arXiv Science⌕ Search

arXiv subjects

Peng Yan

Publications and source records attributed to Peng Yan.

At least 19 recordsLinked to original sources
arXiv2026

DualEvolve: Dual-Level Self-Evolution of Mechanism and Orchestration for Recommendation

Recommender systems are widely used in e-commerce and content platforms, where improvement requires repeated model design, training optimization, and experimental validation. Large language model (LLM) agents increasingly automate this development cycle. Sustained autonomous improvement requires agents to improve both recommendation models and their development workflows. However, task-performance-driven approaches select updates mainly by candidates' final scores, without a separate optimization signal for experimentation quality. We therefore present DualEvolve, a dual-level framework that co-evolves recommendation mechanisms and agent orchestration toward recursive self-improvement. Inner evolution revises model and training code using task feedback to improve recommendation performance. Outer evolution scores the inner-loop experimental process, updates harness and skill rules using these scores and execution traces, and adjusts orchestration choices based on next-round scores to improve the quality of subsequent experiments. Experiments on MovieLens-1M and three Amazon Reviews 2023 subsets cover four recommendation methods and 16 settings. DualEvolve achieves higher test means than Base, OpenEvolve, and DeepEvolve in all 64 setting-metric comparisons, with macro-averaged relative gains over DeepEvolve of 7.6% (HR@5) and 9.0% (NDCG@5).

cs.IR↗
arXiv2026

SPADE: Spatiotemporal Preference Alignment and Diversity Enhancement for Local Search

Proactive query recommendation aims to suggest relevant search queries to users on the homepage of local service apps. Unlike conventional recommendation over a fixed item set, this requires open-set generation from a vocabulary of millions, where outputs must simultaneously align with user preference, match fine-grained spatiotemporal context, cover long-tail intents, and provide diverse alternatives within a few candidate slots. Existing approaches address these axes in isolation: standard tokenization suppresses long-tail queries, coarse location encoding misses fine-grained geographic demand, generic preference optimization fails to align with production impression-click distributions, and standard beam search collapses to near-duplicate outputs. We present SPADE (Spatiotemporal Preference Alignment and Diversity Enhancement), an end-to-end generative framework that jointly optimizes the generation space, contextual representation, preference distribution, and output slate quality through four components: Objective-Aligned BPE (OA-BPE) aligns subword merging with business value signals to boost long-tail query generation probability; Spatio-Temporal Adaptive Representation (STAR) replaces coarse location tokens with continuous multi-scale encodings and soft gating for precise spatiotemporal grounding; MARS-DPO (Multi-dimensional Adaptive Reward Sampling DPO) combines cross-domain pre-training with preference alignment on production impression-click data via composite rewards and adaptive hard negative sampling; and Sigmoid Diversity Beam Search (SDBS) improves intent coverage without sacrificing recall. Offline experiments on large-scale production data show that SPADE outperforms the strongest generative baseline by +5.6% HR@5, -40.3% GeoC@5, and +1.9% Exec@5. A one-week online A/B test confirms production impact: +1.88% Module CTR, +0.177% Search Volume, and +0.105% Search DAU.

cs.IR↗
arXiv2026

Magnetic Quadrupole Lens and Current-Induced Quadrupolar Dynamics of Antiskyrmion

We demonstrate that antiskyrmions can serve as magnetic quadrupole lenses that focus or defocus electric currents traversing them. We showcase the antiskyrmion core as a focal point where electron beams converge for specific injection directions, leading to anisotropic and nonuniform current distributions. Thus the current-induced antiskyrmion dynamics is profoundly modified by the magnetic quadrupole lensing via the spin-transfer torque. By incorporating quantum transport and micromagnetic simulations iteratively, we self-consistently capture the coupled dynamics of conduction electrons and antiskyrmions. Our results unveil a quadrupolar antiskyrmion Hall effect that is highly anisotropic with respect to the current direction and can be effectively tuned by the spin-orbit coupling. In particular, the current focusing (defocusing) dramatically enhances (suppresses) the antiskyrmion speed while simultaneously reducing its Hall angle, thereby overcoming a key challenge for antiskyrmion racetrack memories. Our work establishes the intrinsic current lensing as a powerful mechanism for manipulating topological magnetic textures and opens new avenues for antiskyrmion-based spintronic devices.

