arXiv subjects

Fan Cui

Publications and source records attributed to Fan Cui.

18 recordsLinked to original sources

arXiv2026

Catalog of Altermagnetism in Magnetic Wallpaper/Space Groups and Nonsymmorphic Altermagnets

Conventional altermagnetism, characterized by compensated collinear spin alignment and spin splitting, exhibits identical spin states at opposite momenta. In this work, we employ a non-spatial global symmetry $S$, the spinless time-reversal symmetry, which effectively replaces inversion symmetry in preserving the spin-state equivalence; hence, we systematically extend the classification of altermagnetism to all possible non-centrosymmetric crystals. By analyzing the necessary symmetry conditions, we provide a complete catalog of altermagnetic orders for all 2D magnetic wallpaper groups and all 3D magnetic space groups, identifying 17 altermagnetic wallpaper groups (12 centrosymmetric and 5 non-centrosymmetric) and 422 altermagnetic space groups (160 centrosymmetric and 262 non-centrosymmetric). This catalog assigns each altermagnetic wallpaper and space group to one of the six altermagnetic wave types established in the literature and presents its distinct spin distribution in the Brillouin zone (BZ); notably, the low-energy wave-type description does not necessarily extend throughout the full BZ, since the spin-degenerate nodal lines and planes can be unpinned from the high-symmetry planes. Beyond the catalog, nonsymmorphic symmetries further bring new patterns of the altermagnetic BZs through the emergence of hourglass dispersions, which arise from the compatibility relations between two symmetry-protected degenerate manifolds: same-spin and opposite-spin degeneracies. In both the non-centrosymmetric altermagnetism and the emergence of the hourglass dispersion, the spinless time-reversal symmetry plays the key role. Our work extends the symmetry catalog of altermagnetism and reveals that nonsymmorphic symmetries are essential for realizing altermagnetic band structures beyond the six established wave types, such as an $i$-wave-like spin winding in a tetragonal BZ.

cond-mat.mtrl-sci↗

arXiv2026

Clover: A Neural-Symbolic Agentic Harness with Stochastic Tree-of-Thoughts for Verified RTL Repair

RTL program repair remains a critical bottleneck in hardware design and verification. Traditional automatic program repair (APR) methods rely on predefined templates and synthesis, limiting their bug coverage. Large language models (LLMs) and coding agents based on them offer flexibility but suffer from randomness and context corruption when handling long RTL code and waveforms. We present Clover, a neural-symbolic agentic harness that orchestrates RTL repair as a structured search over code manipulations to explore a validated solution for the bug. Recognizing that different repair operations favor distinct strategies, Clover dynamically dispatches tasks to specialized LLM agents or symbolic solvers. At its core, Clover introduces stochastic tree-of-thoughts, a test-time scaling mechanism that manages the main agent's context as a search tree, balancing exploration and exploitation for reliable outcomes. An RTL-specific toolbox further empowers agents to interact with the debugging environment. Evaluated on the RTL-repair benchmark, Clover fixes 96.8% of bugs within a fixed time limit, covering 94% and 63% more bugs than both pure traditional and LLM-based baselines, respectively, while achieving an average pass@1 rate of 87.5%, demonstrating high reliability and effectiveness.

cs.AR↗

arXiv2026

HWE-Bench: Benchmarking LLM Agents on Real-World Hardware Bug Repair Tasks

Existing benchmarks for hardware design primarily evaluate Large Language Models (LLMs) on isolated, component-level tasks such as generating HDL modules from specifications, leaving repository-scale evaluation unaddressed. We introduce HWE-Bench, the first large-scale, repository-level benchmark for evaluating LLM agents on real-world hardware bug repair tasks. HWE-Bench comprises 417 task instances derived from real historical bug-fix pull requests across six major open-source projects spanning both Verilog/SystemVerilog and Chisel, covering RISC-V cores, SoCs, and security roots-of-trust. Each task is grounded in a fully containerized environment where the agent must resolve a real bug report, with correctness validated through the project's native simulation and regression flows. The benchmark is built through a largely automated pipeline that enables efficient expansion to new repositories. We evaluate seven LLMs with four agent frameworks and find that the best agent resolves 70.7% of tasks overall, with performance exceeding 90% on smaller cores but dropping below 65% on complex SoC-level projects. We observe larger performance gaps across models than commonly reported on software benchmarks, and difficulty is driven by project scope and bug-type distribution rather than code size alone. Our failure analysis traces agent failures to three stages of the debugging process: fault localization, hardware-semantic reasoning, and cross-artifact coordination across RTL, configuration, and verification components, providing concrete directions for developing more capable hardware-aware agents.

