arXiv subjects

Ruiqi Hu

Publications and source records attributed to Ruiqi Hu.

16 recordsLinked to original sources

arXiv2026

GeoTrussRover: Morphological Computation with Contact-Semantic Control Primitives

Reconfigurable robots can change their contact geometry when a fixed body cannot negotiate an obstacle. A variable-geometry truss (VGT) distributes this shape change through a load-bearing structure, but coupling it to a mobile base creates a high-dimensional coordination problem. GeoTrussRover combines an electrically actuated VGT, a wheeled base, and contact-semantic morphology planning and control. We solve one source traversal and extract four contact-semantic primitives that describe coordination among 21 members. Physics-constrained projection adapts them to unseen step heights with the same contact topology. When every phase remains feasible, adaptation does not recompute the complete motion. If one phase violates the new physical constraints, only that phase is recomputed. A full-space QP then tracks the adapted motion and corrects member and wheel errors. For transfer from 0.10m to 0.075m, the method reduces objective-function evaluations by 63.7% relative to full recomputation. Contact-phase feasibility analysis covers step heights from 0.10 to 0.46m, or 1.08 to 4.97 wheel radii, with the upper value near the theoretical feasible boundary. The electric prototype traverses 2.11 wheel radii. The resulting low-dimensional representation stores task coordination in a hyper-redundant, load-bearing morphology and reuses it during locomotion.

cs.RO↗

arXiv2026

EAGG: Embodiment-Aligned Grasp Generation via Geometry-Aware Graph Conditioning

Cross-end-effector grasp generation seeks a unified model that generalizes across objects and across embodiments ranging from parallel grippers to dexterous end effectors. Existing grasp generators are typically designed for a fixed embodiment or encode embodiment identity with a static descriptor, which weakens transfer when topology, actuation coupling, and contact geometry differ substantially. We present EAGG, an embodiment-aligned grasp generator that represents each embodiment with a topology-aware end-effector graph and an embodiment-specific low-dimensional end-effector control space. A frozen end-effector-cognition backbone converts the current articulated state into geometry-aware tokens that act as a reusable morphology prior, and iterative geometry injection refreshes these tokens throughout sampling so that conditioning remains synchronized with the evolving end-effector geometry. On the MultiGripperGrasp benchmark, EAGG reaches 56.17% average success across six training end effectors, remaining within 1.10 percentage points of specialized training while preserving transfer to finetuning and zero-shot end effectors. Iterative geometry injection further reduces the pooled median contact distance from 0.239 cm to 0.189 cm. These results show that cross-end-effector grasp generation is strengthened by aligning embodiment structure inside a shared generator rather than suppressing embodiment differences. Code is available at https://github.com/wanhaoniu/EAGG.

cs.RO↗

arXiv2026

Identification and Structural Characterization of Twisted Atomically Thin Bilayer Materials by Deep Learning

Two-dimensional materials are expected to play an important role in next-generation electronics and optoelectronic devices. Recently, twisted bilayer graphene and transition metal dichalcogenides have attracted significant attention due to their unique physical properties and potential applications. In this study we describe the use of optical microscopy to collect the color space of chemical vapor deposition (CVD) molybdenum disulfide ($\mbox{MoS}_2$), and the application of a semantic segmentation convolutional neural network (CNN) to accurately and rapidly identify thicknesses of $\mbox{MoS}_2$ flakes. A second CNN model is trained to provide precise predictions on the twist angle of CVD-grown bilayer flakes. This model harnessed a dataset comprising over 10,000 synthetic images, encompassing geometries spanning from hexagonal to triangular shapes. Subsequent validation of the deep learning predictions on twist angles was executed through the second harmonic generation and Raman spectroscopy. Our results introduce a scalable methodology for automated inspection of twisted atomically thin CVD-grown bilayer.

