arXiv:2607.15994v1 Announce Type: new
Abstract: A 140-GHz transmitter with direct raised-cosine-like envelope shaping and wide-range phase tuning is presented in 90-nm SiGe BiCMOS. The proposed architecture relaxes conventional baseband pulse shaping and high-speed digital-to-analog converters by directly synthesizing 3-level and 5-level RF envelope states that approximate a raised-cosine waveform, enabled by a 23-dB modulation dynamic range. Measured results demonstrate data rate up to 8 Gbpers with 34-dB sidelobe suppression. To support scalable phased-array applications, an injection-locked-oscillator phase-tuning path is also integrated, providing 22.5deg digital phase resolution together with continuous analog tuning over more than 360 deg. The transmitter achieves up to 2 dBm output power and combines direct RF envelope shaping with wide-range fine phase control for sub-THz wireless links
Science Journals
arXiv:2607.16074v1 Announce Type: new
Abstract: The post-training of Vision-Language-Action (VLA) models is essential due to the diversity of simulators, robot embodiments, and task objectives. Existing compute services, whether offered as direct accelerator rental or batch-workload submission, typically allocate an exclusive set of GPU and CPU resources to a single tenant. While this paradigm maximizes client flexibility, it burdens users with infrastructure adaptation, and the fixed card-hour accounting model renders short or bursty workloads both expensive for tenants and inefficient for the service provider. To address these challenges, we present JoyNexus, a unified service for multi-tenant VLA supervised fine-tuning, reinforcement learning, and evaluation. JoyNexus decouples the Training Model Service, Inference Model Service, and Environment Service, each accessed through APIs and backed by resident shared base models with tenant-specific slots. Tenants can directly invoke high-level semantic APIs for training, rollout, and evaluation, or compose custom algorithms using lower-level APIs and their assigned endpoints. Multiple tenants submit workloads concurrently; their action modules, optimizers, rollout records, and policy versions remain isolated, and the service is scheduled by the global Training Queue and Inference Queue. To further improve multi-tenant training efficiency, JoyNexus introduces group batching for heterogeneous VLA data schemas that share a compatible model-facing prefix, enabling a single shared backbone forward pass over grouped samples. Finally, we evaluate JoyNexus through workload simulation and a group-batching pipeline in a realistic embodied scenario. Results show that, compared with isolated single-tenant execution, JoyNexus reduces aggregate GPU time and improves service utilization via cross-tenant scheduling on shared resources.
arXiv:2607.15762v1 Announce Type: new
Abstract: Modern compiler codebases (e.g., GCC and LLVM) are large and complex, making comprehensive coverage across diverse code regions highly challenging. Most existing test generation techniques ignore characteristics of the target code, producing test programs that exercise only a limited subset of it. Consequently, substantial compiler regions remain insufficiently tested, leaving persistent long-tail coverage gaps that survive across releases. Even existing white-box techniques achieve limited coverage on large-scale compilers. To improve compiler coverage, especially for hard-to-reach edge regions, we present GapForge, a targeted LLM-based test generation technique that reasons about coverage gaps. Unlike program-driven techniques that generate diverse inputs without modeling which regions they exercise, and unlike whole-file summarization that yields coarse guidance, GapForge treats coverage gaps as explicit region-level targets in three steps. First, it prioritizes files via coverage-driven scoring that favors large, undercovered files. Second, it pairs each uncovered line span with its enclosing covered context and performs path-difference analysis to infer fine-grained triggering requirements: the program structures and compilation options needed to reach the uncovered region. Third, it synthesizes prompts from these requirements and previously failed prompts, using coverage feedback to guide next-round selection. On GCC 14.3.0 and LLVM 19.1.0, GapForge significantly outperforms eight state-of-the-art techniques. Within 72 hours, it achieves 68.13% and 69.11% coverage on core compiler modules in GCC and LLVM, surpassing the white-box technique WhiteFox by 24,736 and 19,798 additional lines, respectively. Moreover, GapForge discovers 12 real-world compiler failures (5 in GCC, 7 in LLVM), including 8 crashes and 4 miscompilations, with each component contributing to its performance.
