arXiv:2607.06105v1 Announce Type: new Abstract: High-resolution RGB imagery acquired from low-altitude UAV surveys was processed through a modular pipeline incorporating transformer-based semantic segmentation, connected-component vegetation extraction, fine-grained species classification using a ConvNeXt architecture, and grid-based dominance scoring at 2x2m resolution. The framework targeted two ecologically significant halophytic grasses, Spartina maritima and Puccinellia maritima, and was trained using a curated and manually annotated UAV imagery, along with biodiversity imagery sourced from publicly accessible datasets. In order to identify these plants from the imagery, our segmentation yielded reliable species masks (mean IoU = 0.56; pixel-level accuracy = 0.96), while object-level classification achieved very good discrimination (F1 = 0.99). Dominance estimates closely matched quadrat-based field surveys, with mean absolute differences below 8%, preserving fine-scale spatial structure under realistic survey conditions. The developed system, named EcoVision, establishes a practical foundation for scalable, high-resolution salt marsh monitoring, demonstrating how AI-driven workflows can translate pixel-level predictions into ecologically interpretable metrics.
Science Journals
arXiv:2607.05771v1 Announce Type: new Abstract: The DMRadio-50L experiment is designed to search for axion dark matter in the 5 kHz - 5 MHz frequency range using a lumped-element LC resonator and a toroidal magnet and to serve as a testbed for quantum sensors. This paper describes the custom cryogenic system developed to meet the stringent requirements of the experiment within a standard laboratory environment. The system is designed to cool a 200 kg detector assembly to temperatures as low as 50 mK while providing sufficient cooling power at multiple temperature stages. We present the conceptual design, technical implementation, and measured performance of the hybrid cryogenic system, which combines a horizontal dilution refrigerator with a large vertical payload cryostat.
arXiv:2607.05790v1 Announce Type: new Abstract: Tool-augmented large language models extend their capabilities beyond parametric knowledge through external tools, but tend to invoke them unnecessarily. We investigate whether tool-use decisions have any stable internal representation that can be extracted and manipulated, a question that is non-trivial given that tools exist entirely in context at inference time and have no direct encoding in model weights. We show that steering vectors extracted from heading-anchors positions exert bidirectional causal control over tool-invocation behavior across five open-source models and three domains, suppressing unnecessary tool use most effectively in domains where parametric reasoning suffices. However, geometric analysis reveals that this causal effectiveness does not correspond to clean linear structure: tool-invocation steps exhibit diffuse, bimodal alignment with the suppression vector rather than the consistent negative alignment a linear encoding account would predict, and different tool types recruit largely distinct internal signatures with low cross-tool feature overlap. We hypothesize these geometric properties are indicative of the non-parametric nature of tools, and distinguish tool-use steering vectors from those extracted for parametrically grounded concepts. The relationship between this geometric irregularity and the observed causal effectiveness remains an open question.
arXiv:2607.06162v1 Announce Type: new Abstract: Image outpainting extends an image beyond its original borders, requiring seamless style integration and globally coherent scene completion. Building on the success of diffusion models, recent methods have achieved substantial improvements in visual quality. In practice, however, high-resolution outpainting is commonly performed via progressive expansion around a fixed source image, particularly in artwork scenarios. Despite this progress, existing approaches still suffer from three key limitations: (i) the absence of a reliable global planning mechanism, which leads to structural instability and error accumulation at high resolutions; (ii) limited spatial controllability beyond text prompts, making it difficult to place objects at user-specified locations; and (iii) high inference latency caused by inherently sequential patch generation. To address these issues, we propose a global blueprint-guided two-stage diffusion framework for layout-controllable high-resolution outpainting with efficient parallel synthesis. In Stage 1, we generate a low-resolution global blueprint using a layout adapter that injects bounding-box conditions into a Stable Diffusion inpainting backbone, producing a globally consistent structural plan while extracting global guidance features. In Stage 2, we synthesize high-resolution local patches in parallel by injecting the blueprint-derived global guidance and initializing each patch from the blueprint using the low-frequency preservation property of forward diffusion. This design eliminates sequential dependency while maintaining global coherence. Extensive experiments on large-scale artwork datasets demonstrate improved visual fidelity, stronger semantic consistency, and substantially reduced inference time compared to prior baselines, while uniquely supporting explicit layout control for artwork outpainting.
