SearcharxivSearch

arXiv subjects

Yuchen Peng

Publications and source records attributed to Yuchen Peng.

16 recordsLinked to original sources

Guixu: Valuation-Driven Data Discovery for Autonomous AI Agents with On-Chain Attestation

Autonomous agents increasingly rely on external data to complete downstream tasks such as model training and decision support. However, existing data discovery systems remain largely retrieval-oriented: they surface candidate datasets from heterogeneous sources, but provide limited support for estimating task-specific utility, selecting cost-effective datasets under budget constraints, or incorporating trustworthy feedback from prior usage. This paper presents Guixu, a valuation-driven data discovery system for autonomous agents. Guixu employs a three-phase valuation pipeline with proxy-label propagation and multi-round knapsack optimization for task-aware data valuation. Guixu integrates agentic payment protocol to enable budget-constrained data procurement workflows. Guixu leverages on-chain data market and attestation signals for verifiable data discovery. Our demonstration highlights how Guixu enables an agent to move beyond keyword-based dataset retrieval toward task- and budget-aware, trustworthy data discovery and procurement. Attendees can interactively explore the full workflow, from NL task specification and multi-source search to data valuation and verifiable transaction feedback.

cs.AI

Antichiral hinge states in a higher-order photonic nodal ring semimetal

Antichiral states propagate in the same direction on opposite boundaries, defying the conventional constraint that boundary modes must cancel net chirality. Previously found antichiral states have been limited to first-order topological semimetals. However, antichiral hinge states, the antichiral counterpart of recently discovered higher-order chiral hinge states, remain elusive. Here, we report the observation of antichiral hinge states in a higher-order nodal-ring semimetal made of a three-dimensional gyromagnetic photonic crystal. Near-field scanning measurements reveal a pair of hinge states at two parallel one-dimensional boundaries propagating unidirectionally along the same direction, with robust transport against metallic scatterers. Their spatial positions can be reconfigured by adding or removing photonic layers. The additionally observed antichiral and drumhead surface states manifest a hierarchy of first- and second-order topological boundary states within a single photonic system. Our work extends antichiral states to higher-order topological semimetals and has potential applications in robust and reconfigurable photonic routing.

physics.optics

Learn Before Represent: Bridging Generative and Contrastive Learning for Domain-Specific LLM Embeddings

Large Language Models (LLMs) adapted via contrastive learning excel in general representation learning but struggle in vertical domains like chemistry and law, primarily due to a lack of domain-specific knowledge. This work identifies a core bottleneck: the prevailing ``LLM+CL'' paradigm focuses on semantic alignment but cannot perform knowledge acquisition, leading to failures on specialized terminology. To bridge this gap, we propose Learn Before Represent (LBR), a novel two-stage framework. LBR first injects domain knowledge via an Information Bottleneck-Constrained Generative Learning stage, preserving the LLM's causal attention to maximize knowledge acquisition while compressing semantics. It then performs Generative-Refined Contrastive Learning on the compressed representations for alignment. This approach maintains architectural consistency and resolves the objective conflict between generative and contrastive learning. Extensive experiments on medical, chemistry, and code retrieval tasks show that LBR significantly outperforms strong baselines. Our work establishes a new paradigm for building accurate and robust representations in vertical domains.

cs.IR

SVFusion: A CPU-GPU Co-Processing Architecture for Large-Scale Real-Time Vector Search

Approximate Nearest Neighbor Search (ANNS) underpins modern applications such as information retrieval and recommendation. With the rapid growth of vector data, efficient indexing for real-time vector search has become rudimentary. Existing CPU-based solutions support updates but suffer from low throughput, while GPU-accelerated systems deliver high performance but face challenges with dynamic updates and limited GPU memory, resulting in a critical performance gap for continuous, large-scale vector search requiring both accuracy and speed. In this paper, we present SVFusion, a GPU-CPU-disk collaborative framework for real-time vector search that bridges sophisticated GPU computation with online updates. SVFusion leverages a hierarchical vector index architecture that employs CPU-GPU co-processing, along with a workload-aware vector caching mechanism to maximize the efficiency of limited GPU memory. It further enhances performance through real-time coordination with CUDA multi-stream optimization and adaptive resource management, along with concurrency control that ensures data consistency under interleaved queries and updates. Empirical results demonstrate that SVFusion achieves significant improvements in query latency and throughput, exhibiting a 20.9x higher throughput on average and 1.3x to 50.7x lower latency compared to baseline methods, while maintaining high recall for large-scale datasets under various streaming workloads.

