Searcharxiv⌕ Search

arXiv subjects

Yiran Peng

Publications and source records attributed to Yiran Peng.

At least 19 recordsLinked to original sources
arXiv2026

Foundation Protocol: A Coordination Layer for Agentic Society

Autonomous agents are moving from tools into a layer of social infrastructure: they browse, purchase, deploy software, manage systems, and increasingly interact with one another. As these systems scale, the bottleneck shifts away from raw model capability toward coordination. Agents need to form reliable relationships, organize multi-agent work, exchange value, support an AI economy, and stay safe and accountable under real-world oversight. This paper introduces the Foundation Protocol (FP), a graph-first coordination layer for an emerging human-AI society. FP unifies heterogeneous entities, including agents, tools, resources, humans, institutions, and organizations, and supports native multi-party organization and event-based collaboration. It also provides economic primitives for metering, receipts, and settlement, and treats policy, provenance, and audit as first-class concerns. FP is designed to wrap and bridge existing protocols rather than replace them, enabling incremental adoption while reducing integration and governance overhead. The aim is to keep autonomous agency composable while keeping accountability non-negotiable, so that coordination itself can become shared infrastructure for a human-AI society that is open, pluralistic, and governable.

cs.AI↗
arXiv2026

Harnessing Agentic Evolution

Agentic evolution has emerged as a powerful paradigm for improving programs, workflows, and scientific solutions by iteratively generating candidates, evaluating them, and using feedback to guide future search. However, existing methods are typically instantiated either as fixed hand-designed procedures that are modular but rigid, or as general-purpose agents that flexibly integrate feedback but can drift in long-horizon evolution. Both forms accumulate rich evidence over time, including candidates, feedback, traces, and failures, yet lack a stable interface for organizing this evidence and revising the mechanism that drives future evolution. We address this limitation by formulating agentic evolution as an interactive environment, where the accumulated evolution context serves as a process-level state. We introduce AEvo, a harnessed meta-editing framework in which a meta-agent observes this state and acts not by directly proposing the next candidate, but by editing the procedure or agent context that controls future evolution. This unified interface enables AEvo to steer both procedure-based and agent-based evolution, making accumulated evidence actionable for long-horizon search. Empirical evaluations on agentic and reasoning benchmarks show that AEvo outperforms five evolution baselines, achieving a 26 relative improvement over the strongest baseline. Across three open-ended optimization tasks, AEvo further outperforms four evolution baselines and achieves state-of-the-art performance under the same iteration budget.

cs.AI↗
arXiv2026

Engineering 2D high-temperature ferromagnets with large in-plane anisotropy via alkali-metal decoration in a tetragonal CoSe monolayer

Two-dimensional (2D) ferromagnetic materials with high Curie temperature ($T_{\rm c}$) and large magnetic anisotropy energy (MAE) are critical for nanoscale spintronics but remain rare. We propose, via first-principles calculations, that adsorbing alkali atoms ($A$ = Li, Na, K, Rb, Cs) onto a tetragonal CoSe monolayer transforms it into a series of stable 2D ferromagnetic metals, $A$CoSe, with an in-plane easy axis. Notably, LiCoSe is a half-metal. These functionalized monolayers exhibit dramatically enhanced ferromagnetism compared to the pristine layer, with $T_{\rm c}$ > 300 K and MAE > 800 $\mu$eV/Co. The coupled alkali atoms amplify the local magnetic moment of Co ions, reinforce ferromagnetic Ruderman-Kittel-Kasuya-Yosida (RKKY) and superexchange couplings, and concurrently weaken the direct antiferromagnetic exchange between Co ions. Furthermore, tensile strain can further elevate the MAE (via band shifting) and increase $T_{c}$ (by strengthening the nearest-neighbor exchange $J_1$). Among them, NaCoSe exhibits the highest MAE and excellent strain-modulated $T_{c}$, rendering it the most promising candidate material. Our results establish alkali-metal decoration as an effective strategy for realizing 2D ferromagnets with high $T_{\rm c}$ and large MAE in tetragonal lattices.