cs.AI↗

arXiv2025

R3A: Reliable RTL Repair Framework with Multi-Agent Fault Localization and Stochastic Tree-of-Thoughts Patch Generation

Repairing RTL bugs is crucial for hardware design and verification. Traditional automatic program repair (APR) methods define dedicated search spaces to locate and fix bugs with program synthesis. However, they heavily rely on fixed templates and can only deal with limited bugs. As an alternative, Large Language Models with the ability to understand code semantics can be explored for RTL repair. However, they suffer from unreliable outcomes due to inherent randomness and long input contexts of RTL code and waveform. To address these challenges, we propose R3A, an LLM-based automatic RTL program repair framework upon the basic model to improve reliability. R3A proposes the stochastic Tree-Of-Thoughts method to control a patch generation agent to explore a validated solution for the bug. The algorithm samples search states according to a heuristic function to balance between exploration and exploitation for a reliable outcome. Besides, R3A proposes a multi-agent fault localization method to find fault candidates as the starting points for the patch generation agent, further increasing the reliability. Experiments show R3A can fix 90.6% of bugs in the RTL-repair dataset within a given time limit, which covers 45% more bugs than traditional methods and other LLM-based approaches, while achieving an 86.7% pass@5 rate on average, showing a high reliability.

cs.AR↗

arXiv2025

PHYBench: Holistic Evaluation of Physical Perception and Reasoning in Large Language Models

Current benchmarks for evaluating the reasoning capabilities of Large Language Models (LLMs) face significant limitations: task oversimplification, data contamination, and flawed evaluation items. These deficiencies necessitate more rigorous assessment methods. To address these limitations, we introduce PHYBench, a benchmark of 500 original physics problems ranging from high school to Physics Olympiad difficulty. PHYBench addresses data contamination through original content and employs a systematic curation pipeline to eliminate flawed items. Evaluations show that PHYBench activates more tokens and provides stronger differentiation between reasoning models compared to other baselines like AIME 2024, OlympiadBench and GPQA. Even the best-performing model, Gemini 2.5 Pro, achieves only 36.9% accuracy compared to human experts' 61.9%. To further enhance evaluation precision, we introduce the Expression Edit Distance (EED) Score for mathematical expression assessment, which improves sample efficiency by 204% over binary scoring. Moreover, PHYBench effectively elicits multi-step and multi-condition reasoning, providing a platform for examining models' reasoning robustness, preferences, and deficiencies. The benchmark results and dataset are publicly available at https://www.phybench.cn/.

cs.CL↗

arXiv2024

OriGen:Enhancing RTL Code Generation with Code-to-Code Augmentation and Self-Reflection

Recent studies have demonstrated the significant potential of Large Language Models (LLMs) in generating Register Transfer Level (RTL) code, with notable advancements showcased by commercial models such as GPT-4 and Claude3-Opus. However, these proprietary LLMs often raise concerns regarding privacy and security. While open-source LLMs offer solutions to these concerns, they typically underperform commercial models in RTL code generation tasks, primarily due to the scarcity of high-quality open-source RTL datasets. To address this challenge, we introduce OriGen , a fully open-source framework that incorporates self-reflection capabilities and a novel dataset augmentation methodology for generating high-quality, large-scale RTL code. Our approach employs a code-tocode augmentation technique to enhance the quality of open-source RTL code datasets. Furthermore, OriGen can rectify syntactic errors through a self-reflection process that leverages compiler feedback. Experimental results demonstrate that OriGen significantly outperforms other open-source alternatives in RTL code generation. It surpasses the previous best-performing open-source LLM by 12.8% and even exceeds GPT-4 Turbo in the pass@1 metric on the VerilogEval-Human benchmark. Moreover, OriGen exhibits superior capabilities in self-reflection and error correction, outperforming GPT-4 by 19.9% on a benchmark designed to evaluate self-reflection capabilities.

