arXiv:2607.18171v1 Announce Type: new
Abstract: Real-time multimodal applications, including voice agents and interactive video generation, compose heterogeneous models into pipelines whose efficient deployment requires application-specific decisions about placement, streaming, and intra-model parallelism. Existing serving systems and auto-parallelism compilers commit to limited transformations and fixed workload assumptions, so achieving high performance on a new application requires hand-crafting an efficient implementation. We present FlashRT, an agent harness that guides coding agents to lift simple developer-written reference implementations into optimized multi-GPU deployments that flexibly weigh target metrics like latency and throughput. Using a new chain-of-program paradigm, FlashRT directs a generic coding agent through a multi-pass transformation process where an agent transforms the reference into an intermediate representation (IR) to capture data dependencies and persistent-state scopes, validates this IR via a sequential interpreter, and performs static analyses to identify candidate transformations. Then, the agent iteratively implements, verifies, and benchmarks each candidate under a measurement-gated optimization loop to produce effective deployments that span different hardware budgets. Across various applications, including video world models and multimodal LLMs, FlashRT converts reference implementations into highly efficient deployments, delivering up to ~70x latency reduction and 2.8x throughput improvement on NVIDIA B200 GPUs. On AMD MI355X GPUs, FlashRT matches the peak latency reduction while increasing peak throughput improvement to 3.6x, demonstrating that agent-driven optimization can be more scalable on platforms with less mature expert optimization. In fact, for Qwen3-Omni text-to-audio inference, FlashRT reduces response latency by 65% compared to the expert vLLM-Omni implementation on AMD MI355X.
Science Journals
arXiv:2607.16508v1 Announce Type: new
Abstract: Fish-like swimming has inspired the design of several dozens if not hundreds of bioinspired robots in the last few decades. But the control and motion planning of such robots has been challenging due to the poorly modeled fluid-structure interaction and the nonlinear underactuated dynamics of such robots. While reinforcement learning has allowed significant advances in the context of ground and aerial robots, the lack of a suitable simulation environment with appropriate computational speed and accuracy have prevented similar progress for fish-like robots. We address this two-fold problem by developing a simulation platform that approximates the motion of our fish-like robot with computational efficiency. Then the motion control and path tracking by the robot is performed using PID control where the (variable) gains are learned using back propagation through time and training on a curriculum. The policy learned in the simulation is then applied on the physical platform, demonstrating an excellent match.
arXiv:2607.16897v1 Announce Type: new
Abstract: With the growing demand for robotics, autonomous drones, and wearable extended reality systems, the deployment of Visual SLAM on embedded devices remains challenging. Tracking must sustain high frame rates while preserving compute resources for map extension and maintenance. This paper presents GLidE-SLAM, a monocular hybrid indirect-direct framework that addresses this by architectural separation: the system performs GPU-accelerated direct tracking on intermediate frames, while reserving the full indirect pipeline for map extension and global consistency. We leverage highly parallel image-alignment operations for pose-only estimation without depth optimization or map point creation, making the workload suitable for GPU offloading and freeing CPU resources for backend tasks. We implement the direct tracker using vendor-agnostic OpenGL ES~3.1 compute shaders, enabling deployment across a broader range of commodity embedded platforms without requiring CUDA support. To our knowledge, this is the first complete direct photometric pose estimator realized via compute shaders for embedded-class devices. Experiments on target platforms demonstrate up to 9$\times$ higher frame rates than the CPU-only baseline while maintaining trajectory accuracy and improving practical deployment across commodity resource-constrained hardware.
arXiv:2607.16586v1 Announce Type: new
Abstract: Although standard auction mechanisms help truthfully reveal preferences of bidders, they can inadvertently result in unbounded harms when they fail to account for externalities caused by bid allocations affecting non-bidders. Ad markets, that buy and sell user attention represent such auctions. This research explores a welfare improving auctioneer's audit-and-penalty mechanism that helps screen the worst externalities. We prove this mechanism can internalize externalities formally, then explore social welfare gains empirically.
