arXiv subjects

Lei Chen

Publications and source records attributed to Lei Chen.

At least 163 records · Page 9Linked to original sources

arXiv2025

The Future of Fully Homomorphic Encryption System: from a Storage I/O Perspective

Fully Homomorphic Encryption (FHE) allows computations to be performed on encrypted data, significantly enhancing user privacy. However, the I/O challenges associated with deploying FHE applications remains understudied. We analyze the impact of storage I/O on the performance of FHE applications and summarize key lessons from the status quo. Key results include that storage I/O can degrade the performance of ASICs by as much as 357$\times$ and reduce GPUs performance by up to 22$\times$.

cs.CR↗

arXiv2025

PSD2Code: Automated Front-End Code Generation from Design Files via Multimodal Large Language Models

Design-to-code generation has emerged as a promising approach to bridge the gap between design prototypes and deployable frontend code. However, existing methods often suffer from structural inconsistencies, asset misalignment, and limited production readiness. This paper presents PSD2Code, a novel multi-modal approach that leverages PSD file parsing and asset alignment to generate production-ready React+SCSS code. Our method introduces a ParseAlignGenerate pipeline that extracts hierarchical structures, layer properties, and metadata from PSD files, providing large language models with precise spatial relationships and semantic groupings for frontend code generation. The system employs a constraint-based alignment strategy that ensures consistency between generated elements and design resources, while a structured prompt construction enhances controllability and code quality. Comprehensive evaluation demonstrates significant improvements over existing methods across multiple metrics including code similarity, visual fidelity, and production readiness. The method exhibits strong model independence across different large language models, validating the effectiveness of integrating structured design information with multimodal large language models for industrial-grade code generation, marking an important step toward design-driven automated frontend development.

cs.SE↗

arXiv2025

Phase Diagram and Spectroscopic Signatures of a Supersolid in Quantum Ising Magnet K$_2$Co(SeO$_3$)$_2$

A supersolid is a quantum-entangled state of matter exhibiting the dual characteristics of superfluidity and solidity. Theory predicts that hard-core bosons with repulsive interactions on a triangular lattice can form supersolid phases at half filling and near complete filling. Leveraging an exact mapping between bosons and spin-$\frac{1}{2}$ degrees of freedom, we investigate these phases in the spin-$\frac{1}{2}$ triangular-lattice antiferromagnet \K212 with exchange constants $J_z = 2.96(2)$~meV and $J_{\perp} = 0.21(3)$~meV. At zero field, neutron diffraction reveals the gradual development for $T<15$~K of quasi-two-dimensional $\sqrt{3}\times\sqrt{3}$ magnetic order with $Z_3$ translational symmetry breaking (solidity) albeit with 44(5)% reduced amplitude at $T=0.3$~K indicating strong quantum fluctuations. These are apparent in equidistant bands of continuum neutron scattering for $\hslashω_n\approx n\times J_z$, where $n=0,1,2,3$. The lowest energy ($n=0$) $\bf Q$-dependent continuum has a lower resonant edge and includes a quasi-elastic component at K $(\frac{1}{3}\frac{1}{3})$ consistent with broken $U(1)$ spin rotational symmetry (boson superfluidity). Competing instabilities are apparent in soft albeit finite-energy modes at M $(\frac{1}{2}0)$ and at $\frac{1}{2}$K $(\frac{1}{6}\frac{1}{6})$. For $\bf c$-axis-oriented magnetic fields $17~{\rm T} <μ_0 H< 21~{\rm T}$ that almost saturate the magnetization, corresponding to nearly filling the lattice with bosons, we find a new phase consistent with a second supersolid. These phases are separated by a pronounced 1/3 magnetization plateau that supports coherent spin waves, from which we determine the spin Hamiltonian.

cond-mat.str-el↗

arXiv2025

Automate Strategy Finding with LLM in Quant Investment

We present a novel three-stage framework leveraging Large Language Models (LLMs) within a risk-aware multi-agent system for automate strategy finding in quantitative finance. Our approach addresses the brittleness of traditional deep learning models in financial applications by: employing prompt-engineered LLMs to generate executable alpha factor candidates across diverse financial data, implementing multimodal agent-based evaluation that filters factors based on market status, predictive quality while maintaining category balance, and deploying dynamic weight optimization that adapts to market conditions. Experimental results demonstrate the robust performance of the strategy in Chinese & US market regimes compared to established benchmarks. Our work extends LLMs capabilities to quantitative trading, providing a scalable architecture for financial signal extraction and portfolio construction. The overall framework significantly outperforms all benchmarks with 53.17% cumulative return on SSE50 (Jan 2023 to Jan 2024), demonstrating superior risk-adjusted performance and downside protection on the market.

