arXiv subjects

Ye Lu

Publications and source records attributed to Ye Lu.

At least 19 recordsLinked to original sources

arXiv2026

Efficient Quantum Optimization via Dissipative False Vacuum Decay

Efficient quantum solvers for computationally hard problems are highly desirable for near-term quantum platforms. Here, we propose a cavity-based dissipative quantum solver for a class of NP-hard combinatorial optimization problems encoded in atomic-spin Hamiltonians with a few discrete single-spin-flip transition energies. We show that the cavity-based quantum solver exploits dissipative false-vacuum decay in configuration space to facilitate escape from metastable local minima in rugged energy landscapes. Numerical studies of MaxCut and Maximum Independent Set identify favorable parameter regions for ground-state preparation. Based on the finite-size scalings of the median mixing time and instance-averaged ground-state preparation probability, we estimate time complexities for our solver that are more favorable than those of standard Grover search. These results highlight cavity dissipation as a controllable resource for quantum optimization.

quant-ph↗

arXiv2026

From Speech Tokens to Stolen Voices: Attacker-Chosen-Text Speaker Impersonation via Representation Inversion

Speech interaction systems have gradually evolved from traditional cascaded systems that use text as intermediate information to end-to-end speech language models that directly process speech representations. In some applications, the speech frontend runs on the user's device and sends speech representations to a remote server for inference, while the original audio remains local. However, these representations may still contain speaker identity information. If an attacker obtains them, it may use this information to imitate the user's voice, threatening their voice privacy. To investigate this risk, we propose \stia, a Speech Token Inversion Attack that generates speech with attacker-specified text in a victim's voice using a single short leaked representation, without access to the original audio or its transcript. Specifically, \stia converts discrete, continuous, or mixed speech representations into conditioning sequences and trains a conditional flow-matching speech generation model using pairs of different utterances from the same speaker in public speech datasets. The model combines speaker characteristics contained in the leaked representation with the specified text to generate speech in the victim's voice. Experiments show that, with a Moshi representation leak corresponding to 3 seconds of speech, \stia achieves an average CAM++ speaker cosine similarity of 0.78 and a target-text word error rate of 3.80\%. These results show that speech frontend representations reveal reusable speaker identity information and highlight the need to consider privacy risks when designing speech frontend representation methods.

cs.SD↗

arXiv2026

Steering the Flow: Inverting Face Recognition Models via Gradient-Guided Flow Matching

Model Inversion Attacks (MIAs) aim to reconstruct representative training samples of target identities from face recognition models, exposing critical security vulnerabilities. Existing methods typically rely on indirect guidance or highly stochastic guidance, making it difficult to stably optimize generation trajectories toward target facial images. In this paper, we propose Steering Flow Model Inversion (SFMI), a novel two-stage white-box model inversion method that reformulates inversion as a trajectory-steering task. Specifically, Step I, Learning a Generic Flow Matching Prior, pre-trains a generic unconditional Flow Matching model to encode the manifold of human faces as a robust prior. Step II, Attacking with Progressive Guidance Scheduler (PGS), injects time-dependent target-specific gradients during sampling. By backpropagating through the target model to obtain gradients from intermediate generated states, PGS progressively injects adaptive guidance signals into the vector field. This process effectively steers the current generative flow from random noise toward the high-density regions of the target class. Under an identity-disjoint cross-evaluation setting using the CelebA dataset, SFMI achieves an ACC of 0.9248, an FID of 22.61, and an LPIPS of 0.3874 on the ArcFace target. Extensive experiments on multiple target models demonstrate that SFMI achieves competitive state-of-the-art performance in attack success and visual fidelity under the evaluated white-box protocol.

cs.CV↗

arXiv2026

Reference-Efficient Multimodal Bitstream-Corrupted Video Recovery via Depth-Conditioned Adaptation

Recovering visually faithful content from damaged video bitstreams is important for reliable multimedia communication and storage. Existing methods exploit local residuals and non-local RGB references of the input video sequence, but both sources can become unreliable after decoding, while processing additional references increases receiver computation and memory. This paper studies multimodal bitstream-corrupted video recovery (MBSCVR) using RGB and depth information, where pre-transmission depth-based geometry aids reconstruction when the received appearance is corrupted. We propose a novel MBSCVR framework that selects and effectively utilizes a single selected non-local RGB reference per recovery window to achieve higher performance. It integrates a reliability-aware multimodal selector that combines valid appearance with depth compatibility before reference encoding. Depth-Conditioned Reference Adaptation (DCRA) introduces selected RGB evidence into feature completion and content generation, while frame-aligned depth conditions completion queries. On the complete BSCV benchmark with oracle masks, MBSCVR uses one selected reference to achieve 32.30 dB on YouTube-VOS and 28.93 dB on DAVIS, surpassing full-reference B2SCVR by 0.27 and 0.33 dB and improving SSIM on both subsets. Matched receiver-forward measurements show 36.8\% lower latency and 40.8\% lower peak memory. These results demonstrate how compact geometric guidance and adapted appearance processing improve the quality--resource trade-off of multimedia recovery.

