arXiv:2607.17551v1 Announce Type: new
Abstract: Lung ultrasound (LUS) is a bedside tool for assessing pulmonary edema in patients at risk due to heart failure or impaired kidney function. However, automated LUS analysis remains challenging because of speckle noise, imaging artifacts, and operator-dependent acquisition variability. In this work, we present a deep learning framework for multi-class LUS video classification that explores two components: hierarchy-aware training, and anatomy-guided learning. Starting from a strong baseline, we introduce hierarchical training strategies and then introduce pleural line mask supervision to guide model attention toward anatomically relevant regions. We study four clinically relevant classes--healthy, B-lines, consolidations, and mixed B-lines with consolidations--using an open-access dataset of 1,886 videos from 219 patients, evaluated with patient-level five-fold cross-validation. Results show that hierarchy-aware training improves pathological separation relative to flat classification, while mask-guided attention supervision achieves the highest mean macro-F1 of 65.7\% and produces more localized attention patterns. Transfer experiments on the external COVID-BLUeS dataset further show competitive and parameter-efficient adaptation while preserving pleural-focused attention behavior. These findings suggest that combining clinically structured objectives with anatomy-guided supervision is a practical approach to robust, interpretable LUS video analysis. Code and model implementations are available at https://github.com/Alya-Almsouti/LUS-video-classification.
Science Journals
arXiv:2407.04900v2 Announce Type: replace
Abstract: Numerous existing studies have examined the performance of Sample Average Approximation (SAA) in the fundamental newsvendor problem. Despite these advances, critical gaps remain in two aspects. First, existing works focus on the linear-cost newsvendor problem and heavily rely on the quantile expression of the optimal solution. As a result, their analytical methods limit generalizability to more general inventory problems, where the optimal solution is not a quantile of the demand distribution. Second, even within the linear-cost setting, notable gaps exist between the state-of-the-art regret lower bound and upper bound for SAA under various conditions.
In this paper, we generalize the structure of the newsvendor problem to generic convexity conditions and provide a unified regret analysis of SAA for general sequential stochastic optimization problems. Our approach provides further insights to a broader range of data-driven inventory problems, improves both the upper and lower regret bounds, and establishes the regret rate optimality of SAA. Our lower bound identifies the performance limit achievable by any policy for sequential stochastic optimization and inventory management problems, offering important guidance for future policy design in this area. Moreover, in empirical studies, SAA's performance is frequently used as a benchmark for evaluating new algorithms. The regret rate optimality result provides strong support for its role in assessing other data-driven methods, benefiting both practitioners and researchers. Our new analysis techniques enrich the analysis tools for regret upper and lower bounds for data-driven decision-making problems and other general stochastic optimization problems.
arXiv:2408.14917v2 Announce Type: replace
Abstract: Spiking Neural Networks (SNNs) hold great potential to realize brain-inspired, energy-efficient computational systems. However, current SNNs still fall short in terms of multiscale temporal processing compared to their biological counterparts. This limitation has resulted in poor performance in many pattern recognition tasks with information that varies across different timescales. To address this issue, we put forward a novel spiking neuron model called Parallel Multi-compartment Spiking Neuron (PMSN). The PMSN emulates biological neurons by incorporating multiple interacting substructures and allows for flexible adjustment of the substructure counts to effectively represent temporal information across diverse timescales. Additionally, to address the computational burden associated with the increased complexity of the proposed model, we introduce two parallelization techniques that decouple the temporal dependencies of neuronal updates, enabling parallelized training across different time steps. Our experiments across a wide range of pattern recognition tasks demonstrate that PMSN outperforms state-of-the-art spiking neuron models in temporal processing capacity and training speed. Specifically, compared with the commonly used Leaky Integrate-and-Fire neuron, PMSN offers more than 10 times acceleration and a 30% accuracy improvement on Sequential CIFAR-10 dataset, while maintaining comparable computational cost. Our implementation on neuromorphic hardware further demonstrates the deployability of PMSN and highlights its favorable trade-off between effectiveness and efficiency. Therefore, the proposed PMSN presents a promising solution to harness the computational advantages of detailed biological neurons, enabling high-performance and efficient temporal processing on neuromorphic computing systems. Code is available at https://github.com/xychen-comp/PMSN.