cs.AR↗

arXiv2023

Classification of High-Ordered Topological Nodes towards Moir\'e Flat Bands in Twisted Bilayers

At magic twisted angles, Dirac cones in twisted bilayer graphene (TBG) can evolve into flat bands, serving as a critical playground for the study of strongly correlated physics. When chiral symmetry is introduced, rigorous mathematical proof confirms that the flat bands are locked at zero energy in the entire Moir\'e Brillouin zone (BZ). Yet, TBG is not the sole platform that exhibits this absolute band flatness. Central to this flatness phenomenon are topological nodes and their specific locations in the BZ. In this study, considering twisted bilayer systems that preserve chiral symmetry, we classify various ordered topological nodes in base layers and all possible node locations across different BZs. Specifically, we constrain the node locations to rotational centers, such as {\Gamma} and M points, to ensure the interlayer coupling retains equal strength in all directions. Using this classification as a foundation, we systematically identify the conditions under which Moir\'e flat bands emerge. Additionally, through the extension of holomorphic functions, we provide proof that flat bands are locked at zero energy, shedding light on the origin of the band flatness. Remarkably, beyond Dirac cones, numerous twisted bilayer nodal platforms can host flat bands with a degeneracy number of more than two, such as four-fold, six-fold, and eight-fold. This multiplicity of degeneracy in flat bands might unveil more complex and enriched correlation physics.

cond-mat.str-el↗

arXiv2023

Relate auditory speech to EEG by shallow-deep attention-based network

Electroencephalography (EEG) plays a vital role in detecting how brain responses to different stimulus. In this paper, we propose a novel Shallow-Deep Attention-based Network (SDANet) to classify the correct auditory stimulus evoking the EEG signal. It adopts the Attention-based Correlation Module (ACM) to discover the connection between auditory speech and EEG from global aspect, and the Shallow-Deep Similarity Classification Module (SDSCM) to decide the classification result via the embeddings learned from the shallow and deep layers. Moreover, various training strategies and data augmentation are used to boost the model robustness. Experiments are conducted on the dataset provided by Auditory EEG challenge (ICASSP Signal Processing Grand Challenge 2023). Results show that the proposed model has a significant gain over the baseline on the match-mismatch track.

cs.SD↗

arXiv2023

Exploring Representation Learning for Small-Footprint Keyword Spotting

In this paper, we investigate representation learning for low-resource keyword spotting (KWS). The main challenges of KWS are limited labeled data and limited available device resources. To address those challenges, we explore representation learning for KWS by self-supervised contrastive learning and self-training with pretrained model. First, local-global contrastive siamese networks (LGCSiam) are designed to learn similar utterance-level representations for similar audio samplers by proposed local-global contrastive loss without requiring ground-truth. Second, a self-supervised pretrained Wav2Vec 2.0 model is applied as a constraint module (WVC) to force the KWS model to learn frame-level acoustic representations. By the LGCSiam and WVC modules, the proposed small-footprint KWS model can be pretrained with unlabeled data. Experiments on speech commands dataset show that the self-training WVC module and the self-supervised LGCSiam module significantly improve accuracy, especially in the case of training on a small labeled dataset.

cs.SD↗

arXiv2023

Improving Weakly Supervised Sound Event Detection with Causal Intervention

Existing weakly supervised sound event detection (WSSED) work has not explored both types of co-occurrences simultaneously, i.e., some sound events often co-occur, and their occurrences are usually accompanied by specific background sounds, so they would be inevitably entangled, causing misclassification and biased localization results with only clip-level supervision. To tackle this issue, we first establish a structural causal model (SCM) to reveal that the context is the main cause of co-occurrence confounders that mislead the model to learn spurious correlations between frames and clip-level labels. Based on the causal analysis, we propose a causal intervention (CI) method for WSSED to remove the negative impact of co-occurrence confounders by iteratively accumulating every possible context of each class and then re-projecting the contexts to the frame-level features for making the event boundary clearer. Experiments show that our method effectively improves the performance on multiple datasets and can generalize to various baseline models.