q-fin.PM↗

arXiv2025

Revealing the Potential of Learnable Perturbation Ensemble Forecast Model for Tropical Cyclone Prediction

Tropical cyclones (TCs) are highly destructive and inherently uncertain weather systems. Ensemble forecasting helps quantify these uncertainties, yet traditional systems are constrained by high computational costs and limited capability to fully represent atmospheric nonlinearity. FuXi-ENS introduces a learnable perturbation scheme for ensemble generation, representing a novel AI-based forecasting paradigm. Here, we systematically compare FuXi-ENS with ECMWF-ENS using all 90 global TCs in 2018, examining their performance in TC-related physical variables, track and intensity forecasts, and the associated dynamical and thermodynamical fields. FuXi-ENS demonstrates clear advantages in predicting TC-related physical variables, and achieves more accurate track forecasts with reduced ensemble spread, though it still underestimates intensity relative to observations. Further dynamical and thermodynamical analyses reveal that FuXi-ENS better captures large-scale circulation, with moisture turbulent energy more tightly concentrated around the TC warm core, whereas ECMWF-ENS exhibits a more dispersed distribution. These findings highlight the potential of learnable perturbations to improve TC forecasting skill and provide valuable insights for advancing AI-based ensemble prediction of extreme weather events that have significant societal impacts.

cs.LG↗

arXiv2025

AttentionPredictor: Temporal Patterns Matter for KV Cache Compression

With the development of large language models (LLMs), efficient inference through Key-Value (KV) cache compression has attracted considerable attention, especially for long-context generation. To compress the KV cache, recent methods identify critical KV tokens through static modeling of attention scores. However, these methods often struggle to accurately determine critical tokens as they neglect the temporal patterns in attention scores, resulting in a noticeable degradation in LLM performance. To address this challenge, we propose AttentionPredictor, which is the first learning-based method to directly predict attention patterns for KV cache compression and critical token identification. Specifically, AttentionPredictor learns a lightweight, unified convolution model to dynamically capture spatiotemporal patterns and predict the next-token attention scores. An appealing feature of AttentionPredictor is that it accurately predicts the attention score and shares the unified prediction model, which consumes negligible memory, among all transformer layers. Moreover, we propose a cross-token critical cache prefetching framework that hides the token estimation time overhead to accelerate the decoding stage. By retaining most of the attention information, AttentionPredictor achieves 13$\times$ KV cache compression and 5.6$\times$ speedup in a cache offloading scenario with comparable LLM performance, significantly outperforming the state-of-the-arts. The code is available at https://github.com/MIRALab-USTC/LLM-AttentionPredictor.

cs.CL↗

arXiv2025

Every Step Evolves: Scaling Reinforcement Learning for Trillion-Scale Thinking Model

We present Ring-1T, the first open-source, state-of-the-art thinking model with a trillion-scale parameter. It features 1 trillion total parameters and activates approximately 50 billion per token. Training such models at a trillion-parameter scale introduces unprecedented challenges, including train-inference misalignment, inefficiencies in rollout processing, and bottlenecks in the RL system. To address these, we pioneer three interconnected innovations: (1) IcePop stabilizes RL training via token-level discrepancy masking and clipping, resolving instability from training-inference mismatches; (2) C3PO++ improves resource utilization for long rollouts under a token budget by dynamically partitioning them, thereby obtaining high time efficiency; and (3) ASystem, a high-performance RL framework designed to overcome the systemic bottlenecks that impede trillion-parameter model training. Ring-1T delivers breakthrough results across critical benchmarks: 93.4 on AIME-2025, 86.72 on HMMT-2025, 2088 on CodeForces, and 55.94 on ARC-AGI-1. Notably, it attains a silver medal-level result on the IMO-2025, underscoring its exceptional reasoning capabilities. By releasing the complete 1T parameter MoE model to the community, we provide the research community with direct access to cutting-edge reasoning capabilities. This contribution marks a significant milestone in democratizing large-scale reasoning intelligence and establishes a new baseline for open-source model performance.