arXiv:2607.15932v1 Announce Type: new
Abstract: Advanced static program analysis requires reasoning on the semantics of non-trivial program behaviors (e.g., pointers and complex data structures such as lists and sets, functions, and objects) and how they affect the memory. In most programming languages, static and dynamic allocations are typically managed by the stack and the heap, respectively. However, how allocations behave and how the memory is managed at runtime can vary significantly depending on the programming language being analyzed. Proper handling of these aspects is essential, as an accurate memory model enables the detection of critical issues such as buffer overflows and underflows, use-after-free errors, and null pointer exceptions prior to execution, that is, before such erroneous behaviors occur. In this paper, we propose and formalize a generic memory framework to handle stack and heap memory during the analysis, that is able to support various behaviors from different programming languages (e.g., C, C++, Java, and Python), while remaining parametric, allowing different memory and value analyses to be independently chosen and combined. It relies on the Abstract Interpretation theory and enables sound approximation of different memory models and program behaviors. We introduce a split state abstraction that separates value and memory analyses into two modular abstract domains. These domains interact through a set of memory identifiers, along with a set of operations defined by the domains to manipulate them, allowing the framework to capture both value information and structural memory relationships.
arXiv:2607.16125v1 Announce Type: new
Abstract: Climate services are essential for risk management and economic planning. This paper presents a century-long simulation performed with the ARP-GEM global atmosphere model at 2.6 km horizontal resolution. This simulation was specifically designed to complement and extend the set of simulations used in the latest version of the French climate services. A key feature is its global coverage at kilometer-scale resolution, enabling the representation of phenomena that depend on such fine scales. This is especially relevant for French climate services, as many processes - particularly over islands and resolution-sensitive regions - are currently overlooked in global simulations. Additionally, France's overseas territories are globally dispersed, and this approach allows their representation within a single framework. A limitation of this simulation is the use of prescribed sea surface temperatures due to the lack of ocean coupling; this will be addressed in future work. The present study demonstrates the feasibility of this approach and highlights the benefits of this new generation of climate modeling for climate services.
arXiv:2501.05779v3 Announce Type: replace
Abstract: We introduce a novel toy model for shear flows, exploiting the spatial intermittency and the scale separation between large-scale flows and small-scale structures. The model is highly sparse, focusing exclusively on the most intense structures, which are represented by vortons: dynamically regularized quasi-singularities that experience rapid distortion from the large-scale shear. The vortons, in turn, influence the large-scale flow through the sub-grid stress tensor. Despite its simplicity, the model displays an interesting transition between two distinct regimes: (i) a laminar regime, where dissipation is entirely attributed to the large-scale flow, and the vortons dynamics is essentially diffusive, and (ii) a turbulent regime, in which most of the dissipation arises from the vortons. These regimes correspond to different scalings of dissipation and the Grashof number as functions of the Reynolds number, with power-law relationships that resemble those observed in classical turbulence.
arXiv:2607.15832v1 Announce Type: new
Abstract: Can a model be identified from one experiment? We study analytic systems that are linearly parameterized by a combination of prescribed dictionary terms, such as partial differential operators and dynamical systems. For a single input-response pair, recovery is possible exactly when the evaluated dictionary terms are linearly independent. We prove a sharp zero-one law: either no input uniquely determines the coefficients, or almost every random input sampled from a nondegenerate Gaussian measure does. This dichotomy reduces one-shot system identification to a question about degenerate inputs and provides an a posteriori certificate for any recovered model. Numerical examples recover dynamical systems, nonlinear partial differential equations, and structured matrix families from single trajectory data, while also detecting when an extra probe is necessary.
arXiv:2607.15776v1 Announce Type: new
Abstract: OWL ontologies provide a formal knowledge representation framework that enables semantic reasoning, and have been widely adopted across domains such as healthcare and bioinformatics. In practice, however, real-world ontologies are often incomplete, which pose challenges for reasoning. In this work, we focus on a fundamental subsumption reasoning problem: given an incomplete ontology and a candidate (non-entailed) subsumption, determine whether the subsumption is semantically plausible and, if so, providing a logically sound explanation containing potential missing axioms. This task unifies subsumption verification with ontology abduction, and generalizes the latter by removing the need for a predefined candidate set of missing axioms. To address this subsumption reasoning problem, we propose NeurOWL, an end-to-end neuro-symbolic framework that jointly performs verification and abduction, leveraging both formally defined semantics and textual semantics through Large Language Models and ontology embeddings. We evaluate NeurOWL on real-world ontologies across multiple domains, demonstrating strong and robust performance across different domains.