arXiv:2607.17521v1 Announce Type: new
Abstract: Autonomous driving requires both safe and efficient planning decisions in dynamic 3D environments. Although recent Vision/Video-Action models learn policies directly from visual observations and scale well with advances in vision transformers and large-scale training data, they often lack explicit geometric grounding and future-aware spatial guidance, limiting their ability to balance collision avoidance and driving progress. In this work, we propose GeoWorldAD, a geometry world action model that grounds trajectory planning in ego-aligned 3D space and anticipates short-horizon scene evolution with latent future geometry tokens. Present geometry provides essential spatial constraints for safe planning, while future geometry reveals how surrounding agents and ego-centric free space may evolve, reducing overly conservative decisions without sacrificing safety. To efficiently exploit these geometric cues, GeoWorldAD progressively aggregates multi-scale present geometry and latent future geometry through iterative trajectory refinement. Experiments on NAVSIM v1 and v2 demonstrate state-of-the-art performance, highlighting the effectiveness of explicit 3D geometry grounding and future geometry world modeling for safe and efficient autonomous driving.
arXiv:2607.18178v1 Announce Type: new
Abstract: Software Transactional Memory (STM) systems allow developers to more easily exploit multicore architectures by wrapping arbitrary sequential code in transactions that are executed concurrently. In recent years, the performance of STM systems has approached that of hand-tuned data structures through techniques that avoid unnecessary aborts and exploit the semantics of underlying data structures.
Despite achieving excellent performance, most STM systems do not fully address the concerns they targeted in the first place: safety, usability, and generality. In particular, these systems place restrictions on the data types that may be updated transactionally, such as requiring that these types fit within a word, and can require modification of data layout. Moreover, most STM systems abort transactions in the middle of client code to ensure correctness. This can cause space leaks and other bugs not present in the original code.
We present ustm, a novel STM system addressing all of these shortcomings while still maintaining excellent performance, all within ~300 lines of code. uSTM supports general types while maintaining data layout. Aborts are deferred until the end of the transaction, allowing client code within a transaction to terminate normally. To ensure that uSTM guarantees opacity, we implement a novel timestamping algorithm we call split-increment timestamps.
We compare the performance of uSTM to a variety of state-of-the-art (SOTA) STM systems, demonstrating that uSTM matches or outperforms the SOTA on a variety of workloads.
arXiv:2607.17842v1 Announce Type: new
Abstract: Recent breakthroughs in 3D Gaussian Splatting (3DGS) have advanced neural rendering with high fidelity and speed. However, its performance degrades significantly in large-scale scenes due to the computational burden of tile-based rasterization. Existing optimization efforts either require costly scene re-training or focus on narrow aspects of the pipeline, overlooking critical inefficiencies in real-world deployments. Through a comprehensive analysis, we identify three primary sources of redundancy and low GPU utilization: redundant inter-frame pre-processing, viewpoint-based occlusion redundancy, and severe tile-level load imbalance. To address these issues, we propose CaT-GS, a novel and efficient 3DGS rendering pipeline. CaT-GS introduces a speculative multi-frame preprocessing method to eliminate redundant computations across consecutive frames, and an inter-frame caching mechanism to eliminate viewpoint redundant rendering stages. Furthermore, it refactors rasterization tasks with a dedicated kernel to mitigate tile load imbalance, significantly boosting GPU utilization. Extensive experiments demonstrate that CaT-GS achieves a speedup of up to 10 times over the original 3DGS and up to 70% over previous state-of-the-art methods, establishing a new benchmark for high-fidelity, real-time rendering of large-scale scenes.
arXiv:2607.18189v1 Announce Type: new
Abstract: In this study, we explore how pianists can customize Music Minus One (MMO) concerto accompaniments to match their playing style. Bypassing the need for a symbolic score, often not available digitally, we use three types of audio data: solo piano recordings, MMO orchestra-only recordings, and mixed recordings of both piano and orchestra (e.g., from YouTube). The mixed recording serves as an intermediary reference to align the solo and orchestra parts, with only the orchestral part being adjusted through time-scale modification to synchronize with the user's playing. The main challenge with estimating these alignments is the spectral mismatch between recordings containing different musical parts. Motivated by this application scenario, we introduce Dense-Sparse DTW, a variant of Dynamic Time Warping (DTW) that is designed to improve robustness of alignments to spectral mismatch by focusing on aligning a selected subset of audio frames containing prominent timing cues. We collect and annotate data from four piano concerto movements and establish a framework for generating and evaluating customized accompaniment recordings. On this benchmark, we show that Dense-Sparse DTW has better or comparable performance than more complex approaches based on source separation and spectral subtraction techniques.