cs.CV↗

arXiv2026

Decision Noise in Threshold Selection

Can noisier threshold selection produce a group with higher mean true value? We answer this question by varying the scale of a fixed assessment error. Holding the selection rate fixed, less noise yields a first-order stochastically dominant distribution of selected values. At a fixed threshold, however, more noise can raise the selected mean even as selection loss, which weights selection mistakes by the distance of true value from the threshold, increases. Under joint Gaussianity, we characterize when the selected mean rises or falls with noise. In treatment selection, noise alone can generate subgroup differences in participants' average gains despite identical population gain distributions.

econ.TH↗

arXiv2026

EduAlign: Aligning Educational LLMs for Helpfulness, Personalization, and Creativity

Educational language models should pursue three complementary objectives: helpfulness through responsible guidance, personalization to learner needs, and creativity that supports exploration. General-purpose models do not consistently satisfy these educational requirements, motivating explicit alignment with all three objectives. We present EduAlign, a framework that combines educational reward modeling with multi-objective reinforcement learning. To obtain training signals for these goals across educational scenarios, we develop a rubric-guided data synthesis pipeline that iteratively refines scoring criteria and generation prompts for each subject, grade band, and task type. The resulting supervision trains a multidimensional reward model, Edu-RM, to predict helpfulness, personalization, and creativity scores. To jointly optimize these educational objectives, we introduce Ideal-point Policy Optimization (IdealPO), a multi-objective reinforcement learning algorithm. IdealPO combines ideal-point marginal credit with continuous conflict weighting to guide policy updates that account for uneven progress and conflicting contributions across objectives. Experiments with 7B and 32B models show that EduAlign achieves the highest mean ratings on all three dimensions from both education experts and a model judge, alongside the highest aggregate scores on EduBench, ELMES, and EduValues among the compared methods. Reward-model validation and component ablations further support the effectiveness of the framework.

cs.LG↗

arXiv2026

Interpolating Neural Operator (INO): A Data-Free and Efficient Approach for Learning PDE Solution Operators

Neural operators have become a popular approach to approximate the solution operators of parametric partial differential equations (PDEs). However, existing neural operators either require a large amount of simulation data or a long physics-informed training on GPUs, and they cannot tell how accurate an individual prediction is. In this paper, we propose the Interpolating Neural Operator (INO), a data-free interpolating neural network that is trained directly on the weak form of the PDE. In INO, the Karhunen-Loève coordinates of the input field are treated as additional inputs together with the spatial coordinates, and each input is approximated by a C-HiDeNN sub-network whose trainable parameters are nodal values. Since the network is multilinear in its parameters, training reduces to a sequence of one-dimensional linear solves by greedy alternating least squares. As a result, INO trains on one CPU core and predicts a new solution in microseconds. For coercive problems, the total error of every prediction is bounded by a computable residual bound that requires no reference solution, and the same bound applies to the predictions of other methods that satisfy the boundary conditions exactly. Before each prediction, INO checks whether the leading coordinates of the input lie within the range on which it is trained, and inputs outside this range can be passed to a conventional solver or to an INO trained on a wider range. INO is compared with physics-informed FNO and DeepONet on different benchmarks. INO is the most accurate model on most of these problems, by 15$\times$ on two-dimensional Helmholtz at $65^2$ and 53$\times$ on the diffusion-reaction benchmark, and on the one- and two-dimensional problems its training on one CPU core takes 3-80$\times$ less time than the physics-informed baselines on one GPU.