arXiv:2607.15358v1 Announce Type: cross
Abstract: We demonstrate an asymptotic gate complexity improvement in first-quantized ground-state energy estimation of electronic structure Hamiltonians in a plane wave basis by employing the sum-of-squares spectral gap amplification protocol. The improvement relies on identifying a sum-of-squares representation of the Hamiltonian which provides a lower bound certificate and low cost block encoding that leads to a provably lower quantum phase estimation gate cost. This is achieved by using a sum-of-squares operator generated by the total charge density operator resulting in a block encoding normalization improvement of $\lambda = \mathcal{O}\left(\eta\Delta^{-1.5}+\eta^{1.5}\Delta^{-1} \right)$ compared to prior work $\lambda = \mathcal{O}(\eta\Delta^{-2}+\eta^2\Delta^{-1})$ where $\eta$ is the number of electrons and $\Delta$ is the simulation grid spacing. The asymptotic reduction in block encoding normalization and similar block encoding costs to prior work is demonstrated to reduce resource estimates for materials and chemical systems by a factor of $2 - 44\times$ corresponding to the lowest cost estimates for ab initio materials simulation.
arXiv:2607.15367v1 Announce Type: new
Abstract: Desktop voice assistants are still dominated by cloud pipelines that ship raw audio off the machine and expose a fixed set of skills. We describe AnovaX, a small local-first assistant that runs entirely on the user's computer and treats the desktop itself as its action surface. A single Python process wires together a wake-word gate, a speech pipeline, an LLM planner (Gemini) that emits a JSON plan of tool calls, a whitelist-and-denylist safety layer, a multi-agent orchestrator that translates each plan into typed child agents on a bounded thread pool, and an adaptive recovery loop that takes over whenever a core step fails. Every tool corresponds to a specialized agent class (AppAgent, TypingAgent, BrowserAgent and six others) with its own timeout, retry policy, and shared-resource locks. A recursive MetaAgent lets the planner delegate a sub-goal back to itself, capped at two levels of nesting. The recovery loop uses a compact ReAct-style prompt and hides Gemini's latency behind speculative execution of read-only tools. A companion Flask server exposes a phone-friendly remote over the local WiFi, mirrors every agent lifecycle event to the phone in real time, and streams the laptop's screen back over MJPEG so the user can watch remote commands land as they run. The point of the project is less to compete with Siri or Alexa than to show that a legible, few-thousand-line assistant is enough to open apps, type into them, run searches, coordinate concurrent actions, recover from single-step failures, and be driven entirely from a phone in another room -- without the LLM ever touching the keyboard.
arXiv:2607.15778v1 Announce Type: new
Abstract: Video Large Language Models (Video LLMs) have made significant advancements in various video understanding tasks. However, long-video scenarios remain challenging due to the tension between limited visual token budgets and the need to capture multiple key events. Existing approaches typically process long videos in two stages, i.e., i) select keyframes and ii) perform detailed perception, which exhibit limitations: they lack a modular mechanism for adaptive capacity allocation and self-correction, resulting in unreliable modeling. To tackle these challenges, we propose MoD-VLLM, a novel Modularized Dynamic-Granularity Video LLM framework for multi-event long video understanding, which unifies temporal grounding and semantic understanding iteratively and self-reflectively. Specifically, we propose a Positive-Negative Video Segments Grounding module and a Modularized Dynamic-Granularity Reflection module, which form a closed loop to progressively localize the question-related video segments. The grounding module instructs a Video LLM to distinguish relevant from irrelevant video segments based on the video question. The reflection module employs a modularized scheduler that dynamically selects fine-grained encoding for relevant positive segments to capture detailed perception and coarse-grained encoding for negative segments to maintain global context. We further propose a dynamic-granularity reinforcement learning strategy, allowing MoD-VLLM to learn optimal grounding policies and dynamic granularity visual representation jointly. Moreover, we propose MEventBench, a challenging Multi-Event Long Video Benchmark for complex long video reasoning. Extensive experiments on several long video understanding benchmarks and our MEventBench demonstrate that MoD-VLLM significantly outperforms state-of-the-art baselines.