cond-mat.mtrl-sci↗
arXiv2026

DeepEye: A Steerable Self-driving Data Agent System

Large Language Models (LLMs) have revolutionized natural language interaction with data. The "holy grail" of data analytics is to build autonomous Data Agents that can self-drive complex data analysis workflows. However, current implementations are still limited to linear "ChatBI" systems. These systems struggle with joint analysis across heterogeneous data sources (e.g., databases, documents, and data files) and often encounter "context explosion" in complex and iterative data analysis workflows. To address these challenges, we present DeepEye, a production-ready data agent system that adopts a workflow-centric architecture to ensure scalability and trustworthiness. DeepEye introduces a Unified Multimodal Orchestration protocol, enabling seamless integration of structured and unstructured data sources. To mitigate hallucinations, it employs Hierarchical Reasoning with context isolation, decomposing complex intents into autonomous AgentNodes and deterministic ToolNodes. Furthermore, DeepEye incorporates a database-inspired Workflow Engine (comprising a Compiler, Validator, Optimizer, and Executor) that guarantees structural correctness and accelerates execution via runtime topological optimization. In this demonstration, we showcase DeepEye's ability to orchestrate complex workflows to generate diverse multimodal outputs -- including Data Videos, Dashboards, and Analytical Reports -- highlighting its advantages in transparent execution, automated optimization, and human-in-the-loop reliability.

cs.DB↗
arXiv2026

The giant anomalous Hall and Nernst effects in Kagome permanent magnets RCo5

Kagome lattice materials have attracted considerable attention due to their intriguing topological properties and potential applications in next-generation quantum and spintronic technologies. In particular, rare-earth permanent magnets with Kagome structure provide an ideal platform that combines robust magnetism with nontrivial quantum phenomena. However, their anomalous transport properties, particularly thermoelectric responses, remain insufficiently explored. In this work, we perform systematic first-principles calculations on the anomalous Hall and anomalous Nernst effects in Kagome permanent magnets RCo5 (R = Ce, La, Sm, Gd). We find that CeCo5 exhibits a pronounced anomalous Hall conductivity of about 1500 Omega^-1 cm^-1 while GdCo5 displays a substantial anomalous Nernst conductivity of 11 A m^-1 K^-1 within +/- 0.1 eV of the Fermi energy, both comparable to or surpassing the measured intrinsic values reported in many typical Weyl and Heusler magnets. These exceptional anomalous transport properties originate from Berry curvature hotspots near spin-orbit coupling induced band gaps. If validated, these theoretical predictions would be important for Berry-curvature-driven transport in magnetic intermetallics. Our results establish RCo5 compounds as versatile platforms for exploring Berry curvature-driven transport in tunable magnetic topological materials.

cond-mat.mtrl-sci↗
arXiv2026

AutoWebWorld: Synthesizing Infinite Verifiable Web Environments via Finite State Machines

The performance of autonomous Web GUI agents heavily relies on the quality and quantity of their training data. However, a fundamental bottleneck persists: collecting interaction trajectories from real-world websites is expensive and difficult to verify. The underlying state transitions are hidden, leading to reliance on inconsistent and costly external verifiers to evaluate step-level correctness. To address this, we propose AutoWebWorld, a novel framework for synthesizing controllable and verifiable web environments by modeling them as Finite State Machines (FSMs) and use coding agents to translate FSMs into interactive websites. Unlike real websites, where state transitions are implicit, AutoWebWorld explicitly defines all states, actions, and transition rules. This enables programmatic verification: action correctness is checked against predefined rules, and task success is confirmed by reaching a goal state in the FSM graph. AutoWebWorld enables a fully automated search-and-verify pipeline, generating over 11,663 verified trajectories from 29 diverse web environments at only $0.04 per trajectory. Training on this synthetic data significantly boosts real-world performance. Our 7B Web GUI agent outperforms all baselines within 15 steps on WebVoyager. Furthermore, we observe a clear scaling law: as the synthetic data volume increases, performance on WebVoyager and Online-Mind2Web consistently improves.

cs.AI↗
arXiv2026

AOrchestra: Automating Sub-Agent Creation for Agentic Orchestration

Language agents have shown strong promise for task automation. Realizing this promise for increasingly complex, long-horizon tasks has driven the rise of a sub-agent-as-tools paradigm for multi-turn task solving. However, existing designs still lack a dynamic abstraction view of sub-agents, thereby hurting adaptability. We address this challenge with a unified, framework-agnostic agent abstraction that models any agent as a tuple Instruction, Context, Tools, Model. This tuple acts as a compositional recipe for capabilities, enabling the system to spawn specialized executors for each task on demand. Building on this abstraction, we introduce an agentic system AOrchestra, where the central orchestrator concretizes the tuple at each step: it curates task-relevant context, selects tools and models, and delegates execution via on-the-fly automatic agent creation. Such designs enable reducing human engineering efforts, and remain framework-agnostic with plug-and-play support for diverse agents as task executors. It also enables a controllable performance-cost trade-off, allowing the system to approach Pareto-efficient. Across three challenging benchmarks (GAIA, SWE-Bench, Terminal-Bench), AOrchestra achieves 16.28% relative improvement against the strongest baseline when paired with Gemini-3-Flash. The code is available at: https://github.com/FoundationAgents/AOrchestra