arXiv:2408.15518v3 Announce Type: replace
Abstract: This paper presents Dolphin, a novel decoder-decoder architecture for energy-efficient processing of long contexts in language models. Our approach addresses the significant energy consumption and latency challenges inherent in on-device models. Dolphin employs a compact 0.5B parameter decoder to distill extensive contextual information into a memory embedding, substantially reducing the input length for the primary 7B parameter decoder model. Inspired by vision-language models, we repurpose the image embedding projector to encode long textual contexts, effectively treating extended context as a distinct modality. This innovative method enables processing of substantially longer contexts without the typical computational overhead associated with extended input sequences. Empirical evaluations demonstrate a 10-fold improvement in energy efficiency and a 5-fold reduction in latency compared to conventional full-length context processing methods without losing quality of the response. Our work contributes to the development of more sustainable and scalable language models for on-device applications, addressing the critical need for energy-efficient and responsive AI technologies in resource-constrained environments while maintaining the accuracy to understand long contexts. This research has implications for the broader field of natural language processing, particularly in the domain of efficient model design for resource-limited settings. By enabling more sophisticated AI capabilities on edge devices, Dolphin paves the way for advanced language processing in a wide range of applications where computational resources are at a premium. The Dolphin model is publicly available at https://huggingface.co/NexaAIDev/Dolphin.
arXiv:2409.10067v2 Announce Type: replace
Abstract: Fluid dynamics, and flight in particular, is a domain where organisms challenge our understanding of its physics. Integrating the current knowledge of animal flight, we propose to revisit the use of live animals to study physical phenomena. After a short description of the physics of flight, we examine the broad literature on animal flight focusing on studies of living animals. We start out reviewing the diverse animal species studied so far and then focus on the experimental techniques used to study them quantitatively. Our network analysis reveals how the three clades of animals performing powered flight - insects, birds and bats - are studied using substantially different combinations of measurement techniques. We then combine these insights with a new paradigm for increasing our physical understanding of flight. This paradigm relies on the concept of Animal Learning, where animals are used as probes to study fluid phenomena and variables involved in flight, harnessing their natural adaptability.
arXiv:2409.19109v4 Announce Type: replace
Abstract: Geolocation plays a critical role in understanding the Internet. In this work, we analyze and fix operator-misreported geolocation. Using DNS root servers to detect speed of Internet violations, we conservatively infer that at least 2% of vantage points in the largest community-vantage point collection, RIPE Atlas, were not located at their operator-reported geolocation between January 2019 and April 2026. To increase the accuracy of future studies that use operator-reported geolocation, we open source our simple methodology, implement a reporting campaign of violating probes within RIPE, work with operators to fix the misreported geolocations, and release a continually updated dataset of RIPE vantage points that likely misreport geolocation.
arXiv:2410.01021v4 Announce Type: replace
Abstract: Chains of co-B\"uchi automata (COCOA) have recently been introduced as a new canonical representation of omega-regular languages. The co-B\"uchi automata in a chain assign each omega-word its natural color, which depends only on the language itself and not on the chosen automaton representation. Automata in such a chain can be minimized in polynomial time and are good-for-games, making this representation attractive for verification and reactive synthesis. However, in these applications, specifications are usually given in linear temporal logic (LTL). To make COCOA useful, an LTL specification must first be translated into the chain of automata. The only translation currently known proceeds via deterministic parity automata (LTL$\,{\to}\,$DPA$\,{\to}\,$COCOA), where the first step ignores natural colors and requires involved constructions due to Safra or Esparza et al. This raises the question of whether, by exploiting the definition of the natural color of words, one can avoid such constructions and obtain a direct translation from LTL to COCOA.
In this paper, we present a simple yet optimal translation from LTL to COCOA, as well as a variant that translates LTL into DPA. The translation represents a new path from LTL to DPA and exploits the definition of natural colors. It relies on standard operations on weak alternating automata, the Miyano-Hayashi breakpoint construction, the subset construction, and simple graph algorithms. Starting from weak alternating automata, the procedure also applies to specifications in linear dynamic logic. The procedure runs in asymptotically optimal doubly exponential time and produces automata of asymptotically optimal size.
arXiv:2410.18936v2 Announce Type: replace
Abstract: We consider the foundational problem of maintaining a $(1-\varepsilon)$-approximate maximum weight matching (MWM) in an $n$-node dynamic graph undergoing edge insertions and deletions. We provide a general reduction that reduces the problem on graphs with a weight range of $\mathrm{poly}(n)$ to $\mathrm{poly}(1/\varepsilon)$ at the cost of just an additive $\mathrm{poly}(1/\varepsilon)$ in update time. This improves upon the prior reduction of Gupta-Peng (FOCS 2013) which reduces the problem to a weight range of $\varepsilon^{-O(1/\varepsilon)}$ with a multiplicative cost of $O(\log n)$.