cond-mat.mtrl-sci↗

arXiv2026

Building an Affordable Self-Driving Lab: Practical Machine Learning Experiments for Physics Education Using Internet-of-Things

Machine learning (ML) is transforming modern physics research, but practical, hands-on experience with ML techniques remains limited due to cost and complexity barriers. To address this gap, we introduce an affordable, autonomous, Internet-of-Things (IoT)-enabled experimental platform designed specifically for applied physics education. Utilizing an Arduino microcontroller, a customizable multi-wavelength light emitting diode (LED) array, and photosensors, our setup generates diverse, real-time optical datasets ideal for training and evaluating foundational ML algorithms, including traversal methods, Bayesian inference, and deep learning. The platform facilitates a closed-loop, self-driving experimental workflow, encompassing automated data collection, preprocessing, model training, and validation. Through systematic performance comparisons, we demonstrate the superior ability of deep learning to capture complex nonlinear relationships compared to traversal and Bayesian methods. At approximately $60, this open-source IoT platform provides an accessible, practical pathway for students to master advanced ML concepts, promoting deeper conceptual insights and essential technical skills required for the next generation of physicists and engineers.

physics.ed-ph↗

arXiv2026

Effects of uniaxial strain on monolayer transition-metal dichalcogenides revisited

Using hybrid density functional calculations including spin-orbit coupling, we compute the strain evolution of the band structure of monolayer 1H-phase transition-metal dichalcogenides, MX$_2$ (M= Mo, W; X= S, Se, Te), emphasizing an accurate reproduction of the quasiparticle band gap (as opposed to the excitonic optical gap). We show that tensile uniaxial strain applied along either the armchair or zigzag directions leads to a pronounced reduction of the fundamental gap, with the conduction-band edge generally exhibiting the stronger strain response. Both the conduction-band electron valleys (CBM) and the valence-band hole valleys (VBM) remain degenerate under uniaxial strain, while simultaneously drifting away from the high-symmetry $K$ point under strain ("valley drift"), such that the band extrema occur at nearby off-symmetry wave vectors. A minimal tight-binding model rationalizes the valley drift and the unequal electron- and hole-valley drift rates in the presence of strain, leading to indirect band gaps. In particular, for MoS$_2$ the indirectness increases with tensile strain, providing a natural explanation for the experimentally observed decrease in photoluminescence intensity under uniaxial deformation. These results provide quantitative guidance for tailoring band structures for optoelectronic and quantum-defect applications.

cond-mat.mtrl-sci↗

arXiv2025

LightAgent: Production-level Open-source Agentic AI Framework

With the rapid advancement of large language models (LLMs), Multi-agent Systems (MAS) have achieved significant progress in various application scenarios. However, substantial challenges remain in designing versatile, robust, and efficient platforms for agent deployment. To address these limitations, we propose \textbf{LightAgent}, a lightweight yet powerful agentic framework, effectively resolving the trade-off between flexibility and simplicity found in existing frameworks. LightAgent integrates core functionalities such as Memory (mem0), Tools, and Tree of Thought (ToT), while maintaining an extremely lightweight structure. As a fully open-source solution, it seamlessly integrates with mainstream chat platforms, enabling developers to easily build self-learning agents. We have released LightAgent at \href{https://github.com/wxai-space/LightAgent}{https://github.com/wxai-space/LightAgent}

