arXiv subjects

Kai Liu

Publications and source records attributed to Kai Liu.

At least 145 records · Page 8Linked to original sources

arXiv2025

Model Accuracy and Data Heterogeneity Shape Uncertainty Quantification in Machine Learning Interatomic Potentials

Machine learning interatomic potentials (MLIPs) enable accurate atomistic modelling, but reliable uncertainty quantification (UQ) remains elusive. In this study, we investigate two UQ strategies, ensemble learning and D-optimality, within the atomic cluster expansion framework. It is revealed that higher model accuracy strengthens the correlation between predicted uncertainties and actual errors and improves novelty detection, with D-optimality yielding more conservative estimates. Both methods deliver well calibrated uncertainties on homogeneous training sets, yet they underpredict errors and exhibit reduced novelty sensitivity on heterogeneous datasets. To address this limitation, we introduce clustering-enhanced local D-optimality, which partitions configuration space into clusters during training and applies D-optimality within each cluster. This approach substantially improves the detection of novel atomic environments in heterogeneous datasets. Our findings clarify the roles of model fidelity and data heterogeneity in UQ performance and provide a practical route to robust active learning and adaptive sampling strategies for MLIP development.

cond-mat.mtrl-sci↗

arXiv2025

Intrinsic layer polarization and multi-flatband transport in non-centrosymmetric mixed-stacked multilayer graphene

Graphene multilayers exhibit electronic spectra that depend sensitively on both the number of layers and their stacking order. Beyond trilayer graphene, mixed stacking sequences (alternating Bernal and rhombohedral layers) give rise to multiple coexisting low-energy bands. Here we investigate ABCBC-stacked pentalayer graphene, a less-studied non-centrosymmetric mixed sequence. This stacking can be regarded as an ABC (rhombohedral) trilayer on top of an AB (Bernal) bilayer, so its low-energy band structure contains both a cubic band and a parabolic band that hybridize. In transport measurements, we observe an intrinsic band gap at charge neutrality whose magnitude changes asymmetrically under an applied perpendicular displacement field. This behavior reflects the spontaneous layer polarization inherent to the broken inversion symmetry and mirror symmetry. By tuning the displacement field and carrier density, we drive multiple Lifshitz transitions in the Fermi surface topology and realize Landau levels with different degeneracies arising from the multi-flatband system. Remarkably, a v = -6 quantum Hall state emerges at an exceptionally low magnetic field (~20 mT), indicating the interplay between spontaneous symmetry breaking and Berry curvatures. Our results establish mixed-stacked multilayer graphene as a tunable platform with various broken symmetries and multiple flatbands, suitable for exploring emergent correlated electronic states.

cond-mat.mes-hall↗

arXiv2025

Sliding two-dimensional superconductivity and charge-density-wave state in a bulk crystal

Superconductivity in the two-dimensional (2D) limit is a fertile ground for exotic quantum phenomena-many of which remain elusive in their 3D counterparts. While studies of 2D superconductivity have predominantly focused on mono- or few-layer systems, we demonstrate an alternative route-interlayer sliding in bulk crystals. Through a precisely controlled growth strategy, we engineer interlayer sliding in bulk 3R-NbSe2, deliberately disrupting [001] mirror symmetry and drastically suppressing interlayer coupling. Remarkably, this structural manipulation stabilizes Ising-type superconductivity coexisting with an unconventional charge-density-wave (CDW) state akin to that of monolayer 2H-NbSe2. The sliding phase exhibits a pronounced suppression of the upper critical field at low temperatures, revealing a delicate competition between Ising and Rashba spin-orbit coupling (SOC) in the globally noncentrosymmetric lattice. Intriguingly, the superconducting state displays two-fold symmetry, a signature that may arise from asymmetric SOC or a multi-component pairing order parameter. Our work establishes interlayer sliding as a symmetry-breaking tool to promote 2D superconductivity in bulk materials-without resorting to extrinsic intercalation or doping. More broadly, this approach sets a paradigm for unlocking hidden quantum states in layered materials, offering a new dimension in design of quantum matter.