cs.CL↗

arXiv2025

GoRA: Gradient-driven Adaptive Low Rank Adaptation

Low-Rank Adaptation (LoRA) is a crucial method for efficiently fine-tuning large language models (LLMs), with its effectiveness influenced by two key factors: rank selection and weight initialization. While numerous LoRA variants have been proposed to improve performance by addressing one of these aspects, they often compromise usability or computational efficiency. In this paper, we analyze and identify the core limitations of existing approaches and propose a novel framework--GoRA (Gradient-driven Adaptive Low Rank Adaptation)--that simultaneously adapts both the rank and initialization strategy within a unified framework. GoRA leverages gradient information during training to dynamically assign optimal ranks and initialize low-rank adapter weights in an adaptive manner. To our knowledge, GoRA is the first method that not only addresses the limitations of prior approaches--which often focus on either rank selection or initialization in isolation--but also unifies both aspects within a single framework, enabling more effective and efficient adaptation. Extensive experiments across various architectures and modalities show that GoRA consistently outperforms existing LoRA-based methods while preserving the efficiency of vanilla LoRA. For example, when fine-tuning Llama3.1-8B-Base for mathematical reasoning, GoRA achieves a 5.13-point improvement over standard LoRA and even outperforms full fine-tuning by 2.05 points under high-rank settings. Code is available at: https://github.com/hhnqqq/MyTransformers.

cs.LG↗

arXiv2025

FuXi-Ocean: A Global Ocean Forecasting System with Sub-Daily Resolution

Accurate, high-resolution ocean forecasting is crucial for maritime operations and environmental monitoring. While traditional numerical models are capable of producing sub-daily, eddy-resolving forecasts, they are computationally intensive and face challenges in maintaining accuracy at fine spatial and temporal scales. In contrast, recent data-driven approaches offer improved computational efficiency and emerging potential, yet typically operate at daily resolution and struggle with sub-daily predictions due to error accumulation over time. We introduce FuXi-Ocean, the first data-driven global ocean forecasting model achieving six-hourly predictions at eddy-resolving 1/12° spatial resolution, reaching depths of up to 1500 meters. The model architecture integrates a context-aware feature extraction module with a predictive network employing stacked attention blocks. The core innovation is the Mixture-of-Time (MoT) module, which adaptively integrates predictions from multiple temporal contexts by learning variable-specific reliability , mitigating cumulative errors in sequential forecasting. Through comprehensive experimental evaluation, FuXi-Ocean demonstrates superior skill in predicting key variables, including temperature, salinity, and currents, across multiple depths.

cs.LG↗

arXiv2025

CARPO: Leveraging Listwise Learning-to-Rank for Context-Aware Query Plan Optimization

Efficient data processing is increasingly vital, with query optimizers playing a fundamental role in translating SQL queries into optimal execution plans. Traditional cost-based optimizers, however, often generate suboptimal plans due to flawed heuristics and inaccurate cost models, leading to the emergence of Learned Query Optimizers (LQOs). To address challenges in existing LQOs, such as the inconsistency and suboptimality inherent in pairwise ranking methods, we introduce CARPO, a generic framework leveraging listwise learning-to-rank for context-aware query plan optimization. CARPO distinctively employs a Transformer-based model for holistic evaluation of candidate plan sets and integrates a robust hybrid decision mechanism, featuring Out-Of-Distribution (OOD) detection with a top-k fallback strategy to ensure reliability. Furthermore, CARPO can be seamlessly integrated with existing plan embedding techniques, demonstrating strong adaptability. Comprehensive experiments on TPC-H and STATS benchmarks demonstrate that CARPO significantly outperforms both native PostgreSQL and Lero, achieving a Top-1 Rate of 74.54% on the TPC-H benchmark compared to Lero's 3.63%, and reducing the total execution time to 3719.16 ms compared to PostgreSQL's 22577.87 ms.

cs.DB↗

arXiv2025

SmaRTLy: RTL Optimization with Logic Inferencing and Structural Rebuilding

This paper proposes smaRTLy: a new optimization technique for multiplexers in Register-Transfer Level (RTL) logic synthesis. Multiplexer trees are very common in RTL designs, and traditional tools like Yosys optimize them by traversing the tree and monitoring control port values. However, this method does not fully exploit the intrinsic logical relationships among signals or the potential for structural optimization. To address these limitations, we develop innovative strategies to remove redundant multiplexer trees and restructure the remaining ones, significantly reducing the overall gate count. We evaluate smaRTLy on the IWLS-2005 and RISC-V benchmarks, achieving an additional 8.95% reduction in AIG area compared to Yosys. We also evaluate smaRTLy on an industrial benchmark in the scale of millions of gates, results show that smaRTLy can remove 47.2% more AIG area than Yosys. These results demonstrate the effectiveness of our logic inferencing and structural rebuilding techniques in enhancing the RTL optimization process, leading to more efficient hardware designs.