arXiv:2607.15331v1 Announce Type: new
Abstract: Generalized few-shot 3D point-cloud segmentation (GFS-PCS) asks a model to segment a scene into many base classes seen at training time and a set of novel classes. The state of the art reaches novel classes by reconciling a dense but noisy 3D vision-language prior with the few-shot support, but it pays for this with base 3D labels, per-episode training, and the support annotations themselves. We ask how far the same reconciliation can go with none of these: no training, no 3D labels, and not even the few-shot support. We pair a frozen 3D vision-language model (RegionPLC) as a dense prior with a frozen promptable concept segmenter (SAM3), prompted by the bare novel class names and lifted from posed RGB views, and reconcile the two by cross-view consistency: a point becomes novel only when enough of the views that see it agree. On the ScanNet200 GFS-PCS benchmark this fully training-free, open-vocabulary pipeline improves novel mIoU by +2.6 over the training-free dense prior while holding base accuracy within 0.5, and recovers a third (33%) of the novel-class gap to the trained state of the art that uses far more supervision. We further show that injecting the few-shot support into the pipeline, as a fusion gate and as a prototypical dense classifier, adds nothing over consistency alone and in fact degrades it through the classifier, which is why the method needs no support at all. On the harder ScanNet++ benchmark, where the dense prior is far weaker on novel classes, the same pipeline nearly doubles novel mIoU (+15.7, from 16.2 to 31.9) at a 1.7 base cost, lifting the harmonic mean from 21.5 to 31.1
arXiv:2607.15781v1 Announce Type: new
Abstract: Geospatial datasets support applications from urban planning to climate modeling, yet consistent assessment of FAIR compliance is difficult. Existing evaluators use different rubrics and evidence sources and may fail on JavaScript-rendered pages or repository-specific identifiers. For 50 datasets from 10 repositories, the standard deviation of normalized scores across available tools averages 15.0 percentage points and reaches 30.3 for one dataset. Because these outputs are not equivalent measurements, we use them to characterize disagreement and failure modes, not comparative accuracy. We present AgentFAIR, a multi-agent framework combining structured metadata extraction with 13 sub-principle-specific LLM evaluators. Each produces a 0-3 maturity score, cited evidence, and recommendations; a critic checks evidence and consistency and can request targeted re-evaluation. Mean Findability, Accessibility, Interoperability, and Reusability scores are 79.7%, 70.4%, 45.3%, and 72.0%. Rank correlations with four baseline tools range from 0.31 to 0.61; the FAIR-enough comparison is not statistically significant. On a 10-dataset repeated-run subset, sub-principle agreement averages 89% (standard deviation: 3 percentage points), versus 71% without the critic. A preliminary 15-dataset expert study yields Fleiss' kappa of 0.71 and 82% alignment with expert consensus. API cost is approximately USD 0.054 per dataset. These results support auditability and feasibility, while the limited benchmark, incomplete ablations, and single-model-family validation constrain claims about accuracy and generalization.
arXiv:2607.15782v1 Announce Type: new
Abstract: Designing hardware circuits resistant to side-channel attacks increasingly relies on simulation to predict device leakage before fabrication. Current functional verification simulators are designed for extended correctness-checking runs and are ill-suited for producing large numbers of short trace collections with slight input variants needed for side-channel analysis. We present Vogls: an open-source Verilog simulator built for side-channel analysis, that is the first simulator to combine compiled-code performance, full-timing simulation, and fine-grained control over the simulation state. Vogls simulates a timing annotated gate-level AES design 5.9 times faster than Icarus Verilog and is only 30% slower than Verilator on a PicoRV32 RTL design while also offering a more accurate timing model. Vogls provides a Python interface that allows simulations to be paused, forked, inspected, and mutated around points-of-interest without modifying the hardware design and thereby preserving timing and leakage characteristics. Furthermore, a trace-collection workflow runs setup code once and forks at the point-of-interest, eliminating the need to re-simulate from reset for every trace. A case study demonstrates Vogls on a differential power analysis attack that successfully recovers the key at RTL, GTL and full-timing GTL abstraction levels.