cs.DB

Hybrid skin-topological effect induced by eight-site cells and arbitrary adjustment of the localization of topological edge states

The hybrid skin-topological effect (HSTE) in non-Hermitian systems exhibits both the skin effect and topological protection, offering a novel mechanism for the localization of topological edge states (TESs) in electrons, circuits, and photons. However, it remains unclear whether the HSTE can be realized in quasicrystals, and the unique structure of quasicrystals with multi-site cells may provide novel localization phenomena for TESs induced by the HSTE. We propose an eight-site cell in two-dimensional quasicrystals and realize the HSTE with eight-site nonreciprocal intracell hoppings. Furthermore, we can arbitrarily adjust the eigenfield distributions of the TESs and discover domain walls associated with effective dissipation and their correlation with localization. We present a new scheme to precisely adjust the energy distribution in non-Hermitian quasicrystals with arbitrary polygonal outer boundaries.

cond-mat.mes-hall

Multifrequency and multimode topological waveguides in a Stampfli-triangle photonic crystal with large valley Chern numbers

The multifrequency quantum valley Hall effect (QVHE) with a large valley Chern number has been realized to significantly improve the transmission capacity of topological waveguides and increase the mode density of topological waveguides. However, multifrequency and multimode QVHEs have not been realized simultaneously. In this work, using tight-binding model calculations and numerical simulations, a valley photonic crystal (VPC) consisting of a Stampfli-triangle photonic crystal is constructed, and its multiple degeneracies in the low-frequency and high-frequency bands split simultaneously to realize the QVHE with multiple topological edge states (TESs). The multifrequency and multimode topological transmission with two low-frequency modes and four high-frequency modes is realized by means of simulations and experiments through a Z-shaped waveguide constructed using two VPCs with opposite valley Chern numbers to prove the realization of a large valley Chern number in the two frequency bands. The two low-frequency modes are successfully distinguished with position-dependent selective excitations, which experimentally demonstrates the occurrence of a large valley Chern number. A frequency-dependent multimode beam splitter is theoretically proposed for high-performance integrated photonic device applications. These results provide new ideas for high-efficiency and high-capacity optical transmission and communication devices and their integration; furthermore, they broaden the application range of TESs.

physics.optics

Observation of Topological Corner State Arrays in Photonic Quasicrystals

Recently, the studies of topological corner states (TCSs) are extended from crystals to quasicrystals, which are referred to as higher-order topological quasicrystalline insulators (HOTQIs). However, the TCSs of complete quasi-periodic structure in photonic systems have yet to be demonstrated. Moreover, there is only one TCS in each corner region in higher-order topological insulators (HOTIs). Increasing the number of TCS is expected to increase application potential of TCSs. In this work, HOTQIs in photonic systems are experimentally observed. It is found that HOTQIs possess TCS arrays, and each TCS array contains several TCSs. Furthermore, the universal theoretical framework of the multimer analysis method is improved, and the difference in the average charge density is proposed as a real-space topological index. These results will open up new ideas for investigating highly integrated, multi-region localized TCSs and are expected to provide new ways to explore topological phenomena and the applications of photonic quasicrystals.

cond-mat.mtrl-sci

Aperiodic-quasiperiodic-periodic properties and topological transitions in twisted nested Moiré patterns