cs.AR↗

arXiv2025

Approximate Nearest Neighbor Search of Large Scale Vectors on Distributed Storage

Approximate Nearest Neighbor Search (ANNS) in high-dimensional space is an essential operator in many online services, such as information retrieval and recommendation. Indices constructed by the state-of-the-art ANNS algorithms must be stored in single machine's memory or disk for high recall rate and throughput, suffering from substantial storage cost, constraint of limited scale and single point of failure. While distributed storage can provide a cost-effective and robust solution, there is no efficient and effective algorithms for indexing vectors in distributed storage scenarios. In this paper, we present a new graph-cluster hybrid indexing and search system which supports Distributed Storage Approximate Nearest Neighbor Search, called DSANN. DSANN can efficiently index, store, search billion-scale vector database in distributed storage and guarantee the high availability of index service. DSANN employs the concurrent index construction method to significantly reduces the complexity of index building. Then, DSANN applies Point Aggregation Graph to leverage the structural information of graph to aggregate similar vectors, optimizing storage efficiency and improving query throughput via asynchronous I/O in distributed storage. Through extensive experiments, we demonstrate DSANN can efficiently and effectively index, store and search large-scale vector datasets in distributed storage scenarios.

cs.DB↗

arXiv2025

Pseudo-Random TDM-MIMO FMCW Based Millimeter-Wave Sensing and Communication Integration for UAV Swarm

The integrated sensing and communications (ISAC) can achieve the sharing of hardware and spectrum resources, enabling efficient data transmission and environmental sensing. This fusion is particularly important for unmanned aerial vehicle (UAV) swarms, as it enhances the overall performance, flexibility, and efficiency of such systems. To facilitate the collaborative operations among UAVs, this paper proposes an ISAC solution based on the pseudo-random time-division multiplexing (TDM)-multiple input multiple output (MIMO) millimeter-wave (mmWave) frequency modulated continuous wave (FMCW). Specifically, a novel ISAC chirp waveform is proposed to modulate data in both the delay domain and complex amplitude, while also possessing high-precision sensing capabilities. To address challenges in the TDM-MIMO, we utilize the pseudo-random antenna selection and compressed sensing algorithms, ensuring that the maximum unambiguous velocity is not compromised. Moreover, by employing a chirp-division multiple access scheme, we propose an interference-free multiple antenna transmission scheme to achieve dynamic allocation of time-frequency resources and multi-user transmission. Finally, we propose a communication and sensing fusion-based dynamic iterative computation scheme, simultaneously achieving data demodulation and sensing parameter estimation. Simulation results show that the proposed scheme can achieve ISAC under the dynamic flight scenarios of UAVs. Meanwhile, the scheme outperforms the mmWave-LoRadar in communication and sensing performance, yet its sensing performance is slightly lower than that of the traditional FMCW. Under the urban clutter modeling, the scheme still maintains favorable robustness despite a certain degree of performance degradation.

eess.SP↗

arXiv2025

The normal closure of a bounding pair map in its mapping class group

Johnson showed that the normal subgroup of a mapping class group generated by the genus $1$ bounding pair maps is equal to the Torelli group. Generalizing Johnson's result, we give two descriptions of the normal subgroup generated by the genus $n$ bounding pair maps using the Chillingworth and the Casson--Morita invariants.

math.GT↗

arXiv2025

Bi-level Meta-Policy Control for Dynamic Uncertainty Calibration in Evidential Deep Learning

Traditional Evidence Deep Learning (EDL) methods rely on static hyperparameter for uncertainty calibration, limiting their adaptability in dynamic data distributions, which results in poor calibration and generalization in high-risk decision-making tasks. To address this limitation, we propose the Meta-Policy Controller (MPC), a dynamic meta-learning framework that adjusts the KL divergence coefficient and Dirichlet prior strengths for optimal uncertainty modeling. Specifically, MPC employs a bi-level optimization approach: in the inner loop, model parameters are updated through a dynamically configured loss function that adapts to the current training state; in the outer loop, a policy network optimizes the KL divergence coefficient and class-specific Dirichlet prior strengths based on multi-objective rewards balancing prediction accuracy and uncertainty quality. Unlike previous methods with fixed priors, our learnable Dirichlet prior enables flexible adaptation to class distributions and training dynamics. Extensive experimental results show that MPC significantly enhances the reliability and calibration of model predictions across various tasks, improving uncertainty calibration, prediction accuracy, and performance retention after confidence-based sample rejection.