arXiv:2607.15783v1 Announce Type: new
Abstract: We demonstrate a loss-enabled mechanism that can realize single electromagnetically induced transparency (EIT) analogue in periodic waveguide-resonator structures. Unlike previously reported passband-based EIT analogues, which arise from compressed passbands in the intrinsic-loss-free limit, the proposed mechanism exploits resonator intrinsic loss to generate an isolated transparency mode within the original bandgap. The resulting EIT state remains immune to finite-size mode discretization, enabling robust single-frequency lasing through distributed resonant feedback while requiring a threshold modal gain lower than that of an equally long Fabry-Perot laser. Our findings extend classical EIT analogues from passive spectral phenomena to active laser operation.
arXiv:2410.21817v3 Announce Type: replace
Abstract: We address our attention to the numerical time discretization of stochastic Poisson systems via Poisson integrators. The aim of the investigation regards the backward error analysis of such integrators to reveal their ability of being structure-preserving, for long times of integration. In particular, we first provide stochastic modified equations suitable for such integrators and then we rigorously study them to prove accurate estimates on the long-term numerical error along the dynamics generated by stochastic Poisson integrators, with reference to the preservation of the random Hamiltonian conserved along the exact flow of the approximating Wong-Zakai Poisson system. Finally, selected numerical experiments confirm the effectiveness of the theoretical analysis.
arXiv:2607.15801v1 Announce Type: cross
Abstract: This paper solves the targeting problem focusing on accuracy, computational efficiency, and reliability. The trajectory optimization problem is first recast as a polynomial optimization problem (POP) by leveraging differential algebra to compute high-order Taylor expansions of the nonlinear dynamics and constraints. Moment-sum-of-squares (SOS) optimization is then utilized to solve this POP. A convex formulation based on a second-order expansion of the dynamics is also proposed. For impulsive targeting, the moment-SOS and convex approaches are compared against traditional nonlinear programming (NLP) solvers and map inversion techniques. Results indicate that the moment-SOS approach provides solutions as accurate as traditional NLP, but with the critical advantage of guaranteeing convergence to the global optimum under mild assumptions. Furthermore, the method excels at handling large maneuvers and long propagation times, conditions in which standard linear approximations rapidly degrade. To demonstrate its versatility, the methodology is extended to a continuous low-thrust station keeping (SK) scenario in the Earth-Moon Circular Restricted Three-Body Problem. The algorithm's performance is then evaluated in the presence of significant state errors. The ability to directly handle non-convex constraints and recast complex, nonlinear dynamics into formulations with reliable convergence properties makes the moment-SOS approach suitable for autonomous onboard applications.