cs.AI↗

arXiv2025

FinGAIA: A Chinese Benchmark for AI Agents in Real-World Financial Domain

The booming development of AI agents presents unprecedented opportunities for automating complex tasks across various domains. However, their multi-step, multi-tool collaboration capabilities in the financial sector remain underexplored. This paper introduces FinGAIA, an end-to-end benchmark designed to evaluate the practical abilities of AI agents in the financial domain. FinGAIA comprises 407 meticulously crafted tasks, spanning seven major financial sub-domains: securities, funds, banking, insurance, futures, trusts, and asset management. These tasks are organized into three hierarchical levels of scenario depth: basic business analysis, asset decision support, and strategic risk management. We evaluated 10 mainstream AI agents in a zero-shot setting. The best-performing agent, ChatGPT, achieved an overall accuracy of 48.9\%, which, while superior to non-professionals, still lags financial experts by over 35 percentage points. Error analysis has revealed five recurring failure patterns: Cross-modal Alignment Deficiency, Financial Terminological Bias, Operational Process Awareness Barrier, among others. These patterns point to crucial directions for future research. Our work provides the first agent benchmark closely related to the financial domain, aiming to objectively assess and promote the development of agents in this crucial field. Partial data is available at https://github.com/SUFE-AIFLM-Lab/FinGAIA.

cs.CL↗

arXiv2025

Symmetry-breaking induced surface magnetization in non-magnetic RuO$_2$

Altermagnetism is a newly identified phase of magnetism distinct from ferromagnetism and antiferromagnetism. RuO$_2$ has been considered a prototypical metallic altermagnet with a critical temperature higher than room temperature. Previous interpretations of the unusual magnetic properties of RuO$_2$ relied on the theoretical prediction that local moments on two Ru sublattices, which are connected by four-fold rotational symmetry, are quite significant (approximately 1 $\mu_B$), leading to long-range antiferromagnetic order. However, accumulated experimental data suggest that local moments on Ru in RuO$_2$ are vanishingly small, indicating that the bulk material is likely non-magnetic. This observation is consistent with the delocalized nature of the 4$d$ electrons of Ru and the strong screening effect in the metallic state. In this work, we show that despite the non-magnetic bulk, the RuO$_2$(110) surface exhibits spontaneous magnetization. We attribute this effect to the breaking of local symmetry, which results in electronic redistribution and magnetic moment enhancement. The emergence of surface magnetism gives rise to interesting spectroscopic phenomena, including spin-polarized surface states, spin-polarized scanning probe microscopy images, and potentially spin-dependent transport effects. These findings highlight the important role of surface magnetic structures in the otherwise non-magnetic bulk RuO$_2$.

cond-mat.mtrl-sci↗

arXiv2023

Fermi-level pinning in ErAs nanoparticles embedded in III-V semiconductors

Embedding rare-earth pnictide (RE-V) nanoparticles into III-V semiconductors enables unique optical, electrical, and thermal properties, with applications in THz photoconductive switches, tunnel junctions, and thermoelectric devices. Despite the high structural quality and control over growth, particle size, and density, the underlying electronic structure of these nanocomposite materials has only been hypothesized. Basic questions about the metallic or semiconducting nature of the nanoparticles (that are typically < 3 nm in diameter) have remained unanswered. Using first-principles calculations, we investigated the structural and electronic properties of ErAs nanoparticles in AlAs, GaAs, InAs, and their alloys. Formation energies of the ErAs nanoparticles with different shapes and sizes (i.e., from cubic to spherical, with 1.14 nm, 1.71 nm, and 2.28 nm diameters) show that spherical nanoparticles are the most energetically favorable. As the diameter increases, the Fermi level is lowered from near the conduction band to the middle of the gap. For the lowest energy nanoparticles, the Fermi level is pinned near the mid-gap, at about 0.8 eV above the valence band in GaAs and about 1.2 eV in AlAs, and it is resonant in the conduction band in InAs. Our results show that the Fermi level is pinned on an absolute energy scale once the band alignment at AlAs/GaAs/InAs interfaces is considered, offering insights into the rational design of these nanocomposite materials.