cs.LG↗

arXiv2025

Progressive Uncertainty-Guided Evidential U-KAN for Trustworthy Medical Image Segmentation

Trustworthy medical image segmentation aims at deliver accurate and reliable results for clinical decision-making. Most existing methods adopt the evidence deep learning (EDL) paradigm due to its computational efficiency and theoretical robustness. However, the EDL-based methods often neglect leveraging uncertainty maps rich in attention cues to refine ambiguous boundary segmentation. To address this, we propose a progressive evidence uncertainty guided attention (PEUA) mechanism to guide the model to focus on the feature representation learning of hard regions. Unlike conventional approaches, PEUA progressively refines attention using uncertainty maps while employing low-rank learning to denoise attention weights, enhancing feature learning for challenging regions. Concurrently, standard EDL methods suppress evidence of incorrect class indiscriminately via Kullback-Leibler (KL) regularization, impairing the uncertainty assessment in ambiguous areas and consequently distorts the corresponding attention guidance. We thus introduce a semantic-preserving evidence learning (SAEL) strategy, integrating a semantic-smooth evidence generator and a fidelity-enhancing regularization term to retain critical semantics. Finally, by embedding PEUA and SAEL with the state-of-the-art U-KAN, we proposes Evidential U-KAN, a novel solution for trustworthy medical image segmentation. Extensive experiments on 4 datasets demonstrate superior accuracy and reliability over the competing methods. The code is available at \href{https://anonymous.4open.science/r/Evidence-U-KAN-BBE8}{github}.

eess.IV↗

arXiv2025

$\bf{D^3}$QE: Learning Discrete Distribution Discrepancy-aware Quantization Error for Autoregressive-Generated Image Detection

The emergence of visual autoregressive (AR) models has revolutionized image generation while presenting new challenges for synthetic image detection. Unlike previous GAN or diffusion-based methods, AR models generate images through discrete token prediction, exhibiting both marked improvements in image synthesis quality and unique characteristics in their vector-quantized representations. In this paper, we propose to leverage Discrete Distribution Discrepancy-aware Quantization Error (D$^3$QE) for autoregressive-generated image detection that exploits the distinctive patterns and the frequency distribution bias of the codebook existing in real and fake images. We introduce a discrete distribution discrepancy-aware transformer that integrates dynamic codebook frequency statistics into its attention mechanism, fusing semantic features and quantization error latent. To evaluate our method, we construct a comprehensive dataset termed ARForensics covering 7 mainstream visual AR models. Experiments demonstrate superior detection accuracy and strong generalization of D$^3$QE across different AR models, with robustness to real-world perturbations. Code is available at \href{https://github.com/Zhangyr2022/D3QE}{https://github.com/Zhangyr2022/D3QE}.

cs.CV↗

arXiv2025

LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues

Multi-turn dialogues are essential in many real-world applications of large language models, such as chatbots and virtual assistants. As conversation histories become longer, existing large language models face increasing computational and memory challenges, which hinder their ability to provide efficient and responsive interactions. Most current acceleration methods either compress the context or optimize key value caching, but they often rely on fixed or position-based heuristics that do not adapt well to the dynamic and unpredictable patterns found in actual multi-turn conversations. As a result, these models cannot accurately identify and prioritize the most relevant context, leading to degraded response quality. In this paper, we present LoopServe, an adaptive dual-phase inference acceleration framework for large language models in multi-turn dialogues. LoopServe introduces two main innovations. First, it performs online sparsification during the prefilling phase by dynamically selecting the most important parts of the attention matrix for each new input. Second, it uses progressive key value compression during decoding by adaptively maintaining a relevant and efficient cache based on the most recently generated output tokens. We also propose a new benchmark with eleven multi-turn datasets that reflect realistic query positions and conversational dependencies. Extensive experiments demonstrate that LoopServe consistently achieves superior effectiveness compared to existing baselines and significantly accelerates LLM inference across a wide range of long-context dialogue tasks.

cs.CL↗