arXiv:2607.15533v1 Announce Type: new
Abstract: Haptic interfaces for the wrist and forearm offer an attractive alternative to hand-worn devices as they are simple to wear, leave the hands free for interaction with the real world, and interfere minimally with natural arm motions. To be useful in real-world settings, however, such devices must balance functionality, wearability and comfort, all while being fully untethered with minimal mass and volume. In this work, we present CASAband, a haptic wristband that integrates compliant amplified shape memory alloy actuators (CASA) into a multi-layered textile wristband to deliver spatial and temporal haptic feedback. CASAband operates completely untethered, generates no noise, and has a total mass of 63 g. The device incorporates four actuators that can generate up to 1.7 N of blocked force and 3.2 mm of free displacement with an operating bandwidth ranging from 1.34-6.59 Hz depending on the applied voltage. We conducted a perceptual study and determined that users could identify the location of a single haptic cue around their wrist and discriminate among several patterned cues with over 90% accuracy on average, highlighting that CASAband can be a suitable wearable interface to deliver information for real-world guidance and navigation tasks. To highlight the potential use cases for CASAband, we conducted two demonstrations: a pick and place task where the user relied only on haptic communication from a moderator, and an outdoor pedestrian navigation task where the user relied only on directional cues on the wrist. CASAband is one of the first haptic interfaces that balances the tradeoff between form and function and presents new opportunities for haptic feedback in the real world
arXiv:2607.15477v1 Announce Type: new
Abstract: Sleep apnea diagnosis via polysomnography remains resource intensive and relies on time consuming manual data analysis and scoring. Recent work has demonstrated that central nervous system effects of sleep apnea events can be detected through electroencephalogram (EEG) signals. However, most work uses a single feature type on various datasets combined with different classification algorithms. In this work, we present a comprehensive comparison of deep learning architectures and feature representations for automated sleep apnea detection from multichannel EEG on a single dataset of pediatric subjects. We evaluate Vision Transformers and Graph Attention Networks across distinct signal representations: raw temporal signals, short-time Fourier transform spectrograms, coherence based graphs, and two topological data analysis (TDA) derived features. Using age and sex matching of our train and test sets, we train on 2410 pediatric subjects and test on 575 pediatric subjects. We achieve a best test AUC of 0.750 using a vision transformer based model trained on TDA features. Stratified analysis across patient demographics (age, sex, AHI severity) and sleep stages (N1, N2, N3, REM) reveals significant performance variation. Our results demonstrate the feasibility of EEG based automated OSA screening while highlighting essential challenges for clinical deployment.
arXiv:2607.15794v1 Announce Type: new
Abstract: Classical approaches to event-based egomotion estimation, including those adopted by the top-performing teams of the ELOPE challenge, rely on geometric optimization frameworks such as contrast maximization, homography estimation, or dense optical flow combined with analytic motion inversion. This work investigates the geometric structure that emerges inside a multi-modal network for egomotion estimation. Event tensors, inertial measurements, and range signals are fused through a cross-modal attention architecture and trained in a batch setting. We analyze the latent space geometry and attention dynamics, showing that (i) embeddings lie on low-dimensional manifolds aligned with motion variables, (ii) attention weights adapt with angular excitation and visual reliability, and (iii) the fused representation recovers classical observability cues. These results bridge analytical estimation theory and modern data-driven fusion.
arXiv:2607.15380v1 Announce Type: new
Abstract: Electronic health records combine free-text clinical narratives with structured measurements such as vital signs, laboratory values, and comorbidities. Yet most clinical prediction systems still rely on task-specific fusion architectures, pairing dedicated encoders for each modality with learned combination mechanisms that must be re-engineered for every new task and clinical setting. We propose a simpler alternative: convert all patient data, regardless of modality, into a single natural language sequence and fine-tune a pretrained language model end-to-end, with no architectural modification for fusion. We evaluate this approach across three clinically distinct prediction tasks: in-hospital mortality on MIMIC-III, graft failure prediction using longitudinal data from a German transplant center, and emergency triage classification from ambulance records - comparing encoder-based (ModernBERT) and decoder-based (Llama 3.1, Gemma, DeepSeek-R1-Qwen, Qwen3) fine-tuning against established multimodal baselines and, for graft failure, a gradient boosting model currently used in clinical practice for post-transplant patient management. Across all three tasks, unified textual serialization matches or exceeds task-specific multimodal baselines, and outperforms the clinically deployed gradient boosting system on graft failure prediction. These results indicate that a single serialization-based paradigm, without bespoke fusion architectures, is sufficient for multimodal clinical prediction - substantially reducing system complexity while matching or exceeding specialized designs.
