozen audio encoder with a large language model (LLM) via a trainable projector, incurring the encoder's inference cost and bottlenecking the model through its fixed acoustic features. We present CARD, an encoder-free audio captioning model that removes the encoder at inference: a 13.2M projector feeds a frozen LLM with merged LoRA adapters, while the teacher used to train it is discarded. CARD distills a pretrained audio teacher (CLAP-HTSAT) into the model, but rather than injecting it into the LLM alone, it routes the teacher's representations across components: perceptual stages to the projector and semantic stages to the LLM. This placement improves CIDEr-D by +12.18 over an LLM-only distilled model on AudioCaps and by +5.21 on Clotho, reaching 55.4 against a 66.4 encoder-kept upper bound with no encoder at inference, showing that where a teacher's knowledge is placed matters as much as its presence.
arXiv:2607.04285v1 Announce Type: new Abstract: We propose two numerical domain decomposition methods (DDMs) for elliptic equations on compact Riemannian manifolds, based on physics-informed neural networks (PINNs). Our approach incorporates the DDM technique for manifolds with the advantages of neural networks in high-dimensional settings. The proposed methods are validated through numerical experiments on various manifolds, both with and without boundary, in dimensions ranging from $5$ to $10$.
arXiv:2510.19698v3 Announce Type: replace Abstract: Large Language Models (LLMs) can propose rules in natural language, sidestepping the need for a predefined predicate space in traditional rule learning. Yet many LLM-based approaches ignore interactions among rules, and the opportunity to couple LLMs with probabilistic rule learning for robust inference remains underexplored. We present RLIE, a unified framework that integrates LLMs with probabilistic modeling to learn a set of weighted rules. RLIE has four stages: (1) Rule generation, where an LLM proposes and filters candidates; (2) Logistic regression, which learns probabilistic weights for global selection and calibration; (3) Iterative refinement, which updates the rule set using prediction errors; and (4) Evaluation, which compares the weighted rule set as a direct classifier with methods that inject rules into an LLM. We evaluate multiple inference strategies on real-world datasets. Applying rules directly with their learned weights yields superior performance, whereas prompting LLMs with the rules, weights, and logistic-model outputs surprisingly degrades accuracy. This supports the view that LLMs excel at semantic generation and interpretation but are less reliable for precise probabilistic integration. RLIE clarifies the potential and limitations of LLMs for inductive reasoning and couples them with classic probabilistic rule combination methods to enable more reliable neuro-symbolic reasoning.
arXiv:2607.04362v1 Announce Type: new Abstract: Volterra series feedback linearizes a class of nonlinear hyperbolic PDEs but produces a controller that, even after truncation, demands solving a tower of plant-specific kernel PDEs and evaluating nested integrals. We prove the truncated controller is jointly Lipschitz in plant and state, and learn it as a single neural operator from plant nonlinearity and state to boundary control. Once trained, no kernel is ever solved again, for any plant in the trained class. The closed loop is practically stable in class-$\mathcal{KL}$ form, with a residual ball scaling linearly with training accuracy.
arXiv:2607.04367v1 Announce Type: new Abstract: In the development of cooking robots, mastering the task of cutting is crucial. A significant challenge lies in the diverse properties of food, which necessitate distinct cutting policies and even different knives for optimal processing. This paper presents a perception-manipulation framework for food-cutting tasks. Our system features a knife selection module that utilizes force data from a preliminary fixed trial cut to select the appropriate knife for the given food. This is followed by an adaptive cutting phase using reinforcement learning (RL) to balance cutting speed and energy efficiency. In our experiments, the knife selection module achieved 100% successful rate on unseen food, and we compared the performances of fixed policy, RL policy, with human operators. Our method not only achieves high performance but also demonstrates comparable results to those of human participants.
arXiv:2607.04129v1 Announce Type: new Abstract: Autonomous driving (AD) systems continuously transform multi-rate and asynchronous sensor streams into vehicle actuation through graphs of callbacks, nodes, and middleware components. In such systems, temporal correctness cannot be characterized by the execution time or deadline of an individual task alone: localization and perception chains run in parallel, fuse data with different timestamps, converge at planning, and propagate through control to actuation. Moreover, the demand for high processing capability places AD systems on high-performance processors with multicore parallelism and GPU acceleration, where execution times vary strongly with the input scene, hardware state, and co-running work. Rare deadline misses at runtime therefore cannot be ruled out, and safety is preserved through fail-safe mechanisms such as the minimal-risk maneuver (MRM). This raises a two-sided question: what analytical models are needed to reason about timing in AD systems, and what system software is needed to realize, observe, and enforce those models on real platforms? On the analytical side, real-time research has evolved from periodic/sporadic tasks, directed acyclic graphs (DAGs), pipelines, mixed-criticality systems, and timer-/event-driven models toward end-to-end latency along cause-effect chains, data freshness, timing disparity, probabilistic timing, highest-criticality fail-safe operation, and early deadline-miss detection. On the system-software side, AD stacks and middleware, such as Autoware and ROS~2, expose both the opportunities and limitations of implementing analyzable timing behavior through executors, communication layers, tracing tools, and evaluation frameworks. This paper surveys these two lines of work and identifies the remaining gaps along five dimensions: units of timing constraints, timing metrics, resource models, execution-time variability, and safety integration...
arXiv:2607.04372v1 Announce Type: new Abstract: Event detection tasks in videos, the most important aspect of video surveillance, aim to detect events either at the pixel-level, frame-level, or clip-level. Plenty of methods intended for event detection in different environments, for various applications, and within different acquisition techniques were introduced. Naturally, the attempts were made as well to classify these algorithms in terms of detection of performance or in terms of real-time abilities. Nevertheless, the lack of a large-scale dataset as well as rigorous performance evaluation methods have biased such comparisons as well as the development of the methods. Given the diversity of existing approaches, we believe it is essential for researchers to position their work within such a rich landscape. Thus, we propose a rigorous framework for developing new methods in event detection for videos. Specifically, this framework is based on three main pillars: datasets, performance evaluation, and scenarios for deploying methods.
arXiv:2607.04751v1 Announce Type: new Abstract: Big goals are hard to achieve all at once; breaking them into small steps is wiser. We present Trust Region Policy Distillation (TOP-D), which transforms the notoriously unstable, high-variance On-Policy Distillation (OPD) into a stable training paradigm by dynamically constructing a proximal teacher. Theoretically, we establish a rigorous framework demonstrating that TOP-D inherently controls gradient variance. By providing a formal global convergence analysis alongside a monotonic improvement bound, we mathematically formalize the reliability and stability of the overall training dynamics. Empirically, TOP-D dramatically enhances training stability, sample efficiency, and final performance on mathematical reasoning tasks. More importantly, TOP-D introduces zero additional computational overhead, positioning itself as a promising alternative to the well-established OPD paradigm.
arXiv:2607.04331v1 Announce Type: new Abstract: Evaluating autonomous driving systems in closed-loop settings requires realistic and interactive simulation, yet existing simulators largely rely on log replay or rule-based agents, limiting behavioral diversity and long-tail coverage. We propose an agent-driven simulation framework in which surrounding road participants are controlled by instruction-following large language models through a structured action interface, enabling intentional and reactive behaviors while preserving physical plausibility. Furthermore, we introduce SemanticPlan, a benchmark of closed-loop planning in long-tail and semantically rich scenarios that augment real nuPlan scenes with multiple interactive agents following diverse language instructions. Evaluation results show that state-of-the-art planners still struggle to consistently achieve safe and effective task completion, suggesting that these long-tail scenarios remain challenging.