cond-mat.supr-con↗

arXiv2025

Cascade of Even-Denominator Fractional Quantum Hall States in Mixed-Stacked Multilayer Graphene

The fractional quantum Hall effect (FQHE), particularly at half-filling of Landau levels, provides a unique window into topological phases hosting non-Abelian excitations. However, experimental platforms simultaneously offering large energy gaps, delicate tunability, and robust non-Abelian signatures remain scarce. Here, we report the observation of a cascade of even-denominator FQH states at filling factors $ν$ = ${-5/2}$, ${-7/2}$, ${-9/2}$, ${-11/2}$, and ${-13/2}$, alongside numerous odd-denominator states in mixed-stacked pentalayer graphene, a previously unexplored system characterized by intertwined quadratic and cubic band dispersions. These even-denominator states, representing the highest filling half-filled states reported so far in the zeroth Landau level (ZLL), emerge from two distinct intra-ZLL and exhibit unprecedented displacement field tunability driven by LL crossings in the hybridized multiband structure. At half fillings, continuous quasiparticle phase transitions between paired FQH states, magnetic Bloch states, and composite Fermi liquids are clearly identified upon tuning external fields. Numerical calculations, revealing characteristic sixfold ground-state degeneracy and chiral graviton spectral analysis, suggest the observed even-denominator FQH states belong to the non-Abelian Moore-Read type. These results establish mixed-stacked multilayer graphene as a rich and versatile crystalline platform for exploring tunable correlated topological phases.

cond-mat.mes-hall↗

arXiv2025

Heterogeneous Ensemble Enables a Universal Uncertainty Metric for Atomistic Foundation Models

Universal machine learning interatomic potentials (uMLIPs) are reshaping atomistic simulation as foundation models, delivering near \textit{ab initio} accuracy at a fraction of the cost. Yet the lack of reliable, general uncertainty quantification limits their safe, wide-scale use. Here we introduce a unified, scalable uncertainty metric $U$ based on a heterogeneous model ensemble with reuse of pretrained uMLIPs. Across chemically and structurally diverse datasets, $U$ shows a strong correlation with the true prediction errors and provides a robust ranking of configuration-level risk. Leveraging this metric, we propose an uncertainty-aware model distillation framework to produce system-specific potentials: for W, an accuracy comparable to full-DFT training is achieved using only $4\%$ of the DFT labels; for MoNbTaW, no additional DFT calculations are required. Notably, by filtering numerical label noise, the distilled models can, in some cases, surpass the accuracy of the DFT reference labels. The uncertainty-aware approach offers a practical monitor of uMLIP reliability in deployment, and guides data selection and fine-tuning strategies, thereby advancing the construction and safe use of foundation models and enabling cost-efficient development of accurate, system-specific potentials.

cond-mat.mtrl-sci↗

arXiv2025

Smooth Reading: Bridging the Gap of Recurrent LLM to Self-Attention LLM on Long-Context Tasks

Recently, recurrent large language models (Recurrent LLMs) with linear computational complexity have re-emerged as efficient alternatives to self-attention-based LLMs (Self-Attention LLMs), which have quadratic complexity. However, Recurrent LLMs often underperform on long-context tasks due to their limited fixed-size memory. Previous research has primarily focused on enhancing the memory capacity of Recurrent LLMs through architectural innovations, but these approaches have not yet enabled Recurrent LLMs to match the performance of Self-Attention LLMs on long-context tasks. We argue that this limitation arises because processing the entire context at once is not well-suited for Recurrent LLMs. In this paper, we propose Smooth Reading, a chunk-wise inference method inspired by human reading strategies. Smooth Reading processes context in chunks and iteratively summarizes the contextual information, thereby reducing memory demands and making the approach more compatible with Recurrent LLMs. Our experimental results show that this method substantially narrows the performance gap between Recurrent and Self-Attention LLMs on long-context tasks, while preserving the efficiency advantages of Recurrent LLMs. Our Smooth Reading boosts SWA-3B-4k (a Recurrent LLM) from 5.68% lower to 3.61% higher performance than Self-Attention LLMs on LongBench. Besides, our method maintains the high efficiency, training 3x faster and inferring 2x faster at 64k context compared to Self-Attention LLMs. To our knowledge, this is the first work to achieve comparable performance using Recurrent LLMs compared with Self-Attention LLMs on long-context tasks. We hope our method will inspire future research in this area. To facilitate further progress, we will release code and dataset.