cs.SD↗

arXiv2022

Predicting Multi-Codebook Vector Quantization Indexes for Knowledge Distillation

Knowledge distillation(KD) is a common approach to improve model performance in automatic speech recognition (ASR), where a student model is trained to imitate the output behaviour of a teacher model. However, traditional KD methods suffer from teacher label storage issue, especially when the training corpora are large. Although on-the-fly teacher label generation tackles this issue, the training speed is significantly slower as the teacher model has to be evaluated every batch. In this paper, we reformulate the generation of teacher label as a codec problem. We propose a novel Multi-codebook Vector Quantization (MVQ) approach that compresses teacher embeddings to codebook indexes (CI). Based on this, a KD training framework (MVQ-KD) is proposed where a student model predicts the CI generated from the embeddings of a self-supervised pre-trained teacher model. Experiments on the LibriSpeech clean-100 hour show that MVQ-KD framework achieves comparable performance as traditional KD methods (l1, l2), while requiring 256 times less storage. When the full LibriSpeech dataset is used, MVQ-KD framework results in 13.8% and 8.2% relative word error rate reductions (WERRs) for non -streaming transducer on test-clean and test-other and 4.0% and 4.9% for streaming transducer. The implementation of this work is already released as a part of the open-source project icefall.

eess.AS↗

arXiv2021

Detect what you want: Target Sound Detection

Human beings can perceive a target sound type from a multi-source mixture signal by the selective auditory attention, however, such functionality was hardly ever explored in machine hearing. This paper addresses the target sound detection (TSD) task, which aims to detect the target sound signal from a mixture audio when a target sound's reference audio is given. We present a novel target sound detection network (TSDNet) which consists of two main parts: A conditional network which aims at generating a sound-discriminative conditional embedding vector representing the target sound, and a detection network which takes both the mixture audio and the conditional embedding vector as inputs and produces the detection result of the target sound. These two networks can be jointly optimized with a multi-task learning approach to further improve the performance. In addition, we study both strong-supervised and weakly-supervised strategies to train TSDNet and propose a data augmentation method by mixing two samples. To facilitate this research, we build a target sound detection dataset (\textit{i.e.} URBAN-TSD) based on URBAN-SED and UrbanSound8K datasets, and experimental results indicate our method could get the segment-based F scores of 76.3$\%$ and 56.8$\%$ on the strongly-labelled and weakly-labelled data respectively.

cs.SD↗

arXiv2021

Comprehensive view of microscopic interactions between DNA-coated colloids

The self-assembly of DNA-coated colloids into highly-ordered structures offers great promise for advanced optical materials. However, control of disorder, defects, melting, and crystal growth is hindered by the lack of a microscopic understanding of DNA-mediated colloidal interactions. Here we use total internal reflection microscopy to measure in situ the interaction potential between DNA-coated colloids with nanometer resolution and the macroscopic melting behavior. The range and strength of the interaction are measured and linked to key material design parameters, including DNA sequence, polymer length, grafting density, and complementary fraction. We present a first-principles model that quantitatively reproduces our experimental data without fitting parameters over a wide range of DNA ligand designs. Our theory identifies a subtle competition between DNA binding and steric repulsion and accurately predicts adhesion and melting at a molecular level. Combining experimental and theoretical results, our work provides a quantitative and predictive approach for guiding material design with DNA-nanotechnology and can be further extended to a diversity of colloidal and biological systems.

cond-mat.soft↗

arXiv2021

Effect of Photon Counting Shot Noise on Total Internal Reflection Microscopy

Total internal reflection microscopy (TIRM) measures changes in the distance between a colloidal particle and a transparent substrate by measuring the intensity of light scattered by the particle when it is illuminated by the evanescent field that is created from light totally internally reflected at the substrate interface. From these measurements, the height-dependent effective potential $\varphi(z)$ between the colloidal particle and the substrate can be measured. The spatial resolution with which TIRM can resolve the height $z$ and effective potential $\varphi(z)$ is limited by the intrinsic shot noise of the photon counting process used to measure the scattered light intensity. We develop a model to determine the spatial resolution with which TIRM can measure $\varphi(z)$ and verify its validity with simulations and experiments. We further establish the critical role of photon-counting statistics and the intensity integration time $\tau$ in TIRM measurements, which is a trade-off between narrowing the width of the photon counting distribution and capturing the instantaneous position of the probe particle.

cond-mat.soft↗

arXiv2021

Generalized Fermion Doubling Theorems: Classification of 2D Nodal Systems in Terms of Wallpaper Groups