When combined with a reduction of Bernstein-Dudeja-Langley (STOC 2021) this yields a reduction from dynamic $(1-\varepsilon)$-approximate MWM in bipartite graphs with a weight range of $\mathrm{poly}(n)$ to dynamic $(1-\varepsilon)$-approximate maximum cardinality matching in bipartite graphs at the cost of a multiplicative $\mathrm{poly}(1/\varepsilon)$ in update time, thereby resolving an open problem in [GP'13; BDL'21]. Additionally, we show that our approach is amenable to MWM problems in streaming, shared-memory work-depth, and massively parallel computation models. We also apply our techniques to obtain an efficient dynamic algorithm for rounding weighted fractional matchings in general graphs. Underlying our framework is a new structural result about MWM that we call the "matching composition lemma" and new dynamic matching subroutines that may be of independent interest.
arXiv:2607.17598v1 Announce Type: new
Abstract: Long-document question answering usually forces a choice between loading the whole document into the context window and bolting on a separate retriever. Agentic AI suggests a broader option, giving the agent the document path and letting it decide how and what to read. Agent Skills, a standard for packaging expertise into folders an agent loads on demand, supply a ready mechanism: progressive disclosure, which exposes only what a query needs, from a short description down to the specific passages. Practitioners rapidly adopted this pattern for book-length understanding tasks, but the evidence to support such choices has been anecdotal. We run the first controlled study of the pattern, comparing raw-document navigation and several designs of Agent Skills packs against a classical hybrid retriever across three agent harnesses and three model families on InfiniteBench. On a single book, the gain depends on the harness, running large when the agent navigates the raw document poorly but near zero when a strong agent harness already divides and retrieves on its own. When scaling up to tasks that span many books, raw-document navigation collapses while one-level progressive disclosure degrades more slowly and pulls ahead. A second, deeper routing level never helps and sometimes breaks accuracy outright, so one level is enough. Progressive disclosure buys context, not intelligence: it is redundant while a strong agent can locate the right passages itself, and decisive once the corpus grows too large to navigate by reading.
arXiv:2412.03771v4 Announce Type: replace
Abstract: Zero-shot learning enables models to generalise to unseen classes using semantic information, bridging the gap between training classes and previously unseen test classes. While widely studied in computer vision, its application to environmental audio remains underexplored, and generative approaches have received little attention.
This work presents the first benchmark of generative methods for zero-shot environmental sound classification. Four approaches spanning variational, adversarial, diffusion-based, and denoising paradigms are evaluated. The benchmark includes CADA-VAE and LisGAN, adapted from computer vision, together with two embedding-generation methods introduced in this work: one based on a denoising diffusion probabilistic model (DDPM) and the other on a conditional generative denoising network (CGDN).
Experiments on five environmental audio datasets (ESC-50, ARCA23K-FSD, FSC22, UrbanSound8K, and TAU Urban Acoustic Scenes 2019) and one music dataset (GTZAN) show that generative methods are competitive with established compatibility-based approaches. Among the evaluated generative methods, CGDN achieves the highest average accuracy and is the only one to significantly outperform both the DDPM- and GAN-based methods, while remaining statistically indistinguishable from the strong ALE baseline. These findings suggest that optimisation stability is an important factor in generative zero-shot learning for environmental audio.