arXiv:2607.17541v1 Announce Type: new
Abstract: In robotic manipulation studies, grasping is often treated as a binary success or failure problem, usually defined by whether the object simply stays in the hand. For forceful tool use, however, this view is insufficient because grasp compliance becomes a critical factor governing how the hand and tool behave under load. Compliance arises from coupled kinematics, grasp configuration, passive mechanics, and contact conditions, producing nonlinear behavior in which deformation and interaction forces influence each other. Understanding this relationship is essential for predictive models of how a grasped tool and a compliant hand jointly respond to external loading. In underactuated hands, these effects are amplified: such designs offer low cost and adaptive grasping, but make compliance behavior more difficult to model and predict. Our goal is therefore to develop a predictive model for grasped tool behavior during forceful interactions. To address this challenge, we introduce an analytical model informed neural network (AMINN), a hybrid predictive model that combines an analytical mechanics layer with data driven learning to estimate grasp stability and in hand tool displacement under external loading. The model is evaluated on a three finger underactuated robotic hand and shows strong predictive capability with mechanically meaningful outputs across diverse loading conditions. Compared with a black box multilayer perceptron baseline, AMINN also achieves better energy based physical consistency. Beyond prediction accuracy alone, this framework advances physically interpretable learning for robotic manipulation and supports more reliable, safer, and more trustworthy autonomous tool use in safety critical settings during forceful interaction.
arXiv:2607.17829v1 Announce Type: cross
Abstract: This thesis investigates two central directions in algebraic graph theory, with an emphasis on spectral methods: spectral determination of graphs and transitivity properties of generalized-Hamming graphs and their complements. The first part focuses on graphs that are determined by the spectra of associated matrices. We study spectral determination with respect to the adjacency, Laplacian, signless Laplacian, and normalized Laplacian matrices, with particular emphasis on the adjacency spectrum. We survey existing results on graphs determined by their spectrum and develop new proof techniques for establishing spectral uniqueness. In particular, we present new proofs for the spectral characterization of complete bipartite graphs and Tur\'{a}n graphs, as well as some new results related to the spectral characterization of the important family of strongly regular graphs. In addition, we introduce a new family of graphs, called \emph{the graphs of pyramids}, and prove that they are determined by their adjacency spectrum using tools from matrix analysis, such as Cauchy's interlacing theorem and Schur complements. The second part of the thesis studies generalized-Hamming graphs, a family of Cayley graphs that generalize the sub-family of Hamming graphs, and their complements. We classify the parameters for which these graphs are edge-transitive or even distance-transitive. Our analysis combines spectral methods, group-theoretic arguments, and techniques from the theory of association schemes. As an application, we derive closed-form expressions for the Lov\'{a}sz $\vartheta$-function of generalized-Hamming graphs and their complements whenever either the graph or its complement is edge-transitive. Overall, the results demonstrate how spectral methods provide powerful tools for understanding the structure and symmetry of graphs, and they suggest several directions for further research.
arXiv:2607.16652v1 Announce Type: new
Abstract: This position paper argues that claims about explanation stability are scientifically invalid without cross method validation. Just as statistical significance requires the test statistic to be specified, stability should either be evaluated across multiple attribution paradigms or explicitly scoped to the computational objective of a single method. In controlled chest X ray experiments, DenseNet201, ResNet50V2, and InceptionV3 achieved AUC values above 99%, yet their stability rankings reversed across attribution methods. LayerCAM ranked InceptionV3 as the most stable model, with an IoU of 0.777, whereas GradCAM++ favored DenseNet201 and reduced InceptionV3 stability score by 17.3%. These findings demonstrate that explanation stability is an emergent property of the model method pair rather than an intrinsic characteristic of the model alone. We therefore argue that explanation based claims should be validated across multiple attribution methods and that regulatory submissions should explicitly specify the attribution operators used to avoid creating illusory safety assurances.
arXiv:2607.16808v1 Announce Type: new
Abstract: Event Argument Extraction (EAE) converts documents into structured event records by identifying argument spans and assigning them schema-defined roles. Document-level EAE is challenging due to long-range dependencies between triggers and arguments, cross-sentence context, and strict role constraints, which often lead to boundary errors, uncertainty in roles, and inconsistencies with restricted schemas.