The Nielsen-Ninomiya Theorem has set up a ground rule for the minimal number of the topological points in a Brillouin zone. Notably, in the 2D Brillouin zone, chiral symmetry and space-time inversion symmetry can properly define topological invariants as charges characterizing the stability of the nodal points so that the non-zero charges protect these points. Due to the charge neutralization, the Nielsen-Ninomiya Theorem requires at least two stable topological points in the entire Brillouin zone. However, additional crystalline symmetries might duplicate the points. In this regard, for the wallpaper groups with crystalline symmetries, the minimal number of the nodal points in the Brillouin zone might be more than two. In this work, we determine the minimal numbers of the nodal points for the wallpaper groups in chiral-symmetric and space-time-inversion-symmetric systems separately and provide examples for new topological materials, such as topological nodal time-reversal-symmetric superconductors and Dirac semimetals. This generalized Nielsen-Ninomiya Theorem serves as a guide to search for 2D topological nodal materials and new platforms for twistronics. Furthermore, we show the Nielsen-Ninomiya Theorem can be extended to 2D non-Hermitian systems hosting topologically protected exceptional points and Fermi points for the 17 wallpaper groups and use the violation of the theorem on the surface to classify 3D Hermitian and non-Hermitian topological bulks.

cond-mat.mes-hall↗

arXiv2020

Robust Protection of III-V Nanowires in Water Splitting by a Thin Compact TiO$_2$ Layer

Narrow-bandgap III-V semiconductor nanowires (NWs) with a suitable band structure and strong light-trapping ability are ideal for high-efficiency low-cost solar water-splitting systems. However, due to their nanoscale dimension, they suffer more severe corrosion by the electrolyte solution than the thin-film counterparts. Thus, short-term durability is the major obstacle for using these NWs for practical water splitting applications. Here, we demonstrated for the first time that a thin layer (~7 nm thick) of compact TiO$_2$ deposited by atomic layer deposition can provide robust protection to III-V NWs. The protected GaAs NWs maintain 91.4% of its photoluminescence intensity after 14 months of storage in ambient atmosphere, which suggests the TiO$_2$ layer is pinhole-free. Working as a photocathode for water splitting, they exhibited a 45% larger photocurrent density compared with un-protected counterparts and a high Faraday efficiency of 91%, and can also maintain a record-long highly-stable performance among narrow-bandgap III-V NW photoelectrodes; after 67 hours photoelectrochemical stability test reaction in strong acid electrolyte solution (pH = 1), they show no apparent indication of corrosion, which is in stark contrast to the un-protected NWs that are fully failed after 35-hours. These findings provide an effective way to enhance both stability and performance of III-V NW based photoelectrodes, which are highly important for practical applications in solar-energy-based water splitting systems.

cond-mat.mes-hall↗

arXiv2020

Reemergence of superconductivity in pressurized quasi-one-dimensional superconductor K2Mo3As3

Here we report a pressure-induced reemergence of superconductivity in recently discovered superconductor K2Mo3As3, which is the first experimental case observed in quasi-one-dimensional superconductors. We find that, after full suppression of the ambient-pressure superconducting (SC-I) state at 8.7 GPa, an intermediary non-superconducting state sets in and prevails to the pressure up to 18.2 GPa, however, above this pressure a new superconducting (SC-II) state appears unexpectedly. High pressure x-ray diffraction measurements demonstrate that the pressure-induced dramatic change of the lattice parameter c contributes mainly to the emergence of the SC-II state. Combined with the theioretical calculations on band strcture, our results suggest that the reemergemce of superconductivity is associated with the change of the complicated interplay among different orbital electrons, driven by the pressure-induced unisotropic change of the lattice.

cond-mat.supr-con↗

arXiv2016

Cross-model convolutional neural network for multiple modality data representation

A novel data representation method of convolutional neural net- work (CNN) is proposed in this paper to represent data of different modalities. We learn a CNN model for the data of each modality to map the data of differ- ent modalities to a common space, and regularize the new representations in the common space by a cross-model relevance matrix. We further impose that the class label of data points can also be predicted from the CNN representa- tions in the common space. The learning problem is modeled as a minimiza- tion problem, which is solved by an augmented Lagrange method (ALM) with updating rules of Alternating direction method of multipliers (ADMM). The experiments over benchmark of sequence data of multiple modalities show its advantage.

cs.LG↗