arXiv:2607.15935v1 Announce Type: new
Abstract: Deep reinforcement learning (DRL) has a longstanding tradition in addressing the reach-avoid task problem, especially for controlling robotic arms. While this task serves as a baseline environment within the research community, the ability of DRL to effectively learn the each-avoid task in complex and realistic scenarios beyond simplified and restricted tabletop settings remains uncertain. In this paper, we present, for the first time, a comprehensive benchmark for the reachavoid task that accurately captures real-world complexities without simplifications. We demonstrate a diverse range of settings for robotic arm reach-avoid task, which can be used for evaluating DRL research. We achieved this by utilizing the MuJoCo MJX physics engine and parallelizing both the simulation environment and DRL algorithms using the Brax library. We achieved state-of-the-art results with success rates of 96.1% (UR5e) and 98.8% (Franka Emika Robot) for the reach task and 86.8% (UR5e) and 95.2% (Franka) for the static reachavoid task. Our results indicate that while in previous works DRL agents could solve, for example, a reach task in a simplified setting perfectly, their agents performance collapses when evaluated in realistic scenarios. Overall, this work identifies that additional research is still required to claim the successful resolution of the robotic arm reach-avoid task using DRL. The environment and benchmarking code is available as open source at the following link
arXiv:2607.15600v1 Announce Type: new
Abstract: Monocular depth foundation models have demonstrated remarkable generalization capabilities across diverse environments. However, they continue to struggle with metric depth estimation in diverse environments. This limitation stems from the inherent scale ambiguity of single-view inference, leading to misaligned scale predictions even when the relative geometry is accurate. Conversely, recent multi-view foundation models leverage cross-view cues to learn robust scene-level geometry and consistent scale. Yet, these benefits typically vanish during single-image inference, as the absence of explicit geometric constraints causes performance to degrade. To bridge this gap, we propose a novel framework that transfers the scale-aware geometric priors of multi-view models into monocular depth foundation models. Specifically, we introduce an Epipolar Distillation (EpiDistill), an approach utilizing Rectified Stereo Tokens, which enables the single-view prediction model to retain epipolar attention patterns and maintain geometric consistency without requiring multi-view inputs at inference. Experimental results demonstrate that our method significantly improves zero-shot metric depth estimation, particularly on challenging datasets like ETH3D and DIODE where scale alignment is critical. Furthermore, our approach is model-agnostic, consistently boosting the performance of state-of-the-art ViT-based models, including UniDepthV2 and DepthPro.
arXiv:2607.16003v1 Announce Type: new
Abstract: The adaptive Bregman-Kaczmarz method recovers the exact, noise-free solution of a linear inverse problem even when every measurement it queries is corrupted, provided the corruption is fresh, independent and zero-mean. A block version was proposed for parallel hardware, but whether larger blocks actually converge faster was left open. We show the answer hinges on how the block is used: replacing the block sum with a block average collapses the analysis onto a single positive-semidefinite matrix, through which we prove that the guaranteed convergence improves monotonically with the batch size, with a total gain governed by the stable rank of the system matrix. We then address heterogeneous noise by introducing a noise-aware weighting that down-weights unreliable measurements, and prove it is strictly better than uniform weighting whenever the noise is not proportional to the row norms - a condition that essentially never holds in practice. The two improvements are compatible and their benefits combine. Finally, we show the adaptive step size interpolates automatically between a fast initial phase and a slowly vanishing tail that carries the error exactly to zero, and we explain how its hyperparameters can be estimated without knowing the true solution. Numerical experiments under sparse, heterogeneous noise illustrate these findings and confirm that the heuristic estimates produce effective step sizes.
arXiv:2607.15846v1 Announce Type: new
Abstract: Diffusion Transformers (DiTs) have been widely used in many tasks, including image synthesis, video generation, and content editing. However, their multi-iteration inference process leads to performance inefficiency and high energy consumption. Existing acceleration methods primarily focus on reducing temporal redundancy between adjacent timesteps, but often overlook the specific features of DiTs. As a result, these approaches either suffer from great accuracy degradation or fail to achieve high efficiency.
We present DSTAR, a software-hardware co-design framework that accelerates DiT inference by reducing spatial and temporal redundancy. At the algorithmic level, DSTAR introduces a fine-grained mixed-precision quantization method for differential activations in linear operations, significantly increasing the proportion of low-bit computations. Additionally, DSTAR incorporates a sparse attention reuse mechanism to minimize redundant computation in attention layers. For architectural support, we design a specialized hardware accelerator which achieves high efficiency in both latency and energy consumption. Evaluation on seven typical DiTs demonstrates that DSTAR achieves up to 7.33x latency speedup and 41.89x energy savings compared to an NVIDIA A100 GPU, and achieves up to 2.54x latency speedup and 3.68x energy savings compared to SOTA accelerators, without accuracy degradation.