cond-mat.mes-hall↗
arXiv2026

Instability of two-dimensional nonrelativistic altermagnets

Altermagnets (AMs) are collinear compensated magnets that exhibit momentum-dependent spin splitting without requiring a net magnetization. Since local magnetic moments and exchange-driven order do not rely on relativistic effects, altermagnetism can be naturally formulated in a nonrelativistic spin-group framework. This raises a basic question: can two-dimensional altermagnetic order be stabilized by purely exchange interactions in the absence of spin-orbit-induced anisotropy? We address this question by applying Bogoliubov's inequality to a minimal $d$-wave altermagnetic spin model. We show that the anisotropic exchange pattern responsible for altermagnetism still contributes only a quadratic long-wavelength term to the Bogoliubov denominator. Consequently, short-ranged exchange interactions alone cannot stabilize long-range altermagnetic order in two dimensions when continuous spin-rotation symmetry is preserved. In contrast, the corresponding three-dimensional system has a finite small-momentum contribution and is not ruled out by the Mermin-Wagner argument.

cond-mat.mtrl-sci↗
arXiv2026

Learning from Unreachable Rewards: Hint-Conditioned Reinforcement Learning for Generative Recommendation

Semantic-ID generative recommenders represent each item as a short sequence of discrete semantic tokens and predict the next item by autoregressively generating this token sequence. This paradigm enables a unified generation interface for item IDs, histories, and item text, but it also creates a structured optimization bottleneck during reward-based post-training: when an early semantic token enters the wrong branch of the item-token space, finite rollout groups rarely reach the ground-truth item, so group-relative optimization receives identical zero rewards and produces no useful advantage. We propose Hint-Conditioned Generative Recommendation (HCGRec), a semantic-ID generative recommendation framework that recovers learning signal for such hard training instances. HCGRec diagnoses each instance with checkpoint rollouts and supplies a minimal target-prefix hint only when the current generator cannot reach the correct item. The model then generates the unhinted suffix under the hinted semantic branch, turning zero-reward groups into informative comparisons over item-token completions. Hinting also changes token identity: hinted prefix tokens are oracle-provided item context, while unhinted suffix tokens are sampled generation actions. We therefore introduce hint-aware credit decomposition, using supervised learning to preserve item-semantic and prefix-structure alignment for hinted tokens and GRPO to optimize the sampled suffix. Experiments on sequential recommendation benchmarks show that HCGRec substantially improves over supervised fine-tuning and vanilla reward-based post-training, while reducing zero-advantage training samples from over 70% to below 20%. The code is accessible at https://github.com/WncFht/GRec.

cs.IR↗
arXiv2026

Nonlinear Dynamics of Hopfion for Frequency Multiplication

Hopfions, associated with higher-dimensional topology through the Hopf fibration, exhibit {exotic} features like {complex knot} and improved stability compared to skyrmions, enhancing their appeal for innovative applications. In this paper, we study the nonlinear response of magnetic Hopfion to microwave fields. {We observe the emergence of higher-order harmonics of the driving microwave field as it interacts with the Hopfion}. By carefully selecting the {driving frequency, the corresponding harmonic can efficiently excite localized magnon state of a Hopfion. Our results demonstrate the promising potential of hopfions in nonlinear magnonics.

cond-mat.mes-hall↗
arXiv2026

Engineering nonlinear magnon scattering in artificial spin ice via vertex dipolar control

Artificial spin ice (ASI), composed of geometrically frustrated arrays of interacting nanoislands, provides a versatile platform for reprogrammable magnonic functionality. However, the commonly used geometric control parameters such as island length, width, and aspect ratio simultaneously modify the island footprint, inter island dipolar spacing, and shape anisotropy, making it difficult to tune the nonlinear response independently of the linear spectrum and lattice density. Using micromagnetic simulations of kagome ASI under strong microwave drive, we identify edge curvature as a geometric degree of freedom that separates nonlinear magnon scattering from the island footprint. Sharp tipped islands predominantly generate integer harmonics, whereas dumbbell shaped tips produce a transition toward subharmonic rich spectra by concentrating demagnetizing and exchange fields near the island ends without changing the overall island volume or lattice spacing. By mapping the curvature and drive parameter space, we identify a continuous threshold for subharmonic onset controlled by tip curvature. We further show that the angular dependence of the second harmonic amplitude reverses between sharp and dumbbell geometries, providing an experimentally accessible signature of curvature localized nonlinearity. Width and leg length asymmetry can also modify harmonic amplitudes, but they do not remove the intrinsic trade off between footprint and coupling. These results establish tip curvature as a footprint preserving design parameter for engineering nonlinear magnon scattering in ASI, with implications for reconfigurable magnonic devices.