cs.CE↗

arXiv2026

Rare States and Long-Run Pricing

Macroeconomic crises are rare, yet they can have large effects on asset prices. Because the probability of entering a crisis from normal times is small relative to the probability of recovery, the dynamics are nearly reducible: setting the crisis-entry probability to zero makes crisis states unreachable from normal states. We develop a graph-theoretic and spectral analysis of long-run pricing under reducible and nearly reducible dynamics. We show that, at the reducible limit, physical and pricing dynamics induce the same classes of economic states and accessibility relations, but rank their long-run importance differently: by recurrence and by class-specific pricing rates. Moreover, a claim's price depends only on its pricing corridor---the classes of states lying on directed paths from the initial state to states where the payoff is positive---and the highest class-specific pricing rate within this corridor determines the claim's long-run pricing rate. When the reducible-limit corridor excludes the globally pricing-dominant class, we characterize how restoring rare transitions introduces a contribution from that class with a weight that vanishes as the transitions become rarer. This contribution becomes dominant only beyond a crossover maturity, which increases with the rarity of the connecting paths and decreases with the dominant class's pricing-rate advantage. Finally, a consumption-based disaster-recovery application shows how preferences, consumption risk, cash-flow exposure, and recovery timing jointly shape the long-run pricing dominance of disaster states and crossover maturities.

econ.TH↗

arXiv2026

A4A: Cross-Embodiment Transfer of Action-Oriented 4D Affordances from Human Demonstrations

Human demonstrations contain rich manipulation knowledge, but it remains unclear what information can be transferred effectively to robot control. Existing affordance representations are typically formulated as 2D masks, 3D regions, contact points, or actionability scores, and therefore primarily identify where interaction may occur. However, effective manipulation also requires modeling how interaction-relevant geometry evolves during task execution. To bridge this gap, we introduce action-oriented 4D affordances, which represent the language-conditioned future trajectories of interaction-relevant 3D points. These trajectories capture task-conditioned geometric evolution rather than embodiment-specific actions, enabling transferable interaction priors across humans and robots. Based on this representation, we construct a large-scale action-oriented 4D affordance dataset from existing human--object interaction video data and complementary RGB-D demonstrations, and introduce A4A, an affordance-to-action framework that uses 4D affordance trajectory prediction to pretrain robot policies before manipulation fine-tuning. Experiments in both simulation and the real world validate the effectiveness of A4A, showing that pretraining with action-oriented 4D affordance data consistently improves the manipulation performance of diverse VLA policies. These results establish action-oriented 4D affordances as an effective cross-embodiment representation for transferring manipulation knowledge from human demonstrations to robot control.

cs.RO↗

arXiv2026

Phase-field digital image correlation for integrated displacement and damage measurements

This work presents a novel digital image correlation (DIC) framework for full-field measurements of displacement, strain, and damage, based on a phase field (PF) approach. The idea is to take advantage of the ability of the PF method to track complex crack morphologies and to provide a natural way in DIC to perform damage and crack measurements from experimental speckle images, in addition to displacement and strain fields. Moreover, incorporating the damage variable into DIC can improve the displacement accuracy near the crack tip, and can avoid the need of user-defined masks when dealing with cracked samples, which is advantageous when cracks become complex and the manual application of masks becomes challenging. The theoretical formulation of the proposed framework, namely PF-DIC, was presented in detail in the paper, along with a finite element implementation. Numerical examples have demonstrated the capability of the proposed PF-DIC in terms of capturing different types of cracks while providing similar measurement accuracy to that of masked DIC. Additionally, it is shown that the PF-DIC can be easily adapted to selectively identify critical cracks under specific loading conditions or mechanisms for damage assessment and diagnostic purposes. The proposed DIC framework can be used to characterize material defects, support structural health monitoring, and enable a potential unification of PF simulations and experimental fracture measurements

math.NA↗

arXiv2026

Do Speech Tokens Leak Voiceprints? Speaker Inversion Attacks Against End-to-End Speech Language Models

End-to-end speech language models increasingly represent user speech with speech tokens rather than relying exclusively on cascaded ASR--LLM--TTS pipelines. Although these tokens support expressive and low-latency spoken interaction, they may also preserve sensitive speaker characteristics. We investigate whether exposed speech tokens leak voiceprints and formulate this risk as a speaker inversion attack. We introduce Audio BERT (AuB), a trainable model that constructs token embeddings from discrete codebooks and aggregates them into speaker-sensitive representations, and propose SpInv, a two-stage inversion method built on AuB to recover embeddings in the space of an attacker-specified speaker encoder. We evaluate Moshi, Higgs3, Kimi-Audio, and Qwen3-Omni using speaker-disjoint protocols on the VoxCeleb dataset. Extensive experiments show that, with only three seconds of frontend output, SpInv achieves cosine similarities above 0.70 in the attacker-specified speaker-encoder space.

cs.SD↗

arXiv2026

DPIFrame: A Dual-Level Parallelism Acceleration Framework for CTR Model Inference