arXiv:2607.05408v1 Announce Type: new Abstract: The environmental impact of training large language models (LLMs) is increasingly scrutinised, yet most published estimates focus on operational energy and disclose little about manufacturing (embodied) emissions, water consumption, or the underlying highperformance computing (HPC) infrastructure. We present a life cycle assessment (LCA) of the pre-training of Lucie 7B, an open-source multilingual Foundation Model developed by the OpenLLM-France consortium and trained on the NVIDIA H100 partition of the Jean Zay supercomputer operated by IDRIS (CNRS). The assessment is framed by the AFNOR SPEC 2314 "Frugal AI" reference and applies the Labos 1point5 methodology for greenhouse gas(GHG) accounting in computing. The study scope extends from data preparation to model validation, and integrates the full life cycle of the hardware infrastructure: manufacturing (including raw-material extraction), use (compute, temporary storage, system administration, cooling), and end-of-life. We report (i) an annual footprint of 417.5 tCO2eq for the Jean Zay H100 partition, split almost equally between manufacturing and operation; (ii) an effective intensity of 36.7 gCO2eq per H100 GPU-hour; (iii) a total training footprint of 21 tCO2eq for Lucie 7B (574 564 H100 GPU-hours), inclusive of amortised hardware manufacturing; (iv) on-site water consumption of approximately 76m3 for the training campaign and an annual Water Usage Effectiveness (WUE) of 0.07 L/kWh for IDRIS; (v) a heat-reuse factor (ERF) of 0.37 thanks to waste-heat recovery into the urban heating network. The study contributes one of the few publicly documented LCAs of an LLM training campaign that explicitly couples operational data with embodied emissions decomposed by subsystem (compute, storage, power chain, cooling), and discusses the implications for the design of frugal-by-construction AI systems in Europe.
arXiv:2607.05453v1 Announce Type: new Abstract: Parkinson's disease (PD) affects millions worldwide and causes severe motor symptoms. Adaptive deep brain stimulation (aDBS) delivers physiologically informed stimulation that can track fluctuations in PD motor symptoms, enabling more intelligent DBS control. However, most existing aDBS approaches are primarily algorithm- and software-driven, with limited efforts toward circuit realization, particularly low-power and implantable integrated circuits. This paper presents the Silicon Leaky Integrate-and-Fire Deep Brain Stimulation (SiLIF-DBS) controller, a neuromorphic silicon neuron stimulator implemented with metal-oxide-semiconductor (CMOS) technology. For system-level evaluation, a simplified computational model of the SiLIF-DBS controller is derived and embedded within a Parkinsonian cortico-basal ganglia framework for closed-loop validation. The system is driven by beta-band subthalamic nucleus local field potentials (STN-LFPs), with their average rectified value (Beta ARV) used as the control biomarker. Our SiLIF-DBS controller for aDBS suppresses pathological beta activity while consuming only 25% of the power required by open-loop stimulation and achieving a suppression efficiency of $5.85\%$/$\mu$W. Overall, our SiLIF-DBS controller achieves strong beta suppression at substantially reduced power, delivering high suppression efficiency that demonstrates it is a viable foundation for low-power implantable aDBS.