cs.CL↗

arXiv2025

Origin of insulating-like behavior of Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$ under pressure: A first-principles study

Recent experimental study on Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$ superconductors has revealed an unexpected quantum phase transition from superconducting state to insulatinglike state under pressure [Zhou et al., Nat. Phys. 18, 406 (2022)]. To better understand the physical origin of this pressure-induced phenomenon, here we have studied the structural, electronic, and magnetic properties of undoped and O-doped Bi$_2$Sr$_2$CaCu$_2$O$_{8+x}$ (Bi2212) under pressures based on density-functional theory calculations. We first identified the crystal structure of undoped Bi2212 with the armchair distortions in the BiO layers and reproduced the insulating feature of the parent compound. Then we added an extra O atom to the parent compound to simulate the hole-doping effect and found that the structure with O dopant located in the van der Waals (vdW) gap is energetically the most stable. Further calculations on O-doped (0.125 holes/Cu) Bi2212 revealed that the pressure can induce charge redistributions between CuO$_2$ planes and BiO layers; specifically, Cu-$d_{x^2-y^2}$ orbitals gain electrons and Cu atoms rather than O atoms dominate around the Fermi level under high pressure. Along with the increasing pressure, the density of states at the Fermi level first reaches the maximum at $\sim$ 10 GPa and then shows a valley near the Fermi level above 20 GPa, which may be responsible for the insulatinglike state observed in recent experiments. We suggest that the competition among several factors, such as the increase of electrons in the CuO$_2$ plane, the variation of in-plane hopping due to the shortened Cu-O distance, and the enhanced Coulomb repulsion among the Cu-3$d$ electrons, could lead to the exotic transition under pressure. Our work provides an explanation of the high-pressure behaviors of Bi2212, which may facilitate a comprehensive understanding of cuprate superconductors.

cond-mat.supr-con↗

arXiv2025

MCAM: Multimodal Causal Analysis Model for Ego-Vehicle-Level Driving Video Understanding

Accurate driving behavior recognition and reasoning are critical for autonomous driving video understanding. However, existing methods often tend to dig out the shallow causal, fail to address spurious correlations across modalities, and ignore the ego-vehicle level causality modeling. To overcome these limitations, we propose a novel Multimodal Causal Analysis Model (MCAM) that constructs latent causal structures between visual and language modalities. Firstly, we design a multi-level feature extractor to capture long-range dependencies. Secondly, we design a causal analysis module that dynamically models driving scenarios using a directed acyclic graph (DAG) of driving states. Thirdly, we utilize a vision-language transformer to align critical visual features with their corresponding linguistic expressions. Extensive experiments on the BDD-X, and CoVLA datasets demonstrate that MCAM achieves SOTA performance in visual-language causal relationship learning. Furthermore, the model exhibits superior capability in capturing causal characteristics within video sequences, showcasing its effectiveness for autonomous driving applications. The code is available at https://github.com/SixCorePeach/MCAM.

cs.CV↗

arXiv2025

Operator-based machine learning framework for generalizable prediction of unsteady treatment dynamics in stormwater infrastructure