In this paper, we study whether mid-sized open LLMs can perform schema-constrained EAE reliably at the document level on MAVEN-ARG. Our approach combines (i) role-set injection in prompts for schema compliance, (ii) parameter-efficient supervised fine-tuning (LoRA) using the same JSON-only interface used at inference, and (iii) deterministic decoding with post-processing that validates JSON, filters invalid roles, de-duplicates arguments, and aligns spans to the document window. Under the official MAVEN-ARG evaluator, fine-tuned mid-sized open models outperform previously reported GPT baselines across mention, entity-coreference, and event-coreference evaluations; our best model (Phi-4, 14B) reaches 42.39\% F1 at the event-coreference level. Code to reproduce experiments is publicly available at https://github.com/dessertlab/EAE/.
arXiv:2607.16931v1 Announce Type: new
Abstract: China emits the most methane of any country worldwide, but there are large uncertainties in recent emissions trends, sources, and the potential impacts of policy actions. This study focuses on a period when the government initiated ambitious methane control efforts, linking sectoral policies with atmospheric evidence on sectoral, sub-national, and seasonal emissions during 2019--2024. We quantify daily methane emissions from China using a regional atmospheric inverse model with TROPOMI satellite observations. Our results reveal an average methane emissions increase rate of 0.3 Tg yr$^{-2}$ in Eastern & Central China, likely a milder trend than in the 2010s. Coal industry methane emissions intensity declined for the first time (-3.2% yr$^{-1}$) despite rising production, possibly associated with diverse policy instruments, mandates, and incentives. We further highlight two emerging challenges for future mitigation: leaks from expanding urban gas use amid the energy transition and rising agricultural emission yet with substantial uncertainty in estimates. Lastly, declining emissions intensity of coal mines points to the future role of targeted mandates and incentives in encouraging methane reduction for other sectors.
arXiv:2607.16548v1 Announce Type: new
Abstract: Cross-chain DAOs face unique security challenges that go beyond traditional single-chain vulnerabilities. This paper identifies and categorizes four critical attack vectors in cross-chain DAO governance: bribery attacks, token control exploits, human-computer interaction deceptions, and protocol vulnerabilities. We propose a comprehensive security framework with a multi-layered architecture that integrates cryptographic trust anchors, fraud-resistant consensus mechanisms, and decentralized validation techniques to address these threats. Our framework introduces novel components, including a Governance Kernel with on-chain rule verification, a Cross-Chain Trust Layer using threshold cryptography, and a Resilience Layer offering time-locked decision reversals and progressive dispute resolution. By establishing a structured set of countermeasures, this work lays the foundation for secure, transparent, and attack-resistant governance across diverse blockchain environments.
arXiv:2607.17893v1 Announce Type: new
Abstract: Connected vehicles increasingly embed AI accelerators, offering a substantial yet volatile source of supplemental compute near the network edge. Unlike provisioned MEC hosts, vehicular resources are highly dynamic: vehicles may leave the cell, become locally occupied, or offer heterogeneous compute capacities. Therefore, exploiting vehicular resources requires making admission decisions without knowing the compute capacity that will be available during task execution. We present SMART, an uncertainty-aware admission mechanism that enables an \acs{ETSI} \ac{MEC} orchestrator \textit{to opportunistically exploit vehicular compute under predictive uncertainty.} SMART predicts future vehicular capacity using a \ac{BNN} -- whose uncertainty estimates are the best calibrated among the evaluated forecasters at the nominal 95\% level, and incorporates its calibrated predictive uncertainty into a chance-constrained admission program reformulated through \ac{SAA} and \ac{CVaR} approximations. Under a compute-only admission model, SMART admits 95.6\% of tasks while maintaining a median capacity-violation rate of about 0.81\%. It achieves a favorable admission-violation tradeoff compared with seven reactive, mean-only, and uncertainty-aware baselines, by approaching the performance of a compute-capacity oracle under the modeled assumptions. Finally, the sensitivity analyses show that variability in base-station compute availability is a key determinant of admission performance, highlighting the need for a calibrated admission and resource allocation mechanism tailored to opportunistic base-station compute pooling.