arXiv:2607.05915v1 Announce Type: new Abstract: PCB routing is the task of connecting the nets of a board with copper traces under strict design rules, yet learning-based methods still lag behind rule-based routers. We introduce PCBWorld, an open-source engine-grounded PCB routing environment built on the KiCad EDA engine. As a human engineer does, agents in PCBWorld interactively route a board through the engine's native operations, using its Design Rule Check (DRC) feedback to keep the routing within the design rules. The environment supports both RL policies and tool-using LLM agents. Alongside the environment, PCBWorld-Bench provides three dataset families in KiCad's native board format (.kicad_pcb), covering two types of controllable synthetic instances and 679 real open-source boards. It scores any completed board with eight engine-checked evaluation metrics, regardless of the routing method. In our experiments, agents in PCBWorld consistently outperformed grid-action RL policies and open-loop LLM baselines, and an RL policy trained only on synthetic boards transferred zero-shot to real boards, approaching rule-based routers. These results position the engine-grounded, interactive approach of PCBWorld as a promising foundation for advancing the routing ability of both RL and LLM agents.
arXiv:2607.05933v1 Announce Type: new Abstract: Dynamic Voltage Frequency Scaling (DVFS) on resource-constrained embedded GPU platforms is essential for energy-efficient small language model (SLM) fine-tuning, as privacy- and personalization-driven adaptation increasingly requires local execution and involves repeated forward-backward optimization over many mini-batches, making it substantially more time- and energy-intensive than single-pass inference. To this end, 1) we first characterize the fine-tuning behavior of representative encoder-only SLMs of BERT variants, and autoregressive decoder-only SLMs of Pythia variants on GLUE benchmarks. In addition to the characterizations, 2) we propose a simple yet effective ML-based model selection that selects energy-optimal GPU DVFS settings on resource-constrained embedded platforms. Our results on NVIDIA Jetson AGX Orin demonstrate average 13.11% energy savings (up to 26.73%) over MAXN Mode 0, which has no explicit power cap.
arXiv:2603.04886v2 Announce Type: replace-cross Abstract: Internal--external field separation is crucial for many aspects of geomagnetism, aiming at distinguishing contributions of the magnetic field generated within a given observation surface from those generated in the exterior. When data are available on a full spherical observation surface, this separation is a standard, stable, and widely used procedure dating back to Gauss. However, when data are only available in a subdomain of the observation surface (as is the case for aeromagnetic and ground-based surveys), the situation drastically changes. Here we show that, without prior assumptions, an internal--external field separation is not uniquely possible. Given the geophysically reasonable assumption that the exterior sources, e.g., magnetospheric and ionospheric current systems, are located above a source-free spherical shell, we show that a unique separation becomes possible but that it is highly unstable. The results are based on the spherical Hardy--Hodge decomposition and explain the intrinsic difficulties of regional data-based internal--external potential field separation.
arXiv:2607.05569v1 Announce Type: new Abstract: Arm Permission Overlay Extension (POE) is an intra-process isolation mechanism based on memory protection keys. This mechanism partitions virtual memory into regions whose access permissions can be reconfigured without invoking higher exception levels. Because POE does not enforce privilege separation, POE-based protection domains are applicable only as a software fault-isolation technique. The POE2 is an isolation mechanism that substantially extends POE1. Dealing with memory protection keys, it also introduces new registers and tables that regulate the operations permitted for code executing within protection domains associated with different keys. A permission is determined by the spatial index of the executing code (i.e., the protection key associated with code), the spatial index of the accessed memory, the operation type, and a new attribute of the thread context: its temporal index. POE2 provides stronger security guarantees than POE1, but also introduces considerable architectural complexity. Effective permissions arise from the interaction of multiple hardware components, making direct programming non-trivial and prone to subtle errors. In the paper, we present a detailed analysis of POE2 and introduce a universal programming model with a strong security model for POE2-based systems. The model abstracts the complexity of spatial and temporal indices while enabling typical patterns of partitioned software constructed using intra-process isolation.