arXiv:2607.17140v1 Announce Type: new
Abstract: Unified multimodal models (UMMs) aim to integrate visual understanding and generation within a single architecture, but architectural unification alone does not ensure semantic consistency. A model may describe the intended target correctly while generating an inconsistent edit. This exposes an understanding-generation alignment gap: linguistic and visual outputs live in different spaces, yet should be governed by the same target semantics. We study this gap in image editing, where an instruction defines a target state that can be both described and visually realized. Given a source image and an edit instruction, we compare a UMM's target caption with its edited image to test whether the two outputs converge on the same result. Our analysis shows that existing UMMs remain weakly aligned, especially for fine-grained entities, attributes, spatial relations, and local details, indicating that semantic unification is not achieved by architecture alone. To bridge this gap, we propose STBridge, a shared-target alignment framework that connects understanding and generation through a common target state. Here the target caption expresses the desired visual result, while the edited image realizes it visually, replacing separate task-specific paths with a shared information flow from target expression to target realization. STBridge follows an align-then-optimize strategy: supervised fine-tuning first establishes the shared-target channel, and sequential reinforcement learning further refines target-centered coordination. Across visual understanding, image generation, and image editing benchmarks, STBridge consistently improves over the initialization model. Alignment analysis confirms that STBridge narrows the gap between what the model describes and what it generates, demonstrating shared-target alignment as an effective post-training strategy for bridging understanding and generation in UMMs.
arXiv:2607.17914v1 Announce Type: new
Abstract: Robust multi-agent coordination relies heavily on inter-agent communication, which is frequently disrupted by physical and environmental constraints in real-world deployments. To maintain operation during these intermittent communication failures, agents can employ internal prediction models to estimate missing shared state information. However, predictors trained with standard reconstruction objectives treat all transitions equally. In a Reinforcement Learning context, this forces the model to waste capacity learning stochastic exploration noise and the outdated dynamics of suboptimal policies. In this paper, we propose a value-aware extension of Multi-Agent Observation Sharing under Communication Dropout (MARO) to patch communication gaps; we refer to this method as Value-Aware MARO. By dynamically weighting the predictor's loss function using advantage estimates derived from the underlying actor-critic architecture, our objective explicitly couples the predictor's learning process to the policy's evolution. This formulation focuses the model's capacity on the intentional, high-return dynamics actively reinforced by the agents. We evaluate our framework on several tasks within the Multi-Agent Particle Environment under varying communication reliability levels. Experimental results demonstrate that our approach maintains performance under declining communication reliability, particularly below 40%. While our method performs comparably in tasks where the baseline already maintains high coordination, our value-aware weighting effectively prevents the performance collapse observed in the standard predictor during high-attrition scenarios. In these environments, our method achieves an average improvement in mean returns of more than 20% and reduces performance variance by a mean of 64.7% compared to the standard unweighted baseline.
arXiv:2607.16360v1 Announce Type: new
Abstract: Over-the-air computation (AirComp) exploits multiple-access superposition to compute functions of distributed data without separately decoding all terminal messages. We study a reusable two-input AirComp gate for the exp-minus-log (EML) operation $\eml(u,v)=\exp(u)-\log(v)$, $v>0$. Thus, all internal nodes of a prescribed real-admissible EML tree reuse one gate type, avoiding node-specific nonlinear gate designs. Given an explicit EML tree whose intermediate logarithm arguments remain positive on a given compact domain, we derive additive white Gaussian noise (AWGN) and coherent flat fading implementations under peak-power constraints. We then characterize the number of gate evaluations, the dependency depth, evaluation latency, node-wise feasibility, deterministic error propagation, positivity preservation, and a high-probability AWGN error bound for the complete tree. A four-terminal two-hop example gives explicit positivity and end-to-end error conditions, and a digital interface propagates quantization and gate errors across the tree.
Dependency-Guided Code Generation: Structured Matrix Decomposition and Consistency-Guided Refinement
arXiv:2607.16692v1 Announce Type: new
Abstract: The increasing complexity of modern software systems has made automated code generation a fundamental task in software engineering. However, existing approaches often fail to adequately capture the intricate, multi-level dependencies among code entities, leading to generated code that is logically incomplete or difficult to integrate into real-world systems. To address this limitation, we propose a dependency-aware code generation framework that explicitly models interactions among code entities through a graph-based representation. We decompose dependencies into two complementary components: a quantized matrix that captures strong, explicit relations, and a sparse low-rank factorization that models weaker, implicit interactions. The decomposition is efficiently learned via an alternating optimization procedure. During code generation, the learned dependency structure is incorporated as a constraint, ensuring both semantic coherence and structural consistency of the generated code. Furthermore, we introduce a sparse triplet representation for strong dependencies, significantly improving storage efficiency and computational scalability. Extensive experiments demonstrate that our approach consistently produces code with superior semantic alignment and structural fidelity compared to existing methods.