The Moiré patterns generated by altering the structural parameters in a two or more layers of periodic materials, including single-layer structure, interlayer stacking, and twisting parameters, exhibit prosperous topological physical properties. However, the intricate characteristics of twisted nested Moiré patterns and their relationship with topological transitions remain unclear. In this Letter, based on the proposed twisted nested photonic crystal (TNPC), we derive its spatial geometric functions (SGFs), aperiodic-quasiperiodic-periodic properties in twisted nested Moiré patterns, and the SSHϕ Hamiltonian. We reveal the intrinsic correlation between twisted nested Moiré patterns and topological transitions, obtaining higher-order topological states (HOTSs) with C2z symmetry. This work will provide theoretical references for the design and application of twisted topological PC and their devices.

cond-mat.mes-hall

AgentLens: Visual Analysis for Agent Behaviors in LLM-based Autonomous Systems

Recently, Large Language Model based Autonomous system(LLMAS) has gained great popularity for its potential to simulate complicated behaviors of human societies. One of its main challenges is to present and analyze the dynamic events evolution of LLMAS. In this work, we present a visualization approach to explore detailed statuses and agents' behavior within LLMAS. We propose a general pipeline that establishes a behavior structure from raw LLMAS execution events, leverages a behavior summarization algorithm to construct a hierarchical summary of the entire structure in terms of time sequence, and a cause trace method to mine the causal relationship between agent behaviors. We then develop AgentLens, a visual analysis system that leverages a hierarchical temporal visualization for illustrating the evolution of LLMAS, and supports users to interactively investigate details and causes of agents' behaviors. Two usage scenarios and a user study demonstrate the effectiveness and usability of our AgentLens.

cs.HC

Pseudospin-Induced Asymmetric Field in Non-Hermitian Photonic Crystals with Multiple Topological Transitions

We report the discovery of multiple topological phase transitions induced by the photonic quantum spin Hall effect (PQSHE) in a non-Hermitian photonic crystal (PhC). When increasing the magnitude of the non-Hermitian terms, the system undergoes transitions from topological corner states to topological edge states and subsequently to topological bulk states. The angular momentum of the wave function of the out-of-plane electric field excited by the chiral source acts as the pseudospin degree of freedom in the PQSHE. Therefore, we consider the introduction of chiral sources with different circular polarizations into non-Hermitian PhCs and observe the emergence of asymmetric field responses. These results are expected to enable the multi-dimensional manipulation of topological states, offering a new avenue for the detection of chiral sources.

physics.optics

Simple theoretical model for parity-time-symmetric metasurfaces

Many new possibilities to observe and use novel physical effects are discovered at so called exceptional points (EPs). This is done by using parity-time (PT) -symmetric non-Hermitian systems and balancing gains and losses. When combined with EP-physics, recently, metasurfaces have shown greater abilities for wave manipulation than conventional metasurface systems. However, the solving process for EPs usually requires the transfer matrix method (TMM) or a parametric sweep, which are both complex and time-consuming. In this Letter, we develop a simple theoretical model, which is based on acoustic equivalent-circuit theory and can find the analytic solutions for EPs directly. As a proof of concept, PT-symmetric acoustic metasurfaces are studied to test the theoretical model, which enables unidirectional antireflection effects at EPs. In addition, finite element method (FEM) simulations are performed to study these EP solutions using the theoretical model for different mediums, wavelengths, angles of incidence, and gain-loss ratios. Our work offers a simple and powerful theoretical tool for designing PT-symmetric metasurfaces at EPs and may also be used for other classical wave systems.

physics.app-ph

High-Order Nonreciprocal Add-Drop Filter

Topological photonics have led to the robust optical behavior of the device, which has solved the problem of the influence of manufacturing defects and perturbations on the device performance. Meanwhile, temporal coupled-mode theory (t-CMT) has been developed and applied widely. However, the t-CMT of cascaded coupling cavities (CCC) system and its corresponding high-order filter has yet to be established. Here the t-CMT of CCC system is established based on the existing t-CMT. By combining the CCC with the topological waveguides, a versatile design scheme of the high-order nonreciprocal add-drop filter (HONAF) is proposed. The relationship between coupling effect of cavities and transmission and filtering performance of HONAF is analyzed quantitatively, then a method to improve the transmission efficiency and quality factor of the filter is given. Based on the combination of gyromagnetic photonic crystals and decagonal Penrose-type photonic quasicrystals, a HONAF is proposed. The transmission and filtering performance of the HONAF are numerically analyzed, which verifies the consistency between the theoretical prediction and the numerical simulation. The t-CMT of CCC system established can be widely used in the coupled resonator optical waveguides and their related systems. The designed HONAF can also be applied and compatible to microwave communication system.