arXiv:2607.05573v1 Announce Type: new Abstract: Recent advances in Large Language Models (LLMs) and Vision-Language Models (VLMs) enable the automatic generation of parametric 3D designs from natural-language specifications. This chapter presents an empirical study of foundation models for automatic Computer-Aided Design (CAD) generation of mechanical parts, using a unified evaluation pipeline and a curated benchmark of 97 engineering design problems. We introduce LLMForge, a multi-model text-to-CAD framework integrating JSON-schema validation, analytic feature scoring, mesh synthesis, and multi-round iterative refinement, studied under two critique regimes. IterTracer uses a Phong-shaded ray-trace renderer with analytic visual metrics (silhouette IoU, hole visibility, edge clearance, aspect-ratio conformance) for lightweight geometry-aware feedback across rounds. IterVision replaces the analytic scorer with a VLM semantic critic (Qwen2.5-VL-72B) that evaluates rendered views via chain-of-thought visual reasoning, assessing spatial coherence and design intent. On a benchmark spanning four canonical geometry families (plates with holes and bolt circles, multi-feature boxes, flanged cylinders, and L-brackets), we evaluate seven foundation models: DeepSeek-V3.2, Qwen3-235B-A22B, Llama-3.3-70B, Gemma-3-27B, GLM-4.5, MiniMax-M2.1, and INTELLECT. Under IterTracer, the four highest-ranked models form a tight cluster (overall mean in [0.885, 0.890]) with 98.97% mesh success, showing that compact instruction-tuned models can match substantially larger systems. VLM-based critique in IterVision yields 100% watertight mesh generation on the leading model while surfacing systematic difficulty on rotationally symmetric geometries such as cylinders, where visual and semantic scoring diverge most. We discuss benchmark design, failure modes, CAD-oriented prompting, and implications for industrial workflows and scalable automated mechanical design.
arXiv:2607.06323v1 Announce Type: new Abstract: Real-world learning for dexterous hands remains brittle because high-dimensional hand actions amplify imitation errors and make reinforcement-learning exploration prone to contact-breaking motion. While combining imitation learning (IL) with online reinforcement learning (RL) can reduce manual supervision, unconstrained exploration in raw hand-action spaces is sample-inefficient and risky for physical hardware. We introduce a latent motion prior module (\prior{}) that maps recent hand-action histories to a compact, history-conditioned latent prior and decodes continuous latent commands into executable high-dimensional hand targets. Built on this prior, \method{} is a three-stage real-world dexterous learning framework: it pretrains \prior{} from demonstrations, trains a visuomotor policy that predicts native arm commands and latent hand-action offsets, and improves the policy with online residual RL in the same latent hand-action space. This shared, decodable interface lets residual exploration make local corrections near demonstrated, contact-consistent hand motions rather than perturbing every finger joint independently. We evaluate \method{} on four real-robot dexterous manipulation tasks against raw, linear, and discrete hand-action interfaces. Starting from small task-specific demonstration sets, \method{} achieves a 56.25\% average IL success rate and raises it to 98.75\% after online RL, reaching 100\% final success on three tasks and 95\% on the remaining task.
arXiv:2607.06011v1 Announce Type: new Abstract: This study proposes a metagraph-based domain-decomposed Galerkin reduced-order model (MBDD-G-ROM) for distributed-memory parallel reduced-order analysis of large-scale problems. The method represents domain-decomposed Galerkin reduced-order models over arbitrary domain decompositions using two graph levels: calculation-point graphs for interactions among discretization points and metagraphs for connectivity among local approximation-space subdomains. In the proper orthogonal decomposition (POD)-based implementation, POD computation subdomains are represented as metanodes, while metaedges encode the block-sparsity induced by overlaps between local POD basis supports. Partitioning the metagraph enables the POD computation subdomains to be decoupled from the parallel computation subdomains, allowing distributed-memory parallelization of both offline and online phases, including reduced-system assembly and iterative linear solution, without requiring the two decompositions to coincide. The metagraph also supports static load balancing through metanode weights that approximate computational costs. The method is evaluated for an unsteady diffusion equation and incompressible Navier-Stokes flow around a three-dimensional cylinder. The results show that MBDD-G-ROM preserves reduced-order solution accuracy while achieving high online parallel efficiency. A load-balancing test further demonstrates that cost-based metanode weights can improve computational efficiency.