Stormwater infrastructures are decentralized urban water-management systems that face highly unsteady hydraulic and pollutant loadings from episodic rainfall-runoff events. Accurately evaluating their in-situ treatment performance is essential for cost-effective design and planning. Traditional lumped dynamic models (e.g., continuously stirred tank reactor, CSTR) are computationally efficient but oversimplify transport and reaction processes, limiting predictive accuracy and insight. Computational fluid dynamics (CFD) resolves detailed turbulent transport and pollutant fate physics but incurs prohibitive computational cost for unsteady and long-term simulations. To address these limitations, this study develops a composite operator-based neural network (CPNN) framework that leverages state-of-the-art operator learning to predict the spatial and temporal dynamics of hydraulics and particulate matter (PM) in stormwater treatment. The framework is demonstrated on a hydrodynamic separator (HS), a common urban treatment device. Results indicate that the CPNN achieves R2 > 0.8 for hydraulic predictions in 95.2% of test cases; for PM concentration predictions, R2 > 0.8 in 72.6% of cases and 0.4 < R2 < 0.8 in 22.6%. The analysis identifies challenges in capturing dynamics under extreme low-flow conditions, owing to their lower contribution to the training loss. Exploiting the automatic-differentiation capability of the CPNN, sensitivity analyses quantify the influence of storm event loading on PM transport. Finally, the potential of the CPNN framework for continuous, long-term evaluation of stormwater infrastructure performance is discussed, marking a step toward robust, climate-aware planning and implementation.

cs.CE↗

arXiv2025

Hunyuan-TurboS: Advancing Large Language Models through Mamba-Transformer Synergy and Adaptive Chain-of-Thought

As Large Language Models (LLMs) rapidly advance, we introduce Hunyuan-TurboS, a novel large hybrid Transformer-Mamba Mixture of Experts (MoE) model. It synergistically combines Mamba's long-sequence processing efficiency with Transformer's superior contextual understanding. Hunyuan-TurboS features an adaptive long-short chain-of-thought (CoT) mechanism, dynamically switching between rapid responses for simple queries and deep "thinking" modes for complex problems, optimizing computational resources. Architecturally, this 56B activated (560B total) parameter model employs 128 layers (Mamba2, Attention, FFN) with an innovative AMF/MF block pattern. Faster Mamba2 ensures linear complexity, Grouped-Query Attention minimizes KV cache, and FFNs use an MoE structure. Pre-trained on 16T high-quality tokens, it supports a 256K context length and is the first industry-deployed large-scale Mamba model. Our comprehensive post-training strategy enhances capabilities via Supervised Fine-Tuning (3M instructions), a novel Adaptive Long-short CoT Fusion method, Multi-round Deliberation Learning for iterative improvement, and a two-stage Large-scale Reinforcement Learning process targeting STEM and general instruction-following. Evaluations show strong performance: overall top 7 rank on LMSYS Chatbot Arena with a score of 1356, outperforming leading models like Gemini-2.0-Flash-001 (1352) and o4-mini-2025-04-16 (1345). TurboS also achieves an average of 77.9% across 23 automated benchmarks. Hunyuan-TurboS balances high performance and efficiency, offering substantial capabilities at lower inference costs than many reasoning models, establishing a new paradigm for efficient large-scale pre-trained models.

cs.CL↗

arXiv2025

Robust 6DoF Pose Tracking Considering Contour and Interior Correspondence Uncertainty for AR Assembly Guidance

Augmented reality assembly guidance is essential for intelligent manufacturing and medical applications, requiring continuous measurement of the 6DoF poses of manipulated objects. Although current tracking methods have made significant advancements in accuracy and efficiency, they still face challenges in robustness when dealing with cluttered backgrounds, rotationally symmetric objects, and noisy sequences. In this paper, we first propose a robust contour-based pose tracking method that addresses error-prone contour correspondences and improves noise tolerance. It utilizes a fan-shaped search strategy to refine correspondences and models local contour shape and noise uncertainty as mixed probability distribution, resulting in a highly robust contour energy function. Secondly, we introduce a CPU-only strategy to better track rotationally symmetric objects and assist the contour-based method in overcoming local minima by exploring sparse interior correspondences. This is achieved by pre-sampling interior points from sparse viewpoint templates offline and using the DIS optical flow algorithm to compute their correspondences during tracking. Finally, we formulate a unified energy function to fuse contour and interior information, which is solvable using a re-weighted least squares algorithm. Experiments on public datasets and real scenarios demonstrate that our method significantly outperforms state-of-the-art monocular tracking methods and can achieve more than 100 FPS using only a CPU.