cond-mat.mtrl-sci↗

arXiv2023

Non-trivial topology in rare-earth monopnictides from dimensionality reduction

Thin films of rare-earth monopnictide semimetals are expected to turn into semiconductors due to quantum confinement effect, which lifts the overlap between electron pockets at Brillouin zone edges and hole pockets at the zone center. Instead, taking non-magnetic LaSb as an example, we find the emergence of a quantum spin Hall insulator phase in LaSb(001) films as the thickness is reduced to 7, 5, or 3 monolayers. This is attributed to a strong quantum confinement effect on the in-plane electron pockets, and the lack of quantum confinement on the out-of-plane pocket in reciprocal space projected onto zone center, leading to a band inversion. Spin-orbit coupling opens a sizeable non-trivial gap in the band structure of the thin films. Such effect is shown to be general in rare-earth monopnictides and may lead to interesting phenomena when coupled with the 4f magnetic moments present in other members of this family of materials.

cond-mat.mtrl-sci↗

arXiv2022

Bi2Se3 Growth on (001) GaAs Substrates for Terahertz Integrated Systems

Terahertz (THz) technologies have been of interest for many years due to the variety of applications including gas sensing, nonionizing imaging of biological systems, security and defense, etc. To date, scientists have used different classes of materials to perform different THz functions. However, to assemble an on-chip THz integrated system, we must understand how to integrate these different materials. Here, we explore the growth of Bi2Se3, a topological insulator (TI) material that could serve as a plasmonic waveguide in THz integrated devices, on technologically-important GaAs (001) substrates. We explore surface treatments and find that atomically smooth GaAs surface is critical to achieving high-quality Bi2Se3 films despite the relatively weak film/substrate interaction. Calculations indicate that the Bi2Se3/GaAs interface is likely selenium-terminated and shows no evidence of chemical bonding between the Bi2Se3 and the substrate. These results are a guide for integrating van der Waals materials with conventional semiconductor substrates and serve as the first steps toward achieving an on-chip THz integrated system.

cond-mat.mtrl-sci↗

arXiv2022

Strong coupling between a topological insulator and a III-V heterostructure at terahertz frequency

We probe theoretically the emergence of strong coupling in a system consisting of a topological insulator (TI) and a III-V heterostructure using a numerical approach based on the scattering matrix formalism. Specifically, we investigate the interactions between terahertz excitations in a structure composed of Bi$_{2}$Se$_{3}$ and GaAs materials. We find that the interaction between the Bi$_{2}$Se$_{3}$ layer and AlGaAs/GaAs quantum wells with intersubband transitions (ISBTs) in the terahertz frequency regime creates new hybrid modes, namely Dirac plasmon-phonon-ISBT polaritons. The formation of these hybrid modes results in anti-crossings (spectral mode splitting) whose magnitude is an indication of the strength of the coupling. By varying the structural parameters of the constituent materials, our numerical calculations reveal that the magnitude of splitting depends strongly on the doping level and the scattering rate in the AlGaAs/GaAs quantum wells, as well as on the thickness of the GaAs spacer layer that separates the quantum-well structure from the TI layer. Our results reveal the material and device parameters required to obtain experimentally-observable signatures of strong coupling. Our model includes the contribution of an extra two-dimensional hole gas (2DHG) that is predicted to arise at the Bi$_{2}$Se$_{3}$/GaAs interface, based on density functional theory (DFT) calculations that explicitly account for details of the atomic terminations at the interface. The presence of this massive 2DHG at the TI/III-V interface shifts the dispersion of the Dirac plasmon-ISBT polaritons to higher frequencies. The damping rate at this interface, in contrast, compensates the effect of the 2DHG. Finally, we observe that the phonon resonances in the TI layer are crucial to the coupling between the THz excitations in the TI and III-V materials.