arXiv:2604.24394v2 Announce Type: replace Abstract: The objective of Emergency Medical Services (EMSs) is to promptly respond to calls from citizens for first aid, providing pre-hospital care and, if necessary, to transfer patients to an appropriate Emergency Department (ED) by ambulance. The efficiency of such a system strongly depends on the deployment of ambulance home bases, i.e., locations where ambulances and their crews are strategically positioned, ready to respond to emergency calls. This paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behaviour and workflow of regional ambulance emergency systems. The proposed model incorporates and integrates information collected from different sources, reproducing very accurately the operation of the ambulance system, thus allowing a more comprehensive and realistic analysis. To show the applicability and reliability of the proposed general model, a case study provided by the Azienda Regionale Emergenza Sanitaria - ARES 118 (an Italian Regional Emergency Medical Services Authority - ARES~118}) is presented. It concerns a territory within the Lazio region of Italy, including a medium-size city along with sparsely populated areas. The reported results about scenario analyses highlight how the model we propose can be fruitfully used by the managers to improve effectiveness and quickness of the entire regional EMS system.
arXiv:2605.02965v2 Announce Type: replace Abstract: Artificial intelligence-generated content (AIGC) has emerged as a transformative paradigm for automating the creation of diverse and customized content, giving rise to rapidly growing computational workloads in cloud data centers. It is imperative for AIGC service providers (ASPs) to strategically schedule AIGC workloads to reduce data center energy costs while guaranteeing high-quality content generation. However, the distinctive characteristics of AIGC services pose critical challenges, including model heterogeneity across ASPs, implicit service quality evaluation, and complex inference process control. To tackle these challenges, we propose a joint energy management and coordinated AIGC workload scheduling framework, which introduces an explicit mathematical characterization of service quality to promote both job transfer among ASPs and fine-grained inference process configuration. Moreover, various energy resources within data centers are jointly considered to enhance power usage flexibility. Subsequently, a system utility maximization problem is formulated to balance AIGC service revenue with operational penalties and costs. Nevertheless, the strong coupling among job scheduling decisions induces severe reward sparsity, which limits the effectiveness of existing deep reinforcement learning (DRL) algorithms. To address this issue, we develop a diffusion model-aided reward shaping approach to synthesize complementary reward signals through a multi-step denoising process. This approach is seamlessly integrated with DRL to enable efficient learning of scheduling policies under sparse environmental feedback. Experiments based on real-world models and datasets demonstrate that our scheme effectively accommodates electricity price fluctuations and AIGC model heterogeneity, while achieving superior learning convergence and system utility compared with benchmark methods.
arXiv:2607.05590v1 Announce Type: cross Abstract: Implantable brain-computer interfaces require on-node spike sorting to reduce telemetry bandwidth and power while maintaining reliable neural decoding. This paper presents a hardware-oriented deep binarized neural network (DBNN) spike-sorting system with two binarized hidden layers with 256 neurons and a fixed-point output layer to enable multiplier-free inference dominated by sign-controlled accumulation and bit-wise logic. The proposed classifier operates on compact 16-sample spike waveforms to reduce the implementation cost (16-256-256-3) and achieves a median classification accuracy of 98.7% on both synthetic and in-vivo datasets. An FPGA prototype on a Cyclone V device operates at 50 MHz and requires 528 cycles per spike, corresponding to a 0.01 ms compute latency, while consuming 828 ALMs and 1023 registers with zero DSP blocks. For ASIC feasibility, the DBNN is implemented using FreePDK45-based flow; synthesis in Synopsys Design Compiler indicates an estimated silicon area of 0.014 mm2 and an operating power of 122 nW at 20 kHz under a 1.1 V supply. These results demonstrate that the proposed DBNN spike sorter offers a favorable trade-off between accuracy and implementation cost, supporting low-power, implantable neural interfaces. Overall, the proposed DBNN spike sorter achieves high accuracy (98.7%) with extremely low hardware cost (0.014 mm2, 122 nW at 20 kHz) and multiplier-free operation, making it suitable for low-power, implantable neural interfaces. This paper introduces the first DBNN designed for real-time neural spike sorting, striking an excellent balance between input data size and network complexity.