arXiv:2607.17742v1 Announce Type: new
Abstract: Tables are a critical knowledge source in retrieval-augmented generation (RAG), but a retrieved table may lack sufficient evidence to answer a query, a property we call answerability. While answerability broadly concerns whether a source or collection of sources contains sufficient evidence, retrieval models optimized for semantic relevance do not guarantee it even in the single-source case, creating a fundamental mismatch. To study this, we introduce TCR-Bench, a diagnostic benchmark for Table Content-level Answerability in RAG, built around sibling tables, i.e., tables with highly similar schemas but subtle content differences. On TCR-Bench, the dense retrievers we evaluate persistently exhibit a Semantic-Answerability Gap: they often retrieve the correct sibling group yet struggle to pinpoint the uniquely answerable table within it, dropping QA performance from 0.755 (oracle) to 0.330 (top-5 retrieved). Our analysis suggests this gap is associated with semantic accumulation, schema-level cue dependence, and weak row-column binding. As a diagnostic probe into the source of this gap, we test whether a lightweight two-stage pipeline, Answerability-Aware Reranking (AAR), applying direct query-table answerability judgment, can recover performance: it raises top-1 target retrieval from 18.2% to 57.4%, and this large gain is itself evidence that much of the observed failure reflects a missing answerability verification step, rather than an inherent limitation of model capacity alone.
arXiv:2607.16433v1 Announce Type: new
Abstract: Background: When we model networks, there is a problem of obtaining experimental data to verify other model approaches. And even if there are some experimental data, it is necessary to be sure of their reliability. Purpose: It is necessary to propose methods for obtaining reliable experimental data. Method: By its nature, network equipment is a software and hardware complex. Therefore, a full-scale software model can be considered completely equivalent to real equipment. And a real experiment can be replaced by a nature experiment. The reliability of a nature experiment will be based on its reproducibility. Results A comparison of popular nature network modeling packages was carried out. These packages were divided by functionality and feasibility of reproducible studies. Conclusions: Most software packages meet the reproducibility criteria. The choice of a specific solution depends on non-technical factors: popularity and knowledge of the package.
arXiv:2607.17360v1 Announce Type: new
Abstract: This paper presents a two-dimensional TMz finite-difference time-domain (FDTD) solver based on Yee's scheme for modeling radiation from an infinitely long z-directed line current, with the open region truncated by a Berenger split-field perfectly matched layer (PML). After validating cylindrical-wave propagation and negligible late-time reflections in free space, the solver is applied to three inhomogeneous configurations: (i) diffraction through a one-cell-thick perfectly electrically conducting (PEC) sheet with single and double slits; (ii) scattering from infinitely long PEC cylinders of circular and rectangular cross section; and (iii) scattering from infinitely long dielectric cylinders of varying cross section and permittivity. Beyond qualitative field maps, the diffraction case is characterized quantitatively: a steady-state phasor extracted by a running discrete Fourier transform yields the transmitted intensity, from which the fringe visibility and the far-field pattern are computed and compared against the closed-form Fraunhofer prediction. The single- and double-slit cases are cleanly separated by a visibility that rises from near zero to near unity, and the double-slit interference maxima agree with the grating condition arcsin(m \lambda_0 / d) to within a fraction of a degree. For dielectric cylinders, the field penetrates the obstacle with the expected reduced internal wavelength \lambda_0 / \sqrt{\epsilon_r}, and the scattered field strength grows with permittivity contrast. A reference-subtraction method isolates the scattered field throughout. The results confirm that the FDTD-PML framework accurately captures open-region diffraction and geometry- and material-dependent scattering.