arXiv:2607.16554v1 Announce Type: new
Abstract: In multi-task learning (MTL) negative transfer is often considered as an optimization artifact, but it can also be viewed as a consequence of limited shared capacity and weak task redundancy. We investigate this effect through a Capacity--Redundancy (CR) identity that decomposes the sum of per-task predictive informations into joint predictive information that includes label redundancy defined via total correlation (TC), and a residual coupling term that quantifies interference left unresolved by the shared representation. Additionally, we show two key results: (i) a clustering-gap decomposition that gives a necessary and sufficient condition for clustered sharing to outperform global sharing, and (ii) a gradient--TC bridge in a Gaussian multi-task model that formally justifies gradient cosine similarity as a proxy for redundancy ordering. Empirically, we estimate the residual coupling $\Delta$ from validation residual correlations, showing that clustered LoRA substantially reduces $\widehat{\Delta}$, outperforms size-matched random partitions, and results in statistically significant gains with multi-seed confidence intervals.
arXiv:2607.17371v1 Announce Type: new
Abstract: Processing-in-memory-based (PIM) architectures have emerged as a promising solution for accelerating Convolutional Neural Network (CNN) workloads at the edge. Fine-tuning pre-trained CNNs is a common requirement to enhance the model predictive accuracy after deployment. However, the fine-tuning process is computational and memory-intensive, generating a significant amount of intermediate activations. This leads to frequent off-chip memory access, affecting the overall efficiency of the PIM accelerator. Existing fine-tuning strategies are agnostic to the underlying hardware, as they treat all layers equally. In this paper, we propose a hardware-aware framework called ADEPT to accelerate CNN fine-tuning on PIM architectures. Unlike prior fine-tuning methods, ADEPT adaptively trains the model considering both the training overhead and layer sensitivity. Specifically, ADEPT introduces a novel metric that quantifies the trade-off between a block's gradient-based sensitivity and its hardware architecture-specific Energy-Delay Product (EDP), producing platform-dependent fine-tuning configurations. Overall, ADEPT helps reduce the total trainable parameters and the off-chip data access during fine-tuning, while incurring minimal loss in predictive accuracy compared to full-parameter fine-tuning.
arXiv:2607.18227v1 Announce Type: new
Abstract: In line with the prevailing direction of vision research, we explore the integration of both generation and editing capabilities for video and image modalities within a single model. Current approaches to collecting video editing data typically depend on labour-intensive, time-consuming curated procedures--involving object mask annotation, the use of error-introducing pair synthesis via I2V model and ControlNet-like guidance, and VLM-based quality filtering or refinement--and demonstrate limited task scalability. As a result, the diversity of editing tasks remains substantially narrower than that available for image editing models. We develop a pixel-pair temporal warped flow field that can directly generate corresponding video editing samples in real time from image editing samples, and we demonstrate across multiple levels of video editing tasks that a model can learn video editing using only such data. We regard the image modality as a particular form of the video modality. Accordingly, we design a modality mimic generation loss and a modality mimic editing loss to relatively align the capabilities--and thereby the output distributions--of the two modalities through mutual imitation. Moreover, language-based visual editing entails the comprehension of the editing instruction and the reference visual content, the localization of the region corresponding to that instruction within the reference visual contents, and the modification of that region alone. Existing approaches predominantly rely on external aids, such as fine-tuning an additional MLLM or explicitly supplying a mask sequence as auxiliary input during inference. In contrast, we aspire for the model to internalize this capability. To that end, we introduce sense-related tasks--for instance, referring expression segmentation--along with corresponding editing-region-aware latent-level loss and attention-level loss.