arXiv:2607.05664v1 Announce Type: cross Abstract: We consider the dynamics of inertial particles in unsteady fluid flows. At low Reynolds numbers, where the drag force is linear in the relative velocity, it is well-known that the dynamics admit an attracting, invariant, slow manifold which emerges as the perturbation of a normally hyperbolic critical manifold. However, at high Reynolds numbers, where the drag force is quadratic in the relative velocity, the critical manifold is no longer normally hyperbolic, and therefore its persistence has remained an open problem. Here, we resolve this issue by a particular application of the blowup method, which transforms the equations of motion to a generalized weighted cylindrical coordinate system, thereby desingularizing the dynamics on the critical manifold. We subsequently prove that the critical manifold persists under sufficiently small perturbations and derive the reduced equations of motion on the perturbed slow manifold to arbitrary accuracy. Our reduced equation differs from its linear-drag counterpart in its asymptotic expansion as well as its convergence rate. Using two examples, we demonstrate the validity of our slow manifold reduction. We also showcase an application of the reduced equations to the problem of source inversion in a turbulent dispersion model.
arXiv:2607.05691v1 Announce Type: new Abstract: Every chemical language model reading SMILES begins with a tokenizer, yet the field has inherited byte-pair encoding (BPE) from natural language with little scrutiny. In natural language, BPE's principal alternative, Unigram-LM, is known to build structurally different vocabularies. Whether that contrast survives in chemistry was open. We report a controlled comparison of BPE and Unigram-LM over a fixed 165-token chemistry base, at the small vocabulary sizes where token embeddings are learnable, across three corpus typologies (diverse, drug-like, natural-products) and both pre-tokenization boundary policies. The two do not converge. In all 22 matched conditions they build near-disjoint subword vocabularies: cross-algorithm Jaccard overlap on the learned pieces never exceeds 0.161, and at most 0.05 once weighted toward the high-frequency pieces a model updates most. Unigram-LM also segments held-out molecules into 29-41% more tokens; the arms largely agree on where to cut but not how deeply, so BPE's segmentation is a strict coarsening of Unigram-LM's on 80-99% of molecules. The separation holds across corpus, boundary, and vocabulary size, persisting even at eight times that scale. The subword algorithm is therefore a modeling decision, not a free default. The study trains no language models.
arXiv:2607.06435v1 Announce Type: new Abstract: Data from Singapore indicated that about 31% of the population had evidence of Helicobacter pylori infection. Persistent H. pylori infection is associated with chronic active gastritis and peptic ulcer disease, and its eradication is key to gastric cancer prevention. However, evidence supporting \textit{H. pylori} positivity and H. pylori-associated gastritis may be distributed across heterogeneous coded and free-text report fields and may require contextual interpretation of assertion and negation, limiting keyword search, and making manual review difficult to scale. We conducted a retrospective pilot evaluation of the Nimblemind Multi-Agent System (nMAS), a field-name-driven, evidence-linked extraction workflow, using 54 de-identified gastric biopsy pathology reports from a large healthcare system in Singapore. Four clinician-scoped binary fields were evaluated: gastric/stomach biopsy, biopsy status, H. pylori positivity, and H. pylori-associated gastritis. Across 216 feature-case decisions, nMAS correctly classified 213, corresponding to 98.61% overall accuracy. A separately implemented UMA-style MiniMax M2.5 comparator produced similar aggregate and per-field classification metrics. Although predictive performance was similar, nMAS maintained unified report-level outputs with supporting source sentences; the demonstrated contribution is therefore workflow integration and traceability rather than predictive superiority. Under an illustrative, unmeasured scenario, reviewing 1,000 reports at five minutes per manual review versus five seconds per evidence-linked verification would reduce review time from 83.3 to 1.4 staff-hours, corresponding to 81.9 staff-hours and about USD~6,100 in potential staff-time value. Larger multi-institutional studies should evaluate evidence-span correctness, clinician verification time, and generalizability.