Deep learning technology has enhanced the ability of Click-through rate (CTR) prediction models to learn features and improve prediction accuracy. However, it is challenging to deploy CTR models on GPU smoothly and perform inference efficiently, because there is a huge mismatch between the serial computational pattern and the parallel model structure. In this paper, we propose DPIFrame, the first dual parallelizable framework to accelerate CTR model inference. In DPIFrame, a) a dual parallelizable architecture is proposed to perform parallel CTR model inference in both intra-module and inter-module; b) an efficient multi-table lookup algorithm is presented for embedding operations through anticipating the whole workload in advance; c) a breadth-first stream scheduling strategy is designed for fine-grained management of parallel computation on GPU to further supporting the dual parallel execution. Extensive experiments are conducted on two real-world datasets, and the results highlight that DPIFrame can reduce the embedding latency efficiently by \textbf{23.0$\times$} compared to PyTorch. Compared with PyTorch, TorchRec, HugeCTR, and OneFlow, DPIFrame can achieve state-of-the-art inference performance on GPU with speedups of \textbf{5.83$\times$}, \textbf{4.29$\times$}, \textbf{2.15$\times$}, and \textbf{2.0$\times$}, respectively.

cs.DC↗

arXiv2026

Elmes*: Automated Construction of Fine-Grained Evaluation Rubrics for Large Language Models in Long-Tail Educational Scenarios

Evaluating large language models (LLMs) for education requires measuring how models teach, not only what they know. Existing benchmarks emphasize domain-general correctness or depend on manually designed rubrics that scale poorly to long-tail pedagogical scenarios. We introduce Elmes*, an end-to-end framework for constructing, refining, and applying fine-grained scenario-specific rubrics. Elmes* combines a declarative multi-agent engine for teacher--student--judge interactions with SceneGen, a self-evolving module that co-optimizes evaluation criteria and test data from expert-defined pedagogical dimensions. Using Elmes*, we build Edu-330, covering 330 scenarios across 11 subjects, 3 grade bands, and 10 task types, with over 1{,}000 second-level indicators. Experiments on Edu-330 and four expert-authored gold-standard scenarios show that educational capability is multidimensional: top-tier LLMs differ mainly in creativity and values integration, knowledge-strong models may fail at Socratic scaffolding, and the education-specialized InnoSpark achieves the best human-evaluated average score. LLM judges preserve human-comparable rankings with much lower scoring variance, but exhibit judge-specific biases such as self-preference. Ablations show that expert-scored few-shot anchoring improves human--LLM alignment, while reasoning enforcement and greedy decoding are model-dependent. Elmes* thus provides scalable diagnostic infrastructure for pedagogically grounded LLM evaluation.

cs.LG↗

arXiv2026

Titan's Resources and their Utilization

Saturn's moon Titan is a unique environment in the solar system. It is the only moon with an atmosphere, composed primarily of the gases N2 and CH4. It is also the only world to have abundant surface hydrocarbons CxHy, which are found as both liquids (seas, lakes) and solids (dunes). Meanwhile, oxygen is also readily accessible in the form of crustal water. This combination of abundant reduced carbon, along with available nitrogen and oxygen makes Titan an enticing world rich in resources that can be readily used to make food, fuel, building materials and more - potentially mission-enabling for long-duration voyages or habitats in the outer solar system. At the same time Titan, as an icy moon, is likely to be depleted at the surface in heavier elements including metals, which must therefore be found and brought from elsewhere. In this article we describe both the available resources on Titan, and also their potential uses. We compare and contrast the resource availability and potential in-situ utilization (ISRU) with other destinations suggested for human habitation such as the Moon and Mars. We conclude by discussing what future work will be important to further characterize Titan's resources, and to develop technologies for their utilization.

astro-ph.EP↗

arXiv2026

An efficient and energy stable framework for phase field simulations of grain growth in additive manufacturing

Phase field simulations play a key role in the understanding of microstructure evolution in additive manufacturing. However, they have been found extremely computationally expensive. One of the reasons is the small time step requirement to resolve the complex microstructure evolution during the rapid solidification process. This paper investigates the possibility of using a class of stabilized time integration algorithms to accelerate such phase field simulations by increasing the time steps, based on a phase field model dedicated to simulating the solidification of 316L stainless steel during additive manufacturing, particularly in a regime where the solid-liquid interface is moving fast and there is absolute interfacial stability with negligible composition variations. The specific computational framework, incorporating the finite element method and the stabilized time integration algorithms, was developed. A theoretical analysis on energy stability was conducted, based on a revisited energy law derived for the phase field model. The numerical results confirmed that the proposed framework can effectively enforce the numerical stability and a decreasing energy requirement for the phase field simulations with at least two orders-of-magnitude larger time steps over conventional explicit methods. 2D and 3D phase field simulations have been conducted with relevant physical and kinetic parameters for 316L stainless steel. This computational framework can be easily adapted for different phase field models and open numerous opportunities for efficient phase field simulations.