cond-mat.mes-hall↗
arXiv2026

TrendFact: A Benchmark Towards Hotspot Perception in Automatic Fact-Checking

With the surge of online misinformation, Large Language Models (LLMs) and Reasoning Large Language Models (RLMs) serving as Automatic Fact-Checking (AFC) systems have emerged as a prominent paradigm for reliable, explainable verification. However, our empirical study reveals that this paradigm faces a critical risk asymmetry challenge when deployed in the real world under resource-constrained environments. While Hotspot Perception Ability (HPA), the capacity to dynamically allocate reasoning resources based on social impact, is essential to mitigate this risk, existing benchmarks lack the social metadata and evaluation framework to meet this urgent evaluation needs, thereby hindering the advancement of these AFC systems. To bridge this gap, we introduce TrendFact, the first benchmark capable of evaluating HPA and three fact-checking tasks. It consists of 7,643 curated samples sourced from trending platforms and professional datasets, with an evidence library containing 366,634 entries. To enable HPA assessment, we propose two novel metrics: the Explanation Consistency Score (ECS) to evaluate the reliability of verification reasoning, and the Hotspot Claim Perception Index (HCPI) to quantify the overall HPA of AFC systems. Extensive experiments demonstrate that existing AFC systems exhibit limited performance on TrendFact. Furthermore, our proposed FactISR framework effectively enhances HPA and computational efficiency for RLMs-served AFC systems.

cs.CL↗
arXiv2026

Oxidation-induced ultrafast spin-to-orbital conversion at heavy-metal interfaces

Oxidation engineering provides a route to control orbital degrees of freedom, yet its role in spin-to-orbital conversion remains largely unexplored. Here, we report an efficient spin-to-orbital conversion mechanism driven by interfacial oxidation at heavy-metal interfaces. In W/Co/SiO2 heterostructures, terahertz emission exhibits a time delay that scales linearly with the W thickness, identifying orbital-current transport as the dominant origin. The emission amplitude is approximately three times larger than that of Co/Pt bilayers, indicating highly efficient conversion from spin to orbital angular momentum. Systematic variation of Co thickness, stoichiometry, and interface configuration reveals that the effect originates from oxidation of the W layer at the W/Co interface, which modulates the interfacial orbital texture. We further show that this mechanism is generic across different heavy metals and scales with their spin-orbit coupling strength. These results establish oxidation as an effective handle to engineer spin-to-orbital conversion and provide a general route toward orbitronic terahertz emitters.

cond-mat.mes-hall↗
arXiv2026

Beyond Matching: Category-Guided Latent Intent Reasoning for Generative Retrieval in E-Commerce

Generative retrieval offers a new paradigm for e-commerce search by mapping user queries directly to product Semantic Identifiers (SIDs). However, e-commerce queries are often short, noisy, attribute-heavy, and associated with multiple category-consistent products, creating a substantial representation gap between natural-language shopping intent and artificially constructed item SIDs. Explicit Chain-of-Thought (CoT) reasoning can help bridge this gap, but its extra generation cost is difficult to reconcile with the low-latency requirements of online e-commerce systems. To address this challenge, we propose CaLIR (Category-guided Latent Intent Reasoning), a category-guided latent intent reasoning framework for e-commerce generative retrieval. Rather than generating explicit textual rationales, CaLIR learns continuous latent intent states before SID decoding and uses product category hierarchies as a natural scaffold for coarse-to-fine intent reasoning. Specifically, we introduce hierarchical semantic reasoning to align latent states with category-level shopping intent, and query-wise reasoning enhancement to model diverse intent paths under multi-positive queries. CaLIR further combines a query-specific dynamic prefix trie, assembled from pre-indexed category-level tries, with reasoning-aware constrained decoding. Experiments on multilingual e-commerce search datasets show that CaLIR achieves a better balance between retrieval effectiveness and inference efficiency than existing methods, while also demonstrating transferability and robustness across induced hierarchies and different generative backbones.