arXiv:2607.06457v1 Announce Type: new Abstract: Audio Descriptions (ADs) narrate visual content for Blind and Low Vision (BLV) audiences during gaps in audiovisual media. There is growing momentum around ADs in movies and TV shows, and with mandates from India's Central Board of Film Certification (CBFC), there is a need to expand ADs beyond English. Yet, there is no work that generates ADs for any Indian language. To address this gap, we present the first systematic study of ADs in Hindi, contributing to aspects such as data, generation, and evaluation. We introduce Andha-Dhun, the first dataset of human-authored Hindi ADs collected from 8 full-length movies. We explore two approaches for generating ADs in Hindi: (i) directly from English dense video descriptions, and (ii) translating English ADs into Hindi. We evaluate these approaches using perplexity and LLM-as-a-judge metrics to assess fluency and quality respectively. We also analyze movies that have both English and Hindi human-authored ADs and find that naive translation introduces artifacts and narrows diversity compared to original Hindi ADs. Direct machine translation fails to adapt cultural references, while human-translated ADs do better but still fall short. Our findings emphasize that the purpose of Hindi ADs is accessibility for Indian BLV audiences, and that this requires adapting content for the audience more than strict fidelity to the source.
arXiv:2607.05836v1 Announce Type: cross Abstract: The limited-memory BFGS (L-BFGS) algorithm is a cornerstone of large-scale optimization due to its linear memory and computational costs. However, in ill-conditioned or non-convex landscapes, the implicit inverse Hessian approximation can suffer from an exploding condition number, leading to numerical instability and degraded convergence. To address this, we propose Two-Sided L-BFGS, a safeguarded variant that dynamically constrains the condition number of the inverse Hessian operator via a two-sided geometric envelope. Moreover, we show that Two-Sided L-BFGS preserves accumulated curvature information and maintains standard $O(mn)$ memory and per-iteration time complexities. We prove that this geometric envelope yields a uniform bound on the condition number of every inverse Hessian approximation generated by the algorithm. By tracking the algebraic evolution of the extreme eigenvalues through $m$ consecutive quasi-Newton updates starting from a scaled identity matrix, the resulting bound is expressed explicitly as a function of the memory depth, problem dimension, and envelope hyperparameters. Moreover, we show that Two-Sided L-BFGS preserves asymptotic global convergence in non-convex regimes under standard smoothness and strong Wolfe line-search assumptions, matching the theoretical guarantees of L-BFGS variants utilizing the Li-Fukushima cautious update rule. Numerical experiments on high-dimensional optimization problems demonstrate that the proposed method maintains well-conditioned inverse Hessian approximations and improves robustness and convergence behavior on ill-conditioned benchmarks.
arXiv:2602.14761v2 Announce Type: replace Abstract: Current meta-learning methods are constrained to narrow task distributions with fixed feature and label spaces, limiting applicability. Moreover, the current meta-learning literature uses key terms like "universal" and "general-purpose" inconsistently and lacks precise definitions, hindering comparability. We introduce a theoretical framework for meta-learning which formally defines practical universality and introduces a distinction between algorithm-explicit and algorithm-implicit learning, providing a principled vocabulary for reasoning about universal meta-learning methods. Guided by this framework, we present TAIL, a transformer-based algorithm-implicit meta-learner that functions across tasks with varying domains, modalities, and label configurations. TAIL features three innovations over prior transformer-based meta-learners: random projections for cross-modal feature encoding, random injection label embeddings that extrapolate to larger label spaces, and efficient inline query processing. TAIL achieves state-of-the-art performance on standard few-shot benchmarks while generalizing to unseen domains. Unlike other meta-learning methods, it also generalizes to unseen modalities, solving text classification tasks despite training exclusively on images, handles tasks with up to 20$\times$ more classes than seen during training, and provides orders-of-magnitude computational savings over prior transformer-based approaches.
arXiv:2607.06109v1 Announce Type: new Abstract: Multi-perturbation adversarial training (MAT) aims to achieve robustness against multiple $\ell_p$ perturbations but suffers from robustness trade-offs between different threats. To address this, we employ a mixture of experts (MoE) to route different threats through distinct model pathways. However, naive application of MoE encounters two critical challenges: experts tend to overlook threat-specific features and redundantly capture features shared across threats, and gating networks suffer from threat-agnostic routing where they learn nearly identical routing patterns across threats, thus preventing the construction of threat-specific model pathways. To this end, we propose Robust Mixture of Low-Rank Experts (RoME), where each expert is a low-rank additive update to the shared backbone, allowing it to capture threat-common features while experts focus on threat-specific information. To address threat-agnostic routing, RoME introduces (i) dual-scale gating that exploits threat-discriminative signals from local and global level features, and (ii) threat-guided gating diversification that enforces diverse expert utilization across threats. Extensive experiments demonstrate that RoME outperforms existing state-of-the-art MAT in union robustness and natural accuracy and improves robustness against unseen threats. Codes are available at https://github.com/wkim97/RoME.
arXiv:2607.05879v1 Announce Type: cross Abstract: We study topological relaxation in ordered viscoelastic conformation flows at finite epsilon. In an ordered region, a positive spectral gap selects an oriented principal axis and hence an S^2-valued director with a Hopf class. We show that a change of this class must be accompanied, before the sharp-interface limit, by one of a finite list of costs: exterior gap concentration, ordered-core mass, boundary flux, FENE/collar loss, or a topology exit. The result is proved for a concrete Landau-de Gennes ordered-core closure coupled to an Oldroyd/FENE-type transport law. The structural hypotheses used in the argument are verified up to the first typed exit time: the Morse-Bott ordered well, tubular soft coordinates, massive-mode coercivity, a projected transported Ginzburg-Landau equation, exterior gap control, and tame FENE/collar coefficients. The projected Ginzburg-Landau equation separates translation modes from the remaining residuals. The translation modes give the normal line force, while the orthogonal soft, massive, geometric, and collar terms are absorbed by coercivity or charged to the corresponding exit. A modulated-energy argument propagates a nonempty class of vortex-tube data on regular intervals. On each such interval, the normalized core measures converge to an integral one-varifold satisfying a transported forced Brakke inequality with the computed force. The theorem therefore derives the force projection, open-basin propagation, and Brakke compactness estimates before invoking any limiting Brakke flow, and it records the finite-epsilon cost when the regular ordered-core description breaks down.
arXiv:2607.06151v1 Announce Type: new Abstract: Generalization remains a pivotal challenge in deep learning, where traditional optimizers like Stochastic Gradient Descent (SGD) often converge to sharp minima, leading to overfitting and reduced performance on unseen data. Building on Sharpness-Aware Minimization (SAM), for seeking flat minima associated with improved generalization, we propose the Extragradient-Inspired Sharpness-Aware Minimization (EISAM), a novel optimizer that enhances generalization via the extragradient technique. EISAM uses a two-step update process: a prediction step investigating the geometry of the loss landscape and a perturbation step that refines updates with a base optimizer. This approach achieves better generalization performance than SAM. Crucially, EISAM reduces sensitivity to the perturbation radius, enhancing robustness, and simplifying the tuning across diverse settings. Extensive experiments on benchmark datasets demonstrate that EISAM consistently outperforms SGD, Adaptive Moment Estimation (Adam), and SAM in test accuracy and training efficiency across various architectures. Theoretical analysis further confirms that EISAM tightens the generalization bound by steering parameters toward flatter minima with reduced curvature. Accompanied by a thorough hyperparameter analysis, EISAM offers practical tuning guidance, establishing it as a robust, scalable, and broadly applicable optimization solution that advances both the theory and practice in deep learning.