cs.AI↗
arXiv2025

AutoEnv: Automated Environments for Measuring Cross-Environment Agent Learning

Humans naturally adapt to diverse environments by learning underlying rules across worlds with different dynamics, observations, and reward structures. In contrast, existing agents typically demonstrate improvements via self-evolving within a single domain, implicitly assuming a fixed environment distribution. Cross-environment learning has remained largely unmeasured: there is no standard collection of controllable, heterogeneous environments, nor a unified way to represent how agents learn. We address these gaps in two steps. First, we propose AutoEnv, an automated framework that treats environments as factorizable distributions over transitions, observations, and rewards, enabling low-cost (4.12 USD on average) generation of heterogeneous worlds. Using AutoEnv, we construct AutoEnv-36, a dataset of 36 environments with 358 validated levels, on which seven language models achieve 12-49% normalized reward, demonstrating the challenge of AutoEnv-36. Second, we formalize agent learning as a component-centric process driven by three stages of Selection, Optimization, and Evaluation applied to an improvable agent component. Using this formulation, we design eight learning methods and evaluate them on AutoEnv-36. Empirically, the gain of any single learning method quickly decrease as the number of environments increases, revealing that fixed learning methods do not scale across heterogeneous environments. Environment-adaptive selection of learning methods substantially improves performance but exhibits diminishing returns as the method space expands. These results highlight both the necessity and the current limitations of agent learning for scalable cross-environment generalization, and position AutoEnv and AutoEnv-36 as a testbed for studying cross-environment agent learning. The code is avaiable at https://github.com/FoundationAgents/AutoEnv.

cs.AI↗
arXiv2025

Reasoning via Video: The First Evaluation of Video Models' Reasoning Abilities through Maze-Solving Tasks

Video Models have achieved remarkable success in high-fidelity video generation with coherent motion dynamics. Analogous to the development from text generation to text-based reasoning in language modeling, the development of video models motivates us to ask: Can video models reason via video generation? Compared with the discrete text corpus, video grounds reasoning in explicit spatial layouts and temporal continuity, which serves as an ideal substrate for spatial reasoning. In this work, we explore the reasoning via video paradigm and introduce VR-Bench -- a comprehensive benchmark designed to systematically evaluate video models' reasoning capabilities. Grounded in maze-solving tasks that inherently require spatial planning and multi-step reasoning, VR-Bench contains 7,920 procedurally generated videos across five maze types and diverse visual styles. Our empirical analysis demonstrates that SFT can efficiently elicit the reasoning ability of video model. Video models exhibit stronger spatial perception during reasoning, outperforming leading VLMs and generalizing well across diverse scenarios, tasks, and levels of complexity. We further discover a test-time scaling effect, where diverse sampling during inference improves reasoning reliability by 10--20%. These findings highlight the unique potential and scalability of reasoning via video for spatial reasoning tasks.

cs.CV↗
arXiv2025

Correlation tuned Fermi-arc topology in a Weyl ferromagnet

Electrons on Fermi arcs (FAs), a hallmark of Weyl semimetals, exhibit chiral transport harboring chiral anomaly, negative magnetoresistance, and Majorana zero modes. While FAs were observed in exemplary Weyl semimetal TaAs and Co3Sn2S2, the manipulation of FAs has been rarely explored. Here we take Co3Sn2S2 as an example and demonstrate that tuning the electronic correlation strength is an effective way to control the topology and connectivity of FAs. After achieving a good agreement with experimentally measured band structure by employing combined density functional theory and dynamical mean field theory (DFT+DMFT) calculations, we show that the experimental charge dynamics are well reproduced by DFT+DMFT calculations but not DFT calculations. Electronic correlation renormalizes the bands around the Fermi level and modifies the energy and location of Weyl points, and the resulting FAs. In particular, on the Co-terminated surface, the FAs are formed by connecting Weyl points located in adjacent Brillouin zones in DFT+DMFT calculations and experiments, in strong contrast to the FAs connecting Weyl points within the same Brillouin zone in DFT calculations. We further show the evolution of FAs with correlation and reveal a topological change of the FAs on the Sn-terminated surface at stronger correlation strength. Our study sheds new light on experimental manipulation of FAs to improve the electronic properties of correlated Weyl semimetals.