arXiv:2607.16211v1 Announce Type: new
Abstract: LLM agents augmented with persistent memory can recall past interactions, but existing systems suffer from two limitations: flat, unstructured storage loses relational context needed for multi-hop and temporal reasoning, and reliance on expensive LLM-based classification makes them impractical for latency-sensitive deployment. Without mechanisms to validate new information against stored knowledge, these systems silently accumulate contradictions. We present MOSAIC (Memory-Organized Structured Agent for Information Collection), a structured, conflict-aware long-term memory framework for LLM agents that is substantially more accurate and efficient. MOSAIC introduces three key capabilities: (1) entity-typed graph storage with semantic classification preserving relational structure across events, personas, and relationships, enabling multi-hop and temporal reasoning over conversation history; (2) hash-accelerated dual-path retrieval replacing LLM-based classification with locality-sensitive hashing, achieving near-instantaneous lookup with negligible accuracy loss; and (3) active conflict detection at save time that cross-references new information against existing graph neighbors, triggering updates or deletions for contradictory entries. Evaluated on LoCoMo (long-conversation QA), HaluMem, and a novel clinical-guideline error compounding test, MOSAIC achieves 89.35% accuracy on LoCoMo (+27.21 pp over the best baseline), best HaluMem-Medium extraction F1(86.77%) and HaluMem-Long extraction F1 (85.84%), best QA correctness on both Medium and Long (73.10%, 70.75%), and detects 66% of injected factual conflicts-4.7 times higher than the best baseline (14%)-while hash-accelerated retrieval keeps average search latency at 0.58 s per question.
arXiv:2607.16297v1 Announce Type: new
Abstract: Edge intelligence systems, the intersection of edge computing and artificial intelligence (AI), are pushing the frontier of AI applications. However, the complexity of deep learning models and heterogeneity of edge devices make the design of edge intelligence systems a challenging task. Hardware-agnostic methods face some limitations when implementing edge systems. Thus, hardware-aware methods are attracting more attention recently. In this paper, we present our recent endeavors in hardware-aware design and optimization for edge intelligence. We delve into techniques such as model compression and neural architecture search to achieve efficient and effective system designs. We also discuss some challenges in hardware-aware paradigm.
arXiv:2607.17930v1 Announce Type: new
Abstract: We consider the combination of the two constrained planarity problems Level- and Clustered Planarity. Traditionally, level-planar drawings with convex clusters have been studied in this setting. Fink et al. (EuroCG 2024) recently introduced a different way of combining level- and clustered planarity by mimicking a classic characterization of clustered planarity in the level-planar setting: The problem (y-)monotone Clustered Level Planarity (mCLP) seeks a level-planar drawing in which it is possible to augment each cluster with edges that do not cross cluster boundaries so that it becomes connected while maintaining level-planarity. This is in line with previous research on clustered planarity that poses certain requirements on the augmentation edges that make each cluster connected, e.g., that they form a path. Fink et al. (EuroCG 2024) showed that mCLP is NP-complete even for biconnected single-source graphs and instances with a constant number of levels and clusters.
We further classify the parameterized complexity of the mCLP problem by, on the one hand, showing hardness even for instances that consist of a forest with trees of bounded size, no isolated vertices, and a small constant number of either clusters or levels. This excludes fixed-parameter tractability for almost all graph-structural parameters, except for vertex cover, even in conjunction with the number of clusters. We complement this by showing fixed-parameter tractability when parameterizing by the vertex cover number and the number of clusters. A major obstacle is the fact that mCLP is non-hereditary, i.e., subinstances of yes-instances may be no-instances and vice versa, which makes it challenging to apply usual reduction techniques.
arXiv:2607.17380v1 Announce Type: new
Abstract: Elasto-viscoplasticity provides a unified way of describing yield-stress fluids which may exhibit both solid-like and fluid-like behavior. In this work, we present a finite strain overstress-type elasto-viscoplastic framework designed to facilitate the incorporation of different yield surfaces. Within this framework, we compare several yield-surface choices and assess the associated challenges. We consider three representative yield surfaces: (i) pressure-independent, (ii) pressure-sensitive frictional and (iii) capped surfaces, corresponding to von Mises, Drucker--Prager, and modified Cam--clay models, respectively. In the case of von Mises, the proposed formulation naturally recovers the well-known Bingham and Herschel--Bulkley rheologies which are characterized by a single critical yield stress. We discuss in detail the singularity of the Drucker--Prager yield surface which requires a special treatment. In particular, we show that the modified Cam--clay model can be used to conveniently circumvent this singularity under the right conditions, retrieving the expected solution of Drucker--Prager. Implemented within a hybrid Eulerian--Lagrangian scheme, the general framework presented here enables efficient simulations of elasto-viscoplastic flows in two or three spatial dimensions, not requiring regularizing the solid-fluid transition nor a separate free-surface treatment. Numerical benchmark simulations illustrate how yield surface geometry affects velocity profiles, plug formation and compressibility.