cs.IR↗
arXiv2026

Combinatorial Survey of Structural Phase Distribution and Magnetism in Fe-Ge-Te Composition-spread Thin Film Libraries

Recently, magnetic 2-dimensional (2D) van der Waals (vdW) materials have garnered tremendous attention. The vdW ferromagnet Fe5Ge1Te2 has a Curie temperature Tc of ~ 270 K, which is tailorable by tuning the stoichiometry and the Fe deficiency to reach room temperature. To explore the expanded compositional space, we implemented combinatorial synthesis and high-throughput characterization to investigate the structural phase distribution and ferromagnetism of a Fe-Ge-Te thin film library. The library was prepared by magnetron co-sputtering followed by annealing in vacuum or in an inert environment. Composition and structural phase distribution of the 177 pads in the library were characterized using high-throughput wavelength dispersive spectroscopy (WDS), X-ray diffraction (XRD), and two-point probe resistance measurements. We leverage unsupervised machine learning to cluster the XRD dataset into groups of compositions with similar structural phases, and further study the ferromagnetic properties via SQUID magnetometry and X-ray magnetic circular dichroism (XMCD) across different clusters. The results are compared against magnetization and structural models calculated using DFT. Our results demonstrate that the hexagonal crystal structure is a critical prerequisite for ferromagnetism in this system, and that unexplored materials adopting this structure can be efficiently identified as possible ferromagnetic materials using our high-throughput, ML-assisted framework. This workflow based on the combinatorial strategy allows us to rapidly capture the composition-structure-magnetic property map across a broad compositional landscape of novel magnetic materials.

cond-mat.mtrl-sci↗
arXiv2026

Arbitrary-genus dark soliton gases in the defocusing nonlinear Schrödinger hydrodynamics

The defocusing nonlinear Schrödinger hydrodynamics supports exact dark solitons under finite density boundary conditions. However, the dark soliton gas, an interacting ensemble of dark solitons, has not yet been studied. In this work, we introduce an arbitrary-genus potential of dark soliton gases by considering the limit of the $\mathcal{N}$-dark soliton as $\mathcal{N}\to \infty$. The large-space asymptotics and long-time evolution of this dark soliton gas potential are analytically investigated through Deift-Zhou nonlinear steepest descent approach. The genus-$N$ dark soliton gas potential approaches the genus-$N$ finite-gap solution as $x \to -\infty$ and the background $1$ as $x \to +\infty$. In the long-time evolution, as the self-similar variable $ξ=x/t$ increases, the gas configuration exhibits a cascade of behaviours, passing from unmodulated and modulated genus-$N$ regions and progressively reducing the genus down to the planar region (unmodulated genus-$0$ region). Notably, the evolution of lower-genus soliton gases can be embedded within that of higher-genus gases, exhibiting identical dynamics within specific regimes. This phenomenon is encoded by the underlying spectra. We also include numerical validations, in perfect agreement with the theoretical predictions.

math-ph↗
arXiv2026

Emergence of Nontrivial Topological Magnon States in Skyrmionium Lattices with Zero Topological Charge

We predict the emergence of nontrivial topological magnon states in the skyrmionium lattice with zero topological charge. We propose the concept of weighted magnetic flux, which provides a clear physical picture for this anomalous phenomenon. We also map the skyrmionium lattice onto the Haldane model, offering an alternative framework for interpreting this. Our findings challenge the conventional wisdom that such states are linked to nonzero topological charge in skyrmion lattices, offering a new perspective in topological magnonics. To facilitate experimental validation, we propose two methods for preparing the skyrmionium lattice and calculate the induced magnon thermal Hall conductivity, which is a key indicator in transport measurements.