cs.CV↗

arXiv2025

UAV-DETR: Efficient End-to-End Object Detection for Unmanned Aerial Vehicle Imagery

Unmanned aerial vehicle object detection (UAV-OD) has been widely used in various scenarios. However, most existing UAV-OD algorithms rely on manually designed components, which require extensive tuning. End-to-end models that do not depend on such manually designed components are mainly designed for natural images, which are less effective for UAV imagery. To address such challenges, this paper proposes an efficient detection transformer (DETR) framework tailored for UAV imagery, i.e., UAV-DETR. The framework includes a multi-scale feature fusion with frequency enhancement module, which captures both spatial and frequency information at different scales. In addition, a frequency-focused down-sampling module is presented to retain critical spatial details during down-sampling. A semantic alignment and calibration module is developed to align and fuse features from different fusion paths. Experimental results demonstrate the effectiveness and generalization of our approach across various UAV imagery datasets. On the VisDrone dataset, our method improves AP by 3.1\% and $\text{AP}_{50}$ by 4.2\% over the baseline. Similar enhancements are observed on the UAVVaste dataset. The project page: https://github.com/ValiantDiligent/UAV-DETR

cs.CV↗

arXiv2025

Beyond Fixed Length: Bucket Pre-training is All You Need

Large Language Models (LLMs) have demonstrated exceptional performance across various tasks, with pre-training stage serving as the cornerstone of their capabilities. However, the conventional fixed-length data composition strategy for pre-training presents several practical challenges. When using shorter sequences, documents are often truncated, potentially leading to information loss and affecting the model's ability to capture long-range dependencies. Conversely, longer sequences require concatenation of multiple documents, which can introduce noise and affect the natural document boundaries and semantic coherence as well as require substantial computational overhead. To address these challenges, we first establish three quantitative metrics for evaluating data composition quality: padding ratio, truncation ratio, and concatenation ratio. Building upon these metrics, we propose a novel multi-bucket data composition method that transcends the fixed-length paradigm. Our approach adaptively organizes training data to achieve optimal composition quality as measured by the proposed metrics, offering a more flexible and efficient approach for pre-training. We conduct extensive experiments and the results demonstrate that our proposed method significantly enhances both the efficiency and effectiveness of LLM pre-training.

cs.CL↗

arXiv2025

Full-Gap Superconductivity in BaAs/Ferropnictide Heterostructures

Interfacial interactions often promote the emergence of unusual phenomena in two-dimensional systems, including high-temperature superconductivity. Here, we report the observation of full-gap superconductivity with a maximal spectroscopic temperature up to 26 K in a BaAs monolayer grown on ferropnictide Ba(Fe$_{1-x}$Co$_x$)$_2$As$_2$ (abbreviated as BFCA) epitaxial films. The superconducting gap remains robust even when the thickness of underlying BFCA is reduced to the monolayer limit, in contrast to the rapid suppression of $T_\textrm{c}$ in standalone BFCA thin films. We reveal that the exceptional crystallinity of the BaAs/BFCA heterostructures, featured by their remarkable electronic and geometric uniformities, is crucial for the emergent full-gap superconductivity with mean-field temperature dependence and pronounced bound states within magnetic vortices. Our findings open up new avenues to unravel the mysteries of unconventional superconductivity in ferropnictides and advance the development of FeAs-based heterostructures.

cond-mat.supr-con↗

arXiv2025

Hunyuan3D 2.1: From Images to High-Fidelity 3D Assets with Production-Ready PBR Material

3D AI-generated content (AIGC) is a passionate field that has significantly accelerated the creation of 3D models in gaming, film, and design. Despite the development of several groundbreaking models that have revolutionized 3D generation, the field remains largely accessible only to researchers, developers, and designers due to the complexities involved in collecting, processing, and training 3D models. To address these challenges, we introduce Hunyuan3D 2.1 as a case study in this tutorial. This tutorial offers a comprehensive, step-by-step guide on processing 3D data, training a 3D generative model, and evaluating its performance using Hunyuan3D 2.1, an advanced system for producing high-resolution, textured 3D assets. The system comprises two core components: the Hunyuan3D-DiT for shape generation and the Hunyuan3D-Paint for texture synthesis. We will explore the entire workflow, including data preparation, model architecture, training strategies, evaluation metrics, and deployment. By the conclusion of this tutorial, you will have the knowledge to finetune or develop a robust 3D generative model suitable for applications in gaming, virtual reality, and industrial design.