arXiv:2607.16408v1 Announce Type: new
Abstract: Idle network service cores are treated as wasted compute. This assumption motivates increasingly sophisticated mechanisms that reclaim idle cores at microsecond timescales. We argue that this view no longer matches modern server hardware. On contemporary multicore processors, active cores compete for a shared package level power and thermal budget. Once that budget becomes the limiting resource, an idle core that waits efficiently returns compute capacity that hardware can redistribute to productive work. Measurements on a recent AMD EPYC processor show how waiting strategy, processor topology, and idle duration determine this tradeoff. Our results suggest that reclaiming idle cores often yields less benefit than commonly assumed while introducing substantial scheduling complexity. We propose a budget centric view of service core systems in which power, rather than core occupancy, becomes the fundamental resource and waiting policy becomes a first class systems design choice.
arXiv:2607.16409v1 Announce Type: new
Abstract: Unified Multimodal Large Language Models (MLLMs) offer a promising paradigm for unifying visual understanding and generation, yet they still struggle to follow complex spatial instructions and logical constraints in controllable image generation. To address this gap, we present ATLAS, a unified framework that equips MLLMs with a human-like "Think, Plan, and Paint" paradigm. We adopt layout as the shared representation that connects the three stages, enabling the model to reason about spatial requirements, plan explicit object arrangements, and render the final image. We further improve plan-to-image fidelity with reinforcement-learning-based layout alignment. We instantiate ATLAS at 7B and 80B scales, achieving state-of-the-art performance among MLLMs on image generation benchmarks and an average 65.31% improvement over existing layout-based unified MLLMs. On spatially related tasks, ATLAS obtains an average 23.06% gain over the base models. Through the same layout interface, ATLAS also supports instruction-guided editing and multimodal grounding. We further introduce ATLAS-Reasoning, a benchmark for evaluating generation under complex spatial instructions.
arXiv:2412.10733v3 Announce Type: replace
Abstract: We study the computational power that oblivious robots operating in the plane have under sequential schedulers. We show that this power is much stronger than the obvious capacity these schedulers offer of breaking symmetry, and thus to create a leader. In fact, we prove that under any sequential scheduler, robots are capable of solving problems that are unsolvable even with a leader under the fully synchronous scheduler FSYNC. More precisely, we consider the class of pattern formation problems, and focus on the most general problem in this class, Universal Pattern Formation (UPF), which requires the robots to form every pattern given in input, starting from any initial configuration (where some robots may occupy the same point, hence forming a multiplicity). We first show that UPF is unsolvable under FSYNC, even if the robots are endowed with additional strong capabilities (multiplicity detection, rigid movement, agreement on coordinate systems, presence of a unique leader). On the other hand, we prove that, except for point formation (Gathering), UPF is solvable under any sequential scheduler without any additional assumptions. We then turn our attention to the Gathering problem, and prove that weak multiplicity detection (the ability to detect a multiplicity but not the exact number of robots forming it) is necessary and sufficient for solvability under sequential schedulers. The results obtained show that the computational power of the robots under FSYNC (where Gathering is solvable without any multiplicity detection) and that under sequential schedulers are orthogonal.
arXiv:2412.11475v3 Announce Type: replace
Abstract: We present OmniVLM, a sub-billion-parameter vision-language model for efficient on-device inference. OmniVLM introduces a token compression mechanism that reduces visual token sequence length from 729 to 81 tokens, significantly reducing computational overhead while preserving visual-semantic fidelity. Through a multi-stage training pipeline of pretraining, supervised fine-tuning, and minimal-edit Direct Preference Optimization (DPO), OmniVLM matches the performance of larger models. On multiple benchmarks including ScienceQA, POPE, and MMMU, OmniVLM outperforms existing baselines like nanoLLAVA within a 968M-parameter footprint. Empirical results on the same laptop demonstrate 9.1x faster time-to-first-token (0.75s vs 6.82s) and 1.5x higher decoding speed (29.41 vs 19.20 tokens/s) compared to nanoLLAVA, enabling efficient deployment on edge devices. The model weights can be accessed on huggingface: https://huggingface.co/NexaAIDev/OmniVLM-968M, and the inference examples can be find in Appendix B.