cond-mat.mes-hall↗
arXiv2026

A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions

Agents for computer use (ACUs) are an emerging class of systems capable of executing complex tasks on digital devices -- such as desktops, mobile phones, and web platforms -- given instructions in natural language. These agents can automate tasks by controlling software via low-level actions like mouse clicks and touchscreen gestures. However, despite rapid progress, ACUs are not yet mature for everyday use. In this survey, we investigate the state-of-the-art, trends, and research gaps in the development of practical ACUs. We provide a comprehensive review of the ACU landscape, introducing a unifying taxonomy spanning three dimensions: (I) the domain perspective, characterizing agent operating contexts; (II) the interaction perspective, describing observation modalities (e.g., screenshots, HTML) and action modalities (e.g., mouse, keyboard, code execution); and (III) the agent perspective, detailing how agents perceive, reason, and learn. We review 87 ACUs and 33 datasets across foundation model-based and classical approaches through this taxonomy. Our analysis identifies six major research gaps: insufficient generalization, inefficient learning, limited planning, low task complexity in benchmarks, non-standardized evaluation, and a disconnect between research and practical conditions. To address these gaps, we advocate for: (a) vision-based observations and low-level control to enhance generalization; (b) adaptive learning beyond static prompting; (c) effective planning and reasoning methods and models; (d) benchmarks that reflect real-world task complexity; (e) standardized evaluation based on task success; (f) aligning agent design with real-world deployment constraints. Together, our taxonomy and analysis establish a foundation for advancing ACU research toward general-purpose agents for robust and scalable computer use.

cs.AI↗
arXiv2026

Spin Elasticity

Elasticity has long been regarded as a property exclusive to material media. Here we uncover its hidden existence in the spin degree of freedom. We introduce spin elasticity-an intrinsic mechanism that governs recoverable deformation of spin morphology. This discovery reveals a previously unrecognized universality: elasticity operates in both matter and spin spaces, underpinning structural integrity across physical realms. By establishing the missing spin counterpart, this work completes the elastic picture and points toward a broader paradigm where elasticity transcends its conventional boundaries.

cond-mat.mes-hall↗
arXiv2026

Polarization transfer force on ferroelectric domain walls

We investigate the dynamics of ferroelectric textures driven by polarization currents. We show that, ferrons, the quanta of collective polarization excitations, provide an exotic driving mechanism for domain wall (DW) dynamics, compared with their magnonic counterparts. By mapping the linear polarization dynamics of a DW onto a Schrödinger-like problem with a Pöschl-Teller potential, we show that polarization waves are fully transmitted and therefore do not exert a net force on the DW in the linear regime. However, intrinsic nonlinearities give rise to a negative radiation pressure that pulls the DW toward the source. This mechanism allows efficient DW control by optical excitation and temperature gradients with application potential in ferroelectric memory and logic devices.

cond-mat.mtrl-sci↗
arXiv2026

Stimulated Magnonic Frequency Combs

Magnonic frequency combs, characterized by a series of discrete frequency lines, have emerged as a promising frontier in magnon spintronics, with potential applications in advanced information processing and sensing technologies. Although the three-magnon scattering process is widely recognized as a fundamental mechanism for generating these combs, its experimental realization has remained challenging due to the high threshold power and strict conservation of momentum and energy. In this work, we propose a novel mechanism for the stimulated generation of magnonic frequency combs that overcomes these limitations. Our approach offers precise and efficient control over key comb properties, including spacing between spectral lines and the number of lines, marking a significant advancement in the field. We substantiate this mechanism through a robust combination of theoretical modeling, micromagnetic simulations, and experimental validation. This study not only demonstrates the feasibility of our method but also opens new pathways for integrating magnonic frequency combs into practical spintronic devices.

cond-mat.mes-hall↗
arXiv2026

Voltage-controlled topological spin textures in the monolayer limit

The physics of phase transitions in low-dimensional systems has long been a subject of significant research interest. Long-range magnetic order in the strict two-dimensional limit, whose discovery circumvented the Mermin-Wagner theorem, has rapidly emerged as a research focus. However, the demonstration of a non-trivial topological spin textures in two-dimensional limit has remained elusive. Here, we demonstrate the out-of-plane electric field breaks inversion symmetry while simultaneously modulating the electronic band structure, enabling electrically tunable spin-orbit interaction for creation and manipulation of topological spin textures in monolayer CrI3. The realization of ideal two-dimensional topological spin textures may offer not only an experimental testbed for probing the Berezinskii-Kosterlitz-Thouless mechanism, but also potential insights into unresolved quantum phenomena including superconductivity and superfluidity. Moreover, voltage-controlled spin-orbit interaction offers a novel pathway to engineer two-dimensional spin textures with tailored symmetries and topologies, while opening avenues for skyrmion-based next-generation information technologies.

cond-mat.mes-hall↗