arXiv:2607.18057v1 Announce Type: new
Abstract: AI coding agents increasingly submit complete pull requests (PRs) with minimal human intervention, shifting software development from AI-assisted to autonomous workflows. As these agents become more prevalent, ensuring the code they generate is adequately tested, by existing tests or by tests the agents write, is critical to preventing regressions, yet little is known about testing in agentic PRs. To address this gap, we analyze 4882 agent-generated PRs from the AIDev dataset (532 Java and 4350 Python PRs) produced by five coding agents. We study (i) how often agents include test changes and (ii) how well covered are code changes by existing and agent-written tests. Agents include test changes in only 49.6% of PRs that change code under test files. Existing tests provide an incomplete safety net: they cover 61.5% of agents' changed executable lines in Java and only 27.0% in Python, where 64.8% of PRs have no changed line executed by any existing test. Agent-written tests improve coverage over existing tests, but only in a minority of PRs: 35.9% of Java and 22.5% of Python Code + Tests PRs show a coverage gain. Across both languages, error-handling constructs (e.g., try and catch blocks) are the most consistently under-tested, with miss rates reaching 86.0% in Java and 81.0% in Python. These findings motivate coverage-aware development practices, coverage feedback loops for coding agents, and evaluation benchmarks that measure test quality to better help agents reliably test their own code.
arXiv:2607.16571v1 Announce Type: new
Abstract: Helping seems to make a lock wait-free: wrap the critical section in an idempotent thunk that any process can finish once the holder stalls. Yet helping protects the system, not the call. An overwritable candidate lets later requests bump one another in sequence, so a call can be forced to help newcomer after newcomer and never return, while point contention never exceeds two.
We ask whether a call can instead be charged only for the requests active when it takes its ticket, never for what arrives afterward. We show that the answer is yes. SeniorLock is a deterministic helpable thunk lock in which a call with ticket-time seniority $\beta$ finishes in $O((\beta+1)(T+1))$ local shared-memory steps, where $T$ bounds one thunk's cost, independently of later invocations. We call this guarantee \emph{retrospective wait-freedom}.
The same lock doubles as a universal construction we call SeniorObj: it turns any deterministic sequential object whose operations are bounded, concurrently idempotent thunks into a retrospective wait-free one, with no copying of its representation and, when no senior is active, at essentially the native cost of the operation.
arXiv:2607.17386v1 Announce Type: new
Abstract: Recent advances in Multimodal Large Language Models (MLLMs) have significantly improved remote sensing (RS) multimodal understanding. Language-conditioned segmentation is crucial for fine-grained target understanding in Unmanned Aerial Vehicle (UAV) videos. However, this task remains challenging due to the prevalence of small, visually ambiguous targets and dynamic aerial perspectives. In this paper, we propose SkyVLaM, a multimodal large language model for UAV video understanding. SkyVLaM constructs sparse tokens directly from patch-level video representations through a temporal basis perceiver, regularizes the sparse basis to encourage complementary temporal cues, and adaptively selects a temporally coherent dense segment for high-resolution inspection. The resulting sparse and dense tokens are jointly processed by a large language model for query-conditioned segmentation. We further build SkyVid, consisting of SkyVid-VGCG and SkyVid-RVOS for video grounded conversation generation and referring video object segmentation, respectively. SkyVid contains 101 videos, 33.6K frames, and 1.53M pixel-level object instances. Experiments show that SkyVLaM provides a more effective allocation of the visual token budget and improves language-conditioned video segmentation in UAV scenarios.