cond-mat.mtrl-sci↗
arXiv2025

Trainable Dynamic Mask Sparse Attention

The increasing demand for long-context modeling in large language models (LLMs) is bottlenecked by the quadratic complexity of the standard self-attention mechanism. The community has proposed sparse attention to mitigate this issue. However, position-aware sparse attention methods rely on static sparse structures that lack adaptability to diverse query contexts, while content-aware sparse attention methods depend on heuristic key-value selection, hindering full differentiability. We introduce a trainable dynamic mask sparse attention mechanism, a method that merges the advantages of both position-aware and content-aware approaches. Dynamic Mask Attention (DMA) achieves this through three key innovations: First, it leverages value vector representations to generate content-aware dynamic masks, enabling the model to adaptively identify and attend to critical information. Second, it computes position-aware sparse weights in a hardware-friendly manner, efficiently skipping unnecessary computational regions. Finally, we demonstrate that the introduced dynamic mask and sparse weights do not obstruct gradients, supporting end-to-end training. We have validated the performance of DMA through comprehensive experiments. A large body of experimental evidence shows that DMA consistently holds a Pareto advantage over state-of-the-art sparse attention baselines in tasks including scaling laws, multi-query associative recall, standard benchmarks, and needle in a haystack tests, while also delivering up to a 10x overall speedup. These results highlight its ability to effectively balance model efficiency with long-context modeling capabilities. Our computational kernel code is now open-source at https://github.com/SmallDoges/flash-dmattn to encourage further research and application by the community.

cs.AI↗
arXiv2023

Room-temperature non-volatile optical manipulation of polar order in a charge density wave

Utilizing ultrafast light-matter interaction to manipulate electronic states of quantum materials is an emerging area of research in condensed matter physics. It has significant implications for the development of future ultrafast electronic devices. However, the ability to induce long-lasting metastable electronic states in a fully reversible manner is a long-standing challenge.Here, by using ultrafast laser excitations, we demonstrate the capability to manipulate the electronic polar states in the charge-density-wavematerial EuTe4 in a non-volatile manner. The process is completely reversible and is achieved at room temperature with an all-optical approach. Each induced non-volatile state brings about modifications to the electrical resistance and second harmonic generation intensity. The results point to layer-specific phase inversion dynamics by which photoexcitation mediates the stacking polar order of the system. Our findings extend the scope of non-volatile all-optical control of electronic states to ambient conditions, and highlight a distinct role of layerdependent phase manipulation in quasi-two-dimensional systems with inherent sublayer stacking orders.

cond-mat.str-el↗
arXiv2022

Correlation driven near-flat band Stoner excitations in a Kagome magnet

Among condensed matter systems, Mott insulators exhibit diverse properties that emerge from electronic correlations. In itinerant metals, correlations are usually weak, but can also be enhanced via geometrical confinement of electrons, that manifest as `flat' dispersionless electronic bands. In the fast developing field of topological materials, which includes Dirac and Weyl semimetals, flat bands are one of the important components that can result in unusual magnetic and transport behaviour. To date, characterisation of flat bands and their magnetism is scarce, hindering the design of novel materials. Here, we investigate the ferromagnetic Kagomé semimetal Co$_3$Sn$_2$S$_2$ using resonant inelastic X-ray scattering. Remarkably, nearly non-dispersive Stoner spin excitation peaks are observed, sharply contrasting with the featureless Stoner continuum expected in conventional ferromagnetic metals. Our band structure and dynamic spin susceptibility calculations, and thermal evolution of the excitations, confirm the nearly non-dispersive Stoner excitations as unique signatures of correlations and spin-polarized electronic flat bands in Co$_3$Sn$_2$S$_2$. These observations serve as a cornerstone for further exploration of band-induced symmetry-breaking orders in topological materials.

cond-mat.str-el↗
arXiv2022

Correlation-corrected band topology and topological surface states in iron-based superconductors

Iron-based superconductors offer an ideal platform for studying topological superconductivity and Majorana fermions. In this paper, we carry out a comprehensive study of the band topology and topological surface states of a number of iron-based superconductors using a combination of density functional theory (DFT) and dynamical mean field theory. We find that the strong electronic correlation of Fe 3d electrons plays a crucial role in determining the band topology and topological surface states of iron-based superconductors. Electronic correlation not only strongly renormalizes the bandwidth of Fe 3d electrons, but also shifts the band positions of both Fe 3d and As/Se p electrons. As a result, electronic correlation moves the DFT-calculated topological surface states of many iron-based superconductors much closer to the Fermi level, which is crucial for realizing topological superconducting surface states and observing Majorana zero modes as well as achieving practical applications, such as quantum computation. More importantly, electronic correlation can change the band topology and make some iron-based superconductors topologically nontrivial with topological surface states whereas they have trivial band topology and no topological surface states in DFT calculations. Our paper demonstrates that it is important to take into account electronic correlation effects in order to accurately determine the band topology and topological surface states of iron-based superconductors and other strongly correlated materials.

