arXiv subjects

Yuqi Li

Publications and source records attributed to Yuqi Li.

At least 73 records · Page 4Linked to original sources

arXiv2025

The Structural Scalpel: Automated Contiguous Layer Pruning for Large Language Models

Although large language models (LLMs) have achieved revolutionary breakthroughs in many fields, their large model size and high computational cost pose significant challenges for practical deployment on resource-constrained edge devices. To this end, layer pruning has been proposed to reduce the computational overhead by directly removing redundant layers. However, existing layer pruning methods typically rely on hand-crafted metrics to evaluate and remove individual layers, while ignoring the dependencies between layers. This can disrupt the model's information flow and severely degrade performance. To address these issues, we propose CLP, a novel continuous layer pruning framework that introduces two key innovations: a differentiable concave gate algorithm that automatically identifies the best continuous layer segments for pruning via gradient-based optimization; and a cutoff endpoint tuning strategy that effectively restores model performance by fine-tuning only the layers adjacent to the pruned segments. Extensive experiments across multiple model architectures (including LLaMA2, LLaMA3 and Qwen) and sizes (from $7$B to $70$B parameters) show that CLP significantly outperforms existing state-of-the-art baselines. For example, at a pruning rate of $20\%$, CLP achieves an average performance retention of $95.34\%$ on LLaMA3-70B, outperforming baselines by $4.29\%$-$30.52\%$. Furthermore, CLP can be seamlessly combined with quantization to further compress the model with only a slight performance loss.

cs.LG↗

arXiv2025

Robust and Hyper-Efficient Multi-dimensional Optical Fiber Semantic Communication

The growing demands of artificial intelligence and immersive media require communication beyond bit-level accuracy to meaning awareness. Conventional optical systems that focused on syntactic precision suffer significant inefficiencies. Here, we introduce a multi-dimensional semantic communication framework that bridges this gap by directly mapping high-level semantic features onto the orthogonal physical dimensions of light, frequency, polarization, and intensity, within a multimode fiber. This synergistic co-design of semantic logic and the photonic channel achieve an unprecedented equivalent spectral efficiency approaching 1000 bit/s/Hz. Moreover, it demonstrates profound resilience, maintaining high-fidelity reconstruction even when the physical-layer symbol error rate exceeds 36%, a condition under which conventional communication systems fail completely. Crucially, this deeply integrated co-design of semantic encoding and physical-layer modulation enables full semantic demodulation with only single-ended intensity detection, therefore significantly reducing system complexity and cost. This work establishes a validated pathway toward hyper-efficient, error-resilient optical networks for the next generation of data-intensive computing.

physics.optics↗

arXiv2025

Epitaxial Electrodeposition of Fe with Controlled In-Plane Variants for Reversible Metal Anode in Aqueous Electrolyte

The development of reversible metal anodes is a key challenge for advancing aqueous battery technologies, particularly for scalable and safe stationary energy storage applications. Here we demonstrate a strategy to realize epitaxial electrodeposition of iron (Fe) on single-crystal copper (Cu) substrates in aqueous electrolytes. We compare the electrodeposition behavior of Fe on polycrystalline and single-crystalline Cu substrates, revealing that the latter enables highly uniform, dense, and crystallographically aligned Fe growth. Comprehensive electron backscatter diffraction (EBSD) and X-ray diffraction (XRD) analysis confirms the formation of Fe with specific out-of-plane and in-plane orientations, including well-defined rotational variants. Our findings highlight that epitaxial electrodeposition of Fe can suppress dendritic growth and significantly enhance Coulombic efficiency during plating/stripping cycles. This approach bridges fundamental crystallography with practical electrochemical performance, providing a pathway toward high-efficiency aqueous batteries utilizing Earth-abundant materials.

cond-mat.mtrl-sci↗

arXiv2025

ClusterRCA: An End-to-End Approach for Network Fault Localization and Classification for HPC System

Network failure diagnosis is challenging yet critical for high-performance computing (HPC) systems. Existing methods cannot be directly applied to HPC scenarios due to data heterogeneity and lack of accuracy. This paper proposes a novel framework, called ClusterRCA, to localize culprit nodes and determine failure types by leveraging multimodal data. ClusterRCA extracts features from topologically connected network interface controller (NIC) pairs to analyze the diverse, multimodal data in HPC systems. To accurately localize culprit nodes and determine failure types, ClusterRCA combines classifier-based and graph-based approaches. A failure graph is constructed based on the output of the state classifier, and then it performs a customized random walk on the graph to localize the root cause. Experiments on datasets collected by a top-tier global HPC device vendor show ClusterRCA achieves high accuracy in diagnosing network failure for HPC systems. ClusterRCA also maintains robust performance across different application scenarios.

cs.DC↗

arXiv2025

PanoLora: Bridging Perspective and Panoramic Video Generation with LoRA Adaptation

Generating high-quality 360° panoramic videos remains a significant challenge due to the fundamental differences between panoramic and traditional perspective-view projections. While perspective videos rely on a single viewpoint with a limited field of view, panoramic content requires rendering the full surrounding environment, making it difficult for standard video generation models to adapt. Existing solutions often introduce complex architectures or large-scale training, leading to inefficiency and suboptimal results. Motivated by the success of Low-Rank Adaptation (LoRA) in style transfer tasks, we propose treating panoramic video generation as an adaptation problem from perspective views. Through theoretical analysis, we demonstrate that LoRA can effectively model the transformation between these projections when its rank exceeds the degrees of freedom in the task. Our approach efficiently fine-tunes a pretrained video diffusion model using only approximately 1,000 videos while achieving high-quality panoramic generation. Experimental results demonstrate that our method maintains proper projection geometry and surpasses previous state-of-the-art approaches in visual quality, left-right consistency, and motion diversity.

cs.CV↗

arXiv2025

LoRALib: A Standardized Benchmark for Evaluating LoRA-MoE Methods

As a parameter efficient fