arXiv:2604.19570v2 Announce Type: replace Abstract: Accurate medical image segmentation requires both long-range contextual reasoning and precise boundary delineation, a task where existing transformer- and diffusion-based paradigms are frequently bottlenecked by quadratic computational complexity and prohibitive inference latency. We propose RF-HiT, a Rectified Flow Hierarchical Transformer that integrates an Hourglass Transformer backbone with a multi-scale hierarchical encoder for anatomically guided feature conditioning. Unlike prior diffusion-based approaches that rely on hundreds of denoising steps, RF-HiT leverages rectified flow with efficient transformer blocks, achieving linear complexity and requiring only a few discretization steps. The model further fuses conditioning features at each resolution via learnable interpolation, enabling effective multi-scale feature integration with minimal computational overhead. As a result, RF-HiT achieves a strong efficiency-performance trade-off, requiring only 10.14 GFLOPs, 13.6M parameters, and inference in as few as 3 steps. Despite its compact design, RF-HiT attains 91.27% mean Dice on ACDC and 87.40% on BraTS 2021, achieving performance comparable to or exceeding that of significantly more intensive architectures. These results suggest that RF-HiT is a promising, computationally efficient foundation for clinical image segmentation.
Science Journals
arXiv:2605.00366v4 Announce Type: replace Abstract: High-capacity associative memories based on Kernel Logistic Regression (KLR) exhibit strong storage capabilities, but the dynamical and geometric mechanisms underlying their stability remain poorly understood. This paper investigates the global geometry of attractor basins and the mechanisms governing the storage limit in KLR-trained Hopfield networks. We combine empirical evaluations using random sequences and real-world image embeddings (CIFAR-10) with morphing experiments and statistical Signal-to-Noise Ratio (SNR) analysis. Our experiments show that the network achieves a storage capacity for random sequences up to $P/N \approx 16$, while maintaining stable retrieval for structured data at effective loads near $P/N \approx 20$. Morphing analysis indicates that attractors on the "Ridge of Optimization" are separated by sharp, phase-transition-like boundaries, characterized by steep effective potential barriers and critical slowing down. Furthermore, by comparing an SNR analysis with a geometric reference point inspired by Cover's theorem, we show that the practical storage limit is governed primarily not by a lack of geometric separability in the feature space, but by the loss of dynamical stability against crosstalk noise. These findings suggest that KLR networks function as highly localized exemplar-based memories that operate near the onset of dynamical collapse, providing a useful perspective on the design of robust, large-scale retrieval systems.
arXiv:2605.04035v3 Announce Type: replace Abstract: We propose HeadsUp, a scalable feed-forward method for reconstructing high-quality 3D Gaussian heads from large-scale multi-camera setups. Our method employs an efficient encoder-decoder architecture that compresses input views into a compact latent representation. This latent representation is then decoded into a set of UV-parameterized 3D Gaussians anchored to a neutral head template. This UV representation decouples the number of 3D Gaussians from the number and resolution of input images, enabling training with many high-resolution input views. We train and evaluate our model on an internal dataset with more than 10,000 subjects, which is an order of magnitude larger than existing multi-view human head datasets. HeadsUp achieves state-of-the-art reconstruction quality and generalizes to novel identities without test-time optimization. We extensively analyze the scaling behavior of our model across identities, views, and model capacity, revealing practical insights for quality-compute trade-offs. Finally, we highlight the strength of our latent space by showcasing two downstream applications: generating novel 3D identities and animating the 3D heads with expression blendshapes.
arXiv:2605.05453v2 Announce Type: replace Abstract: Silicon detectors typically require an insensitive area around their periphery to accommodate guard rings, which help maintain the electric field uniformity around edge pixels and isolate the high leakage current from the physical edges of the detector. Minimization of this insensitive region is desirable for applications in high-energy physics, X-ray experiments, and medical imaging. Existing active edge technology offers a solution for reduction or total elimination of the insensitive region, via a continuation of the highly doped backside up the sidewalls of the device. However, current methods for realizing this technology are complex and expensive. We propose a new technique that simplifies the fabrication of highly doped edges using side ion implantation and microwave annealing. Tests demonstrating the feasibility of this proposed process were performed on a set of sensors, and current versus bias voltage measurements probing the edge effects were performed before and after the edge implantation and annealing. To aid in interpretation of the results, TCAD simulations of the test devices were performed. Significant improvement in the edge leakage current is observed, indicating the promise of this simplified process for fabrication of active edge sensors.
arXiv:2605.09253v4 Announce Type: replace Abstract: While recent work in Reinforcement Learning with Verifiable Rewards (RLVR) has shown that a small subset of critical tokens disproportionately drives reasoning gains, an analogous token-level understanding of On-Policy Distillation (OPD) remains largely unexplored. In this work, we investigate high-loss tokens, a token type that--as the most direct signal of student-teacher mismatch under OPD's per-token KL objective--should progressively diminish as training converges according to existing studies; however, our empirical analysis shows otherwise. Even after OPD training reaches apparent saturation, a substantial subset of tokens continues to exhibit persistently high loss; these tokens, which we term Rock Tokens, can account for up to 18\% of the tokens in generated outputs. Our investigation reveals two startling paradoxes. First, despite their high occurrence frequency providing a disproportionately large share of total gradient norms, Rock Tokens themselves remain stagnant throughout training, resisting teacher-driven corrections. Second, through causal intervention, we find that these tokens provide negligible functional contribution to the model's actual reasoning performance. These findings suggest that a vast amount of optimization bandwidth is spent on structural and discourse residuals that the student model cannot or need not internalize. By deconstructing these dynamics, we demonstrate that strategically bypassing these ``stumbling blocks'' can significantly streamline the alignment process, challenging the necessity of uniform token weighting and offering a more efficient paradigm for large-scale model distillation.
arXiv:2605.12887v2 Announce Type: replace Abstract: Web-enabled LLM agents are changing how online information influences search outcomes. Existing Generative Engine Optimization (GEO) studies mainly focus on individual webpages. However, agentic web search is not a single-document setting: an agent may issue queries, crawl pages, follow links, reformulate searches, and synthesize evidence across multiple browsing steps. Influence therefore depends not only on page content, but also on how pages are organized, connected, and encountered along the agent's browsing trajectory. We study this shift through Ecosystem Generative Engine Optimization (EcoGEO), which treats GEO as an environment-level influence problem for web-enabled LLM agents. To instantiate this perspective, we propose TRACE, a Trajectory-Aware Coordinated Evidence Ecosystem. Given a recommendation query and a fictional target product, our method builds a controlled evidence environment that coordinates an agent-facing navigation entry page with heterogeneous support pages. These pages use shared terminology, internal links, and consistent product attributes to introduce, verify, and reinforce the target product. We evaluate our method on OPR-Bench, a benchmark for open-ended product recommendation. Experiments show that it consistently outperforms page-level GEO baselines in final target recommendation. Trajectory-level metrics further show increased initial target-result crawls, target-specific follow-up searches, and internal-link crawls, suggesting that the gains come from shaping the agent's evidence-acquisition process rather than merely adding more target-related content. Overall, our findings support an ecosystem research paradigm for GEO, where web-enabled LLM agents are studied in relation to the broader evidence environments that guide search, browsing, and answer synthesis.
arXiv:2607.00288v1 Announce Type: new Abstract: This article presents a communication-aware and risk-aware predictive latent control (CRPL) framework for unmanned aerial vehicle (UAV) systems operating under partial observability and uncertain environment dynamics. CRPL integrates a joint-embedding predictive architecture (JEPA) with probabilistic communication and safety constraints to jointly optimize UAV motion and transmission power. The learned latent model generates recursive multi-step rollouts, enabling the controller to anticipate future motion, channel degradation, and collision risk. These predictions are incorporated into a unified safety-aware optimization framework for proactive, energy-aware trajectory and communication adaptation. Simulation results show that CRPL closely approaches the performance of an oracle analytical predictive controller and outperforms reactive constrained and unconstrained baselines under limited bandwidth and dynamic uncertainty. In the bandwidth-limited regime, CRPL reduces terminal error, i.e., the final UAV-to-goal distance, by up to a factor of approximately $3$ and outage duration by up to approximately $18$, while also lowering communication energy and collision risk. These improvements are achieved with only a moderate motion-energy overhead, demonstrating a favorable trade-off among mobility effort, communication reliability, and operational safety.
arXiv:2607.00356v1 Announce Type: new Abstract: For multiple-input multiple-output (MIMO) channels with phase modulation, we recently proposed a method of unitarily transforming the channel matrix into a certain row-echelon form, by which the original MIMO channel can be converted into a certain number of scalar sub-channels with two phase inputs, thereby forming an annulus constellation geometry, and corrupted by both the additive white Gaussian noise and weak self-interference. In this paper, several bounds are derived to evaluate the fundamental limit of such a specific transceiver architecture. Two upper bounds are obtained by upper-bounding the capacity of a scalar channel with an annulus support constraint from the perspective of the convex geometry, while a lower bound is obtained by the standard entropy power inequality. Numerical results show that the gaps between these bounds are small at high signal-to-noise ratios for the MIMO phase-modulated channels over the Rayleigh fading and the single-input multiple-output symbiotic communication system assisted by a reconfigurable intelligent surface.
arXiv:2605.14939v2 Announce Type: replace Abstract: Reliable position and shape control in tokamak plasmas requires accurate real-time regulation of several strongly coupled shape parameters. The control vectors that disentangle these couplings, referred to as \textit{virtual circuits} (VCs), enable independent shape parameter control for a specific Grad--Shafranov (GS) equilibrium. Numerical calculation of VCs is not currently feasible in real time, therefore VCs are usually computed prior to each experiment, using a small number of reference GS equilibria sampled along the desired scenario trajectory, with each VC used to control the plasma within a preset time interval. While effective near the reference equilibrium, this approach can lead to degraded performance as the plasma departs from the reference equilibrium and/or from the desired trajectory, and it complicates the design of robust control strategies for rapidly evolving plasma configurations. In this paper, we construct neural-network-based emulators of plasma shape parameters from which VCs can be derived, to provide the MAST Upgrade (MAST-U) plasma control system with state-aware VCs in real-time. To do this, we develop an extensive library of over a million simulated GS equilibria, covering a substantial portion of the MAST-U operational space. These emulators provide differentiable functions whose gradients can be rapidly computed, enabling the derivation of accurate VCs for real-time shape control. We perform extensive verification of the emulated VCs by testing whether they disentangle the control problem. The neural-network-based approach delivers high accuracy and orthogonality across a diverse range of equilibria. This work establishes the physical validity of emulated VCs as a scalable and general alternative to schedules of precomputed VCs.
arXiv:2510.03976v2 Announce Type: replace Abstract: Equilateral triangular waveguides are one of the very few special kind of waveguides, whose field solutions can be constructed without necessarily solving the Maxwell's equations. Solutions can be obtained simply by superposing some plane wave solutions and requiring the solutions to obey the necessary boundary conditions for each modes, as would be expected of any Maxwell equation's solution or that of any Helmholtz equation resulting from the problem of Heat flow, membrane vibrations, etc. In fact, this was how one of the earlier solutions or eigen functions found by Gabriel Lam\'e, were obtained. Julian Schwinger, performed the same analysis, but used the superposition of complex exponential functions, in order to obtain the eigen functions of an equilateral triangular waveguide. The solutions exhibit other special symmetric properties that leave their solutions invariant and these special symmetries and applications, are what we investigate in this work, particularly as it relates to their attenuation characteristics and quality factors. Finally, by employing the number theory of Eisenstein integers or primes we have obtained a well ordered and more rigorous eigenvalues of the equilateral triangular waveguides, which is definitely something missing in the literature.
arXiv:2510.13537v2 Announce Type: replace Abstract: On-device deployment of Large Language Models (LLMs) frequently leverages Low-Rank Adapters (LoRAs) to support diverse downstream tasks under tight resource constraints. To address the limited storage capacity of mobile devices, recent works have explored model merging techniques to fuse multiple LoRAs into a single one. In practice, however, LoRAs are often delivered incrementally, as users request support for new tasks (e.g., novel problem types or languages). This scenario introduces a new challenge: on-device online continual merging, where the objective is to incorporate new LoRAs while preserving the performance on previously supported tasks. In this paper, we propose a data-free and computationally efficient strategy for selecting and merging LoRAs when a new one becomes available, assuming the device can store only a limited number of adapters. Extensive experiments across real-world tasks demonstrate the superiority of our approach compared to alternative strategies while adhering to the storage budget and compute limitations of on-device settings. The project page is available at: https://donaldssh.github.io/K-Merge.
arXiv:2605.15099v2 Announce Type: replace Abstract: Classifying interactions is key in the physical sciences, and bonding mechanisms in matter-antimatter systems remain particularly enigmatic. Here we focus on a paradigmatic example of positronium hydride (PsH) dimer composed of two protons, two positrons, and four electrons, whose bonding nature has been previously described as either ionic, covalent, or van der Waals-like. Accurate quantum Monte Carlo calculations show that the two positrons occupy a delocalized molecular orbital that envelopes the two hydrogen anions and responds as a collective dipole to an applied electric field. This positronic bonding stems from quantum correlations that resemble a single covalent bond formed between negatively charged pseudo-nuclei, but with a bond strength commensurate with the traditional van der Waals interaction. Our findings suggest that the ability to form delocalized proto-bonds is a more general property of quantum systems, and could be present in a broader class of particles, antiparticles, and quasi-particles interacting with matter.
arXiv:2605.16813v3 Announce Type: replace Abstract: The generation of production-ready quad-dominant meshes is a cornerstone of modern 3D content creation. Generating anisotropic quad-dominant meshes from point clouds is challenging, as existing methods are typically limited to producing either pure triangular meshes or pure quadrilateral meshes with isotropic densities. In this paper, we present QuadLink, a unified framework consisting of three stages for quad-dominant mesh generation by linking points into structured faces. QuadLink formulates polygonal mesh generation as a hybrid centroid-conditioned vertex linking model: it first predicts a unified set of anchors (vertices and face centroids), then learns centroid-conditioned links that associate vertices with face centroids, and finally assembles polygonal faces with a quad-first strategy guided by robust geometric verification strategies. This link-based formulation enables efficient generation of sparse and anisotropic quad-dominant meshes with coherent edge flow and meanwhile supporting hybrid polygonal topology. To construct training data for this model, we further introduce a Tri-to-Quad Operator that converts artistic triangle meshes into quad-dominant training data via global merge selection. Extensive experiments show that QuadLink produces production-ready quad-dominant meshes from point clouds and achieves improved geometric fidelity and topological quality compared to prior baselines. Our method natively supports hybrid polygonal topology, generalizing to arbitrary n-gon meshes without architectural changes.
arXiv:2605.20855v2 Announce Type: replace Abstract: Photoinduced force microscopy (PiFM) provides a force-based probe of nanoscale polarization-dependent molecular excitations. Here, we theoretically investigate PiFM images of zinc phthalocyanine assemblies using the discrete dipole approximation with nonlocal molecular susceptibilities. Under linearly polarized illumination, intermolecular dipole coupling splits the molecular resonance into bonding and antibonding modes. These modes produce distinct force distributions: bonding modes enhance signals at molecular termini, whereas antibonding modes localize responses in intermolecular regions, reflecting collective intermolecular polarization modes. Under circularly polarized illumination, intermolecular coupling and anisotropic molecular packing generate spatially varying local circular dichroism signals with enhanced asymmetric force factors. These results establish PiFM as a real-space probe of collective polarization modes and coupling-induced local chiral optical responses.
arXiv:2605.22584v2 Announce Type: replace Abstract: Arguably the most widely used approaches for obtaining highly accurate molecular ground-state energies are coupled cluster methods. Despite introducing two layers of approximation, a linear and a nonlinear one, coupled cluster methods remain computationally intensive, with the complexity scaling as $O(poly(N))$, where $N$ is the number of electrons. Moreover, this method must be applied over a large set of different nuclear coordinates in order to study certain chemical phenomena. Therefore, in this work, we investigate the regularity of single-reference coupled cluster amplitudes with respect to nuclear coordinate displacements, with the aim of enabling interpolation or extrapolation approaches that rely on only a limited number of reference geometries. We show that, in theory, under certain non-degeneracy assumptions on the Hartree-Fock level of theory, and the coupled cluster level of theory the amplitudes behave real analytic. Furthermore, we analyze the artifacts that arise in practical calculations that use canonical orbitals, which hinder this high degree of regularity, and suggest strategies to mitigate these issues. Finally, we validate our findings through numerical experiments by interpolating the amplitudes and comparing the performance of the interpolants with that of the exact amplitudes.
arXiv:2605.24661v3 Announce Type: replace Abstract: Despite remarkable progress on reasoning benchmarks, current LLM evaluation practice remains anchored to final-answer correctness, providing limited insight into how models reason, how reliably they behave under contextual variation, or how efficiently they reach conclusions. This paper proposes a unified multi-dimensional framework for measuring LLM reasoning quality from a behavioral perspective, operationalizing six theoretically grounded dimensions rooted in cognitive science: Correctness (CQ), Consistency (CS), Robustness (RS), Local Logical Coherence (LS), Efficiency (ES), and Stability (SS). The framework introduces deployment-aware aggregation, enabling context-specific model selection beyond accuracy-based leaderboards. Experiments across multiple LLMs and benchmarks reveal behaviors systematically concealed by single-metric evaluation, including the orthogonality of local logical coherence and correctness, deployment-context-dependent ranking inversions, and non-trivial dimensional profiles in small locally-deployed models. Discriminant validity analysis confirms that the proposed dimensions capture largely non-redundant signals. The resulting pipeline provides a foundation for diagnosing LLM reasoning behavior across deployment contexts, with domain-specific validation as a direction for future work.
arXiv:2605.25237v3 Announce Type: replace Abstract: The shortest path multiplicity, here denoted by $\mu$, is an important metric of complex networks. For real networks $\mu$ is high and it correlates with the network community structure. Since local network evolution induces network communities, it is possible that a high shortest path multiplicity is the natural expectation of local evolution rules. Here I demonstrate, by means of numerical simulations, that this is indeed the case. For random graphs with arbitrary degree distributions $p_k$, $\langle\mu\rangle\sim \langle k(k-1)\rangle / (\langle k\rangle e)$, growing with the network size when $p_k\sim k^{-\gamma}$ and $\gamma\leq3$. For networks generated by local rules, $\langle\mu\rangle$ increases with the network size and it does so faster than what is observed in their randomized versions. Furthermore, the number of communities increases with the network size and the correlation with $\langle \mu\rangle$ follows.
arXiv:2605.26542v4 Announce Type: replace Abstract: Tool-using agents increasingly operate in open-ended deployment environments, where they compose file systems, web APIs, code interpreters, and enterprise services at runtime. This creates a safety gap in tool composition: an agent can satisfy every per-tool permission check and still produce an unsafe end-to-end effect, such as reading a confidential document, summarizing it, and sending the summary to an external endpoint. We call this failure mode permission laundering. ChainCaps addresses it with a runtime rule: every value carries a sink-specific capability budget, and tool composition propagates budgets by intersection. A value can preserve or lose authority as it moves through a tool chain, but it cannot gain new authority through composition. We implement ChainCaps as a transparent MCP proxy that requires no changes to the agent or tool servers. On 82 tasks across five frontier models from three providers, ChainCaps reduces attack success rate from 25-68% to 0-4.8% while preserving 96-100% benign completion. In replay experiments, it also outperforms scalar-IFC and per-function-isolation baselines. Manifest quality is the dominant deployment bottleneck: expert manifests reach 100% attack blocking, while naive manifests fall to 27.3%. Our claims are limited to explicit-flow composition safety under trusted manifests and proxy-visible data movement, a practical gap in deployed tool-using agents today.
arXiv:2605.28380v2 Announce Type: replace Abstract: In this paper, we develop an efficient preconditioned unfitted finite element method for the elliptic interface problem, based on the reconstructed discontinuous approximation. The key idea is to impose suitable constraints on the local least squares reconstruction. These constraints ensure the stability near cut interface elements and, more importantly, establish a norm equivalence between the high-order space and the lowest-order piecewise constant space. This result allows us to construct an optimal preconditioner directly from the lowest-order system on the same unfitted mesh for any high-order scheme. The resulting method combines a cut discontinuous Galerkin formulation with Nitsche's penalty technique. The approximation space achieves arbitrarily high order accuracy with only one degree of freedom per element. We prove optimal error estimates and show that the condition number of the preconditioned system is uniformly bounded independently of the mesh size, coefficient contrast, and the location of the interface relative to the mesh. Multigrid algorithms are further designed to efficiently approximate the inverse of the lowest-order system matrix. Numerical experiments in two and three dimensions confirm the optimal convergence rates and demonstrate the robustness and efficiency of the proposed preconditioning method.
arXiv:2607.00171v1 Announce Type: new Abstract: Text embeddings are standard for semantic similarity tasks, yet their evaluation remains an open challenge. Current benchmarks are static, cover only a limited set of languages, are often domain-specific, susceptible to overfitting, and poorly representative of low-resource languages. To address these limitations, we introduce ALEE, a framework that extends Sentence Smith (Li et al., 2025) to the cross-lingual and paragraph level. ALEE uses Abstract Meaning Representations (AMR) to generate English minimal pairs with controlled, fine-grained semantic shifts, which are paired with translations in target languages. This approach enables targeted diagnostics for models in any language with English parallel data. We conduct a large-scale empirical study across a diverse set of embedding models and 275+ languages spanning three parallel datasets. On ALEE, performance varies substantially across languages, text lengths, and linguistic phenomena, exposing persistent gaps in cross-lingual semantic representation that track language prevalence in training resources and subword tokenization. We release ALEE at https://github.com/Andrian0s/any-lang-embed-eval
arXiv:2602.04724v2 Announce Type: replace Abstract: High-reflective crystalline $GaAs/Al_{0.92}Ga_{0.08}As$ coatings show reduced Brownian noise compared to conventional dielectric coatings. However, several ultra stable laser systems observed additional noise sources that hinder the realization of the expected improvements in frequency stability. These additional noise sources are related to the birefringence of the coatings and its modification by intracavity light. The origin of the birefringence is not yet well understood and its modification via illumination remains unexplained. Here we present an extensive study on the steady-state and transient modification of the birefringence by intracavity light and by uniform illumination at various wavelengths using an optical cavity at room temperature. We find a unified description that suggests a primary two-photon process for photon energies below the bandgap of GaAs, or a single-photon process at higher energies. Adding external illumination allows to reduce noise induced by laser power fluctuations by balancing the photo-thermal-optic response of the mirrors and the photo-birefringent effect at a more favorable low intracavity power.
arXiv:2605.28418v3 Announce Type: replace Abstract: With the rise of tabular foundation models alongside traditional models still performing well on many tasks, choosing the right model for a tabular dataset remains difficult. We investigate whether dataset meta-features can explain performance gaps between model families on tabular prediction tasks. Using the TabArena benchmark results, we analyze dataset-level performance gaps and relate them to model-agnostic meta-features. After strict statistical tests with false discovery control, we find that (1) for neural network vs. tree gaps, no meta-feature survives false discovery control, (2) for non-foundation vs. foundation model gaps, one association is robust but does not generalize when tested in leave-one-dataset-out prediction, and (3) for TabICLv2 vs. TabPFN-2.6, one robust association also improves held-out prediction. Furthermore, we conduct a leave-one-dataset-out analysis and find that meta-feature predictors fail to improve meaningfully over a simple baseline. Overall, our results show the heterogeneity of tabular datasets and that global meta-feature approaches are not robust enough to offer explanations on the 51 TabArena datasets.
arXiv:2605.28673v2 Announce Type: replace Abstract: We have developed a methodology for analyzing multifrequency oscillating magnetic fields using neutron spin interferometry. A theoretical formulation was derived for the contrast and the phase of the interference pattern for an input oscillating magnetic field consisting of a series of sine functions with multiple frequencies. According to the formulation, the contrast decays because of the dispersion of the Larmor precession induced by an oscillating field. The formulation also predicts that the phase of the interference pattern is not constant in contrast to single-frequency cases. We conducted an experiment on the JRR-3 C3-1-2-2 beamline to confirm the formulation in the cases of double-frequency oscillating fields in the frequency range from 2500 to 10000~Hz. The measured interference patterns showed reasonable agreement with the theoretical predictions.
arXiv:2607.00382v1 Announce Type: new Abstract: We propose the first compression approach for image-to-shape Diffusion Transformers (DiTs) that substantially reduces model size while preserving geometric fidelity. Despite remarkable progress in 3D shape generation, large DiT-based models remain computationally prohibitive in resource-constrained settings. Furthermore, it is difficult to directly transfer existing diffusion model compression strategies developed for different domains to 3D generation, and prior 3D efficiency approaches focus primarily on inference speed rather than backbone compression. To address this limitation, we build a geometry-aware compression framework tailored to image-to-shape DiTs. Guided by the observation that 3D DiT layers exhibit non-uniform importance for geometry synthesis, we introduce a vitality-guided framework integrating structured pruning, adaptive quantization, and targeted fine-tuning. Our method achieves up to 66% model-size reduction across state-of-the-art image-to-3D models while maintaining synthesis fidelity comparable to full-sized counterparts. This highlights the potential of our framework as a plug-and-play solution for efficient 3D shape generation across diverse models.
arXiv:2607.00389v1 Announce Type: new Abstract: We study longest-common-extension (LCE) queries and lexicographic minimizer maintenance on the suffix tree of a sliding window. The main difficulty is that a sliding suffix tree is maintained in an implicit Ukkonen-style form: some suffixes of the current window are not represented by leaves. We show that the longest implicit (i.e. non-leaf) suffix induces a periodic representative map that folds every implicit suffix to an explicit suffix leaf in constant time. Combined with leaf pointers [Leonard et al., PSC 2026] and a dynamic LCA data structure [Cole & Hariharan, SICOMP 2005], this yields a linear-space data structure with amortized constant-time window shifts and worst-case constant-time LCE queries over a constant-size alphabet. For minimizers, the LCE structure gives a direct exact solution, but it uses more machinery than fixed-depth comparisons require. We therefore give an alternative LCE-free algorithm that reports minimizers in constant time per window shift, which is built on BP-linked suffix trees [Sumiyoshi et al, SPIRE 2024] and a standard order maintenance data structure (e.g. [Bender et al., ESA 2002]).