physics.optics

Coupled cavity-waveguide system based on topological corner state and edge state

Topological corner state (TCS) and topological edge state (TES) have provided new approaches to control the propagation of light. The construction of topological coupled cavity-waveguide system (TCCWS) based on TCS and TES is worth looking forward to, due to its research prospects in realizing high-performance micro-nano integrated photonic devices. In this Letter, TCCWS is proposed in two-dimensional (2D) photonic crystal (PC), which possesses strong optical localization, high quality factor and excellent robustness compared with the conventional coupled cavity-waveguide system (CCCWS). This work will provide the possibility to design high-performance logic gates, lasers, filters and other micro-nano integrated photonics devices and expand their applications.

physics.optics

Frequency tunable topological edge states of two-dimensional honeycomb lattice photonic crystals

In this paper, the photonic quantum spin Hall effect (PQSHE) is realized in dielectric two-dimensional (2D) honeycomb lattice photonic crystal (PC) by stretching and shrinking the honeycomb unit cell. Combining two honeycomb lattice PCs with a common photonic band gap (PBG) but different band topologies can generate a topologically protected edge state at the combined junction. The topological edge states and their unidirectional transmission as the scatterers with triangular, pentagonal, and heptagonal shapes are researched. Meanwhile, the unidirectional transmission in an inverted Ω-shaped waveguide with large bending angle is realized, and verifies the characteristics of the topological protection by adding different kind of defects. Moreover, the frequency varies significantly when changing the scatterers shape, which shows that the PC with various scatterers shape can tune the frequency range of the topological edge state significantly. In other words, it can adjust the frequency of unidirectional transmission and increase the adjustability of the topological edge state.

physics.optics

High-fidelity quantum state preparation using neighboring optimal control

We present an approach to single-shot high-fidelity preparation of an $n$-qubit state based on neighboring optimal control theory. This represents a new application of the neighboring optimal control formalism which was originally developed to produce single-shot high-fidelity quantum gates. To illustrate the approach, and to provide a proof-of-principle, we use it to prepare the two qubit Bell state $|β_{01}\rangle = (1/\sqrt{2})\left[\, |01\rangle + |10\rangle\,\right]$ with an error probability $ε\sim 10^{-6}$ ($10^{-5}$) for ideal (non-ideal) control. Using standard methods in the literature, these high-fidelity Bell states can be leveraged to fault-tolerantly prepare the logical state $|\overlineβ_{01}\rangle$.

quant-ph

Improving quantum gate performance through neighboring optimal control

Successful implementation of a fault-tolerant quantum computation on a system of qubits places severe demands on the hardware used to control the many-qubit state. It is known that an accuracy threshold $P_{a}$ exists for any quantum gate that is to be used in such a computation. Specifically, the error probability $P_{e}$ for such a gate must fall below the accuracy threshold: $P_{e} < P_{a}$. Estimates of $P_{a}$ vary widely, though $P_{a}\sim 10^{-4}$ has emerged as a challenging target for hardware designers. In this paper we present a theoretical framework based on neighboring optimal control that takes as input a good quantum gate and returns a new gate with better performance. We illustrate this approach by applying it to all gates in a universal set of quantum gates produced using non-adiabatic rapid passage that has appeared in the literature. Performance improvements are substantial, both for ideal and non-ideal controls. Under suitable conditions detailed below, all gate error probabilities fall well below the target threshold of $10^{-4}$.

quant-ph