cond-mat.supr-con↗
arXiv2022

Electronic structure and magnetism of the Hund insulator CrI3

CrI3 is a two-dimensional ferromagnetic van der Waals material with a charge gap of 1.1-1.2 eV. In this study, the electronic structure and magnetism of CrI3 are investigated by using density functional theory and dynamical mean-field theory. Our calculations successfully reproduce a charge gap of 1.1 eV in the paramagnetic state when a Hund coupling JH = 0.7 eV is included with an on-site Hubbard U = 5 eV. In contrast, with a large U value of 8 eV and negligible Hund coupling JH, CrI3 is predicted to be a moderately correlated metal in the paramagnetic state. We conclude that CrI3 is a Mott-Hund insulator due to the half-filled configuration of the Cr 3d t2g orbitals. The Cr 3d eg orbitals are occupied by approximately one electron, which leads to strong valence fluctuations so that the Cr 3d orbitals cannot be described by a single state. Moreover, at finite temperature, the calculated ordered static magnetic moment in the ferromagnetic state is significantly larger in the R3 phase than in the C2/m phase. This observation indicates that the structural phase transition from the C2/m phase to the R3 phase with decreasing temperature is driven by ferromagnetic spin fluctuations.

cond-mat.str-el↗
arXiv2022

Correlation-enhanced electron-phonon coupling and superconductivity in (Ba,K)SbO$_3$ superconductors

The electronic structure, lattice dynamics, and electron-phonon coupling (EPC) of the newly discovered (Ba,K)SbO$_3$ superconductors are investigated by first-principles calculations. The EPC of (Ba,K)SbO$_3$ is significantly enhanced by considering non-local electronic correlation using the Heyd-Scuseria-Ernzerhof hybrid exchange-correlation functional (HSE06). The EPC strength λ of Ba$_{0.35}$K$_{0.65}$SbO$_3$ is strongly increased from 0.33 in local-density approximation calculations to 0.59 in HSE06 calculations, resulting in a superconducting transition temperature Tc of about 14.9 K, which is in excellent agreement with experimental value of ~ 15 K. Our findings suggest (Ba,K)SbO$_3$ are extraordinary conventional superconductors, where non-local electronic correlation expands the bandwidth, enhances the EPC, and boosts the Tc. Moreover, we find both λ and Tc depend crucially on the K-doping level for (Ba,K)SbO$_3$ and (Ba,K)SbO$_3$ compounds. (Ba,K)SbO$_3$ have stronger EPC strength and higher Tc than those of (Ba,K)SbO$_3$ at the same K-doping level.

cond-mat.supr-con↗
arXiv2022

Ta2NiSe5: a candidate topological excitonic insulator with multiple band inversions

The electronic structures and topological properties of the orthorhombic and monoclinic phases of the quasi-one-dimensional excitonic insulator Ta2NiSe5 are investigated based on density functional theory. In contrast to a single parity or band inversion across the Fermi level in many topological insulators studied previously, there are multiple parity and band inversions with or without spin-orbit coupling in both phases of Ta2NiSe5, resulting in more complex and topologically nontrivial electronic structures. The Dirac cone type surface states of the low-temperature monoclinic phase are also obtained. In this paper, we demonstrate that Ta2NiSe5 is a promising candidate as a three-dimensional topological excitonic insulator.

cond-mat.mtrl-sci↗
arXiv2021

Computational design of a new layered superconductor LaOTlF2

A new layered compound LaOTlF2 is designed and investigated using first-principles calculations in this work. The parent compound is an insulator with an indirect band gap of 2.65 eV. Electron-doping of the parent compound makes the material metallic. In the meantime, several lattice vibrational modes couple strongly to the conduction band, leading to a large electron-phonon coupling constant and conventional superconductivity. The highest superconducting transition temperature Tc is predicted to be approximately 8.6 K with λ about 1.25 in the optimally doped LaO0.95F0.05TlF2, where λ is calculated using the Wannier interpolation technique.

cond-mat.supr-con↗