cs.CV↗

arXiv2025

Universal machine learning interatomic potentials poised to supplant DFT in modeling general defects in metals and random alloys

Recent advances in machine learning, combined with the generation of extensive density functional theory (DFT) datasets, have enabled the development of universal machine learning interatomic potentials (uMLIPs). These models offer broad applicability across the periodic table, achieving first-principles accuracy at a fraction of the computational cost of traditional DFT calculations. In this study, we demonstrate that state-of-the-art pretrained uMLIPs can effectively replace DFT for accurately modeling complex defects in a wide range of metals and alloys. Our investigation spans diverse scenarios, including grain boundaries and general defects in pure metals, defects in high-entropy alloys, hydrogen-alloy interactions, and solute-defect interactions. Remarkably, the latest EquiformerV2 models achieve DFT-level accuracy on comprehensive defect datasets, with root mean square errors (RMSE) below 5 meV/atom for energies and 100 meV/Å for forces, outperforming specialized machine learning potentials such as moment tensor potential and atomic cluster expansion. We also present a systematic analysis of accuracy versus computational cost and explore uncertainty quantification for uMLIPs. A detailed case study of tungsten (W) demonstrates that data on pure W alone is insufficient for modeling complex defects in uMLIPs, underscoring the critical importance of advanced machine learning architectures and diverse datasets, which include over 100 million structures spanning all elements. These findings establish uMLIPs as a robust alternative to DFT and a transformative tool for accelerating the discovery and design of high-performance materials.

cond-mat.mtrl-sci↗

arXiv2025

A high-sensitivity frequency counter for free-induction-decay signals

Real-time frequency readout of time-dependent pulsed signals with a high sensitivity are key elements in many applications using atomic devices, such as FID atomic magnetometers. In this paper, we propose a frequency measurement algorithm based on the Hilbert transform and implement such a scheme in a FPGA-based frequency counter. By testing pulsed exponential-decay oscillation signals in the frequency range of 10 to 500 kHz, this frequency counter shows a frequency sensitivity better than 0.1 mHz/Hz^(1/2) at 10 Hz, with an output rate of 200 Hz. When the output rate is increased to 1000 Hz, the sensitivity remains better than 0.4 mHz/Hz^(1/2) at 10 Hz. The performance on frequency sensitivity is comparable with results obtained by off-line nonlinear fitting processes. In addition, this frequency counter does not require the pre-knowledge of the analytic expression of the input signals. The realization of such a device paves the way for practical applications of highly-sensitive FID atomic magnetometers.

physics.ins-det↗

arXiv2025

A Dynamic Transformer Network for Vehicle Detection

Stable consumer electronic systems can assist traffic better. Good traffic consumer electronic systems require collaborative work between traffic algorithms and hardware. However, performance of popular traffic algorithms containing vehicle detection methods based on deep networks via learning data relation rather than learning differences in different lighting and occlusions is limited. In this paper, we present a dynamic Transformer network for vehicle detection (DTNet). DTNet utilizes a dynamic convolution to guide a deep network to dynamically generate weights to enhance adaptability of an obtained detector. Taking into relations of different information account, a mixed attention mechanism based channel attention and Transformer is exploited to strengthen relations of channels and pixels to extract more salient information for vehicle detection. To overcome the drawback of difference in an image account, a translation-variant convolution relies on spatial location information to refine obtained structural information for vehicle detection. Experimental results illustrate that our DTNet is competitive for vehicle detection. Code of the proposed DTNet can be obtained at https://github.com/hellloxiaotian/DTNet.

cs.CV↗