physics.comp-ph↗

arXiv2026

AcademiClaw: When Students Set Challenges for AI Agents

Benchmarks within the OpenClaw ecosystem have thus far evaluated exclusively assistant-level tasks, leaving the academic-level capabilities of OpenClaw largely unexamined. We introduce AcademiClaw, a bilingual benchmark of 80 complex, long-horizon tasks sourced directly from university students' real academic workflows -- homework, research projects, competitions, and personal projects -- that they found current AI agents unable to solve effectively. Curated from 230 student-submitted candidates through rigorous expert review, the final task set spans 25+ professional domains, ranging from olympiad-level mathematics and linguistics problems to GPU-intensive reinforcement learning and full-stack system debugging, with 16 tasks requiring CUDA GPU execution. Each task executes in an isolated Docker sandbox and is scored on task completion by multi-dimensional rubrics combining six complementary techniques, with an independent five-category safety audit providing additional behavioral analysis. Experiments on six frontier models show that even the best achieves only a 55\% pass rate. Further analysis uncovers sharp capability boundaries across task domains, divergent behavioral strategies among models, and a disconnect between token consumption and output quality, providing fine-grained diagnostic signals beyond what aggregate metrics reveal. We hope that AcademiClaw and its open-sourced data and code can serve as a useful resource for the OpenClaw community, driving progress toward agents that are more capable and versatile across the full breadth of real-world academic demands. All data and code are available at https://github.com/GAIR-NLP/AcademiClaw.

cs.AI↗

arXiv2026

Relation Reasoning with LLMs in Expensive Optimization

Expensive optimization problems (EOPs) are black-box tasks with costly objective evaluations and no gradient access, making the evaluation budget the key bottleneck. Surrogate-assisted evolutionary algorithms (SAEAs) reduce evaluations via surrogate predictions, but conventional surrogates often require frequent retraining as populations evolve, incurring overhead. This paper proposes R2SAEA, a reinforcement-trained relation-based large language model (LLM) surrogate assisted evolutionary algorithm. We cast relation-based surrogate modeling as an in-context pairwise reasoning task. To enable efficient inference in evolutionary loops, we develop an anchor-based iterative context construction strategy that reduces prompt complexity from quadratic to linear in population size, and a voting-based aggregation scheme that converts predicted relations into scores for offspring selection. We further build an RL pipeline from evolutionary trajectories and fine-tune Qwen2.5 with GRPO. Experiments on single- and multi-objective benchmarks show improved relation prediction and state-of-the-art optimization performance over strong SAEA baselines and general LLMs. Quantization also enables efficient edge deployment, supporting a zero-shot surrogate paradigm without per-generation retraining. Code and models are available at https://github.com/Septend9/R2SAEA.

cs.NE↗

arXiv2026

Accelerating Diffusion-based Video Editing via Heterogeneous Caching: Beyond Full Computing at Sampled Denoising Timestep

Diffusion-based video editing has emerged as an important paradigm for high-quality and flexible content generation. However, despite their generality and strong modeling capacity, Diffusion Transformers (DiT) remain computationally expensive due to the iterative denoising process, posing challenges for practical deployment. Existing video diffusion acceleration methods primarily exploit denoising timestep-level feature reuse, which mitigates the redundancy in denoising process, but overlooks the architectural redundancy within the DiT that many attention operations over spatio-temporal tokens are redundantly executed, offering little to no incremental contribution to the model output. This work introduces HetCache, a training-free diffusion acceleration framework designed to exploit the inherent heterogeneity in diffusion-based masked video-to-video (MV2V) generation and editing. Instead of uniformly reuse or randomly sampling tokens, HetCache assesses the contextual relevance and interaction strength among various types of tokens in designated computing steps. Guided by spatial priors, it divides the spatial-temporal tokens in DiT model into context and generative tokens, and selectively caches the context tokens that exhibit the strongest correlation and most representative semantics with generative ones. This strategy reduces redundant attention operations while maintaining editing consistency and fidelity. Experiments show that HetCache achieves a noticeable acceleration, including a 2.67$\times$ latency speedup and FLOPs reduction over commonly used foundation models, with negligible degradation in editing quality.

cs.CV↗