arXiv:2607.17389v1 Announce Type: new
Abstract: The paper investigates the possibility of predicting function-inlining decisions in compact compilers, source-to-source tools, and interpreters where the reuse of GCC or LLVM optimization infrastructure is impractical. The relevance of this work is determined by the need to transfer mature inlining heuristics to systems with limited compiler infrastructure, restricted runtime dependencies, and reduced access to target-specific analysis. Existing production compilers already contain strong inliners, but their decisions depend on internal intermediate representations (IRs), pass ordering, target models, and analysis stacks that are difficult to reproduce in lightweight systems. To overcome these constraints, we propose a portable inlining-prediction framework. Production compiler diagnostics serve as supervision; a separate extractor reconstructs caller-callee callsites, prepares sterile source snippets, normalizes them into a universal AST, optionally lowers them to a lightweight structural IR, and exports scalar features for model training. Thus, a trained predictor can be emitted as ordinary C code without a compiler-runtime dependency. To evaluate the proposed framework, we constructed a dataset comprising 336,938 callsites from fifteen open-source C projects, including 79,287 compiler-reported inline events. A comparison of several tabular models is performed using project-aware validation. Under leave-one-project-out validation, CatBoost reaches ROC-AUC 0.928 and PR-AUC 0.713; after threshold tuning, F1 improves from 0.670 to 0.729 and the false-positive rate drops from 0.192 to 0.084. Feature analysis shows that most signal is concentrated in source locality, explicit inline intent, callee size, side effects, branch and call structure, signature shape, and callsite argument shape.
arXiv:2607.18231v1 Announce Type: new
Abstract: Vision-language-action (VLA) models have achieved impressive generalization in robotic manipulation, and recent memory-augmented VLAs have relaxed the Markovian assumption by conditioning on past images or language summaries. Vision-based memory approaches address this by conditioning on sampled past image frames, but they are computationally expensive and fundamentally limited when temporal events are visually ambiguous, e.g., pushing a button multiple times with small movements. We propose FM-VLA, a VLA model with force-based memory, enabling temporal context reasoning for non-Markovian, contact-rich manipulation. We encode force histories into compact force memory tokens with a variational autoencoder (VAE) pretrained with force time series reconstruction. By projecting force latent representations and short state history as additional conditioning tokens to the action expert module, we enable VLAs to leverage accumulated contact event history to guide manipulation. We evaluate FM-VLA on three memory-dependent tasks, including finding a hidden block, pressing a button, and wiping a dish for a specific number of times. Our lightweight force memory achieves over 80% success rate with minimal inference overhead, significantly outperforming baseline approaches. Project page: https://qft-333.github.io/FM-VLA-Page/
arXiv:2607.16487v1 Announce Type: new
Abstract: As IoT and embedded devices proliferate across various domains, securing their firmware has become critical. Fuzzing offers a systematic approach to uncovering vulnerabilities in firmware, and coverage feedback can improve its effectiveness by guiding exploration. However, many devices make coverage information impossible to obtain by preventing firmware extraction, instrumentation, or accurate emulation; in such cases, testers are left with only inefficient black-box fuzzing. In this paper, we present an approach that leverages electromagnetic (EM) side-channel emanations to guide firmware fuzzing in purely black-box settings. However, turning raw EM measurements into reliable guidance is challenging: EM traces are noisy, and timing jitter causes corresponding features in different traces to shift in time. We address these challenges by combining frequency band selection based on the activity-to-idle signal contrast with dynamic time warping to align per-input traces and detect sustained divergence, while maintaining scalability by organizing executions in a tree structure based on their divergence times. We evaluate our approach on four real firmware targets and demonstrate that EM-derived feedback enhances path exploration, yielding higher code coverage than unguided fuzzing.
arXiv:2607.17098v1 Announce Type: new
Abstract: Speech recordings used for dementia detection inherently expose speaker identity, raising critical privacy concerns. Existing methods typically address only singular threats and fail to resolve the privacy--utility trade-off. We propose a multi-level framework designed to neutralize two distinct eavesdropping vectors. At the signal level, a Cumulative Signal Attack (CSA) concentrates perturbations in keyword-aligned regions to maximize transcription error (Word Error Rate WER = 1.00) while preserving vital prosodic biomarkers. At the feature level, a Gradient Reversal Layer (GRL) with Mutual Information (MI)-guided noise injection suppresses speaker-discriminative dimensions while retaining dementia-relevant diagnostic structure. Evaluated on the DementiaBank Pitt Corpus, our framework achieves near-chance speaker identification (Equal Error Rate EER = 0.59, F1 = 0.003) while maintaining strong dementia classification performance (F1 = 0.78, AUC = 0.86).