cond-mat.mtrl-sci↗

arXiv2019

Learning Graph Embedding with Adversarial Training Methods

Graph embedding aims to transfer a graph into vectors to facilitate subsequent graph analytics tasks like link prediction and graph clustering. Most approaches on graph embedding focus on preserving the graph structure or minimizing the reconstruction errors for graph data. They have mostly overlooked the embedding distribution of the latent codes, which unfortunately may lead to inferior representation in many cases. In this paper, we present a novel adversarially regularized framework for graph embedding. By employing the graph convolutional network as an encoder, our framework embeds the topological information and node content into a vector representation, from which a graph decoder is further built to reconstruct the input graph. The adversarial training principle is applied to enforce our latent codes to match a prior Gaussian or Uniform distribution. Based on this framework, we derive two variants of adversarial models, the adversarially regularized graph autoencoder (ARGA) and its variational version, adversarially regularized variational graph autoencoder (ARVGA), to learn the graph embedding effectively. We also exploit other potential variations of ARGA and ARVGA to get a deeper understanding on our designs. Experimental results compared among twelve algorithms for link prediction and twenty algorithms for graph clustering validate our solutions.

cs.LG↗

arXiv2019

Attributed Graph Clustering: A Deep Attentional Embedding Approach

Graph clustering is a fundamental task which discovers communities or groups in networks. Recent studies have mostly focused on developing deep learning approaches to learn a compact graph embedding, upon which classic clustering methods like k-means or spectral clustering algorithms are applied. These two-step frameworks are difficult to manipulate and usually lead to suboptimal performance, mainly because the graph embedding is not goal-directed, i.e., designed for the specific clustering task. In this paper, we propose a goal-directed deep learning approach, Deep Attentional Embedded Graph Clustering (DAEGC for short). Our method focuses on attributed graphs to sufficiently explore the two sides of information in graphs. By employing an attention network to capture the importance of the neighboring nodes to a target node, our DAEGC algorithm encodes the topological structure and node content in a graph to a compact representation, on which an inner product decoder is trained to reconstruct the graph structure. Furthermore, soft labels from the graph embedding itself are generated to supervise a self-training graph clustering process, which iteratively refines the clustering results. The self-training process is jointly learned and optimized with the graph embedding in a unified framework, to mutually benefit both components. Experimental results compared with state-of-the-art algorithms demonstrate the superiority of our method.

cs.LG↗

arXiv2019

Adversarially Regularized Graph Autoencoder for Graph Embedding

Graph embedding is an effective method to represent graph data in a low dimensional space for graph analytics. Most existing embedding algorithms typically focus on preserving the topological structure or minimizing the reconstruction errors of graph data, but they have mostly ignored the data distribution of the latent codes from the graphs, which often results in inferior embedding in real-world graph data. In this paper, we propose a novel adversarial graph embedding framework for graph data. The framework encodes the topological structure and node content in a graph to a compact representation, on which a decoder is trained to reconstruct the graph structure. Furthermore, the latent representation is enforced to match a prior distribution via an adversarial training scheme. To learn a robust embedding, two variants of adversarial approaches, adversarially regularized graph autoencoder (ARGA) and adversarially regularized variational graph autoencoder (ARVGA), are developed. Experimental studies on real-world graphs validate our design and demonstrate that our algorithms outperform baselines by a wide margin in link prediction, graph clustering, and graph visualization tasks.

cs.LG↗

arXiv2018

Universal Network Representation for Heterogeneous Information Networks

Network representation aims to represent the nodes in a network as continuous and compact vectors, and has attracted much attention in recent years due to its ability to capture complex structure relationships inside networks. However, existing network representation methods are commonly designed for homogeneous information networks where all the nodes (entities) of a network are of the same type, e.g., papers in a citation network. In this paper, we propose a universal network representation approach (UNRA), that represents different types of nodes in heterogeneous information networks in a continuous and common vector space. The UNRA is built on our latest mutually updated neural language module, which simultaneously captures inter-relationship among homogeneous nodes and node-content correlation. Relationships between different types of nodes are also assembled and learned in a unified framework. Experiments validate that the UNRA achieves outstanding performance, compared to six other state-of-the-art algorithms, in node representation, node classification, and network visualization. In node classification, the UNRA achieves a 3\% to 132\% performance improvement in terms of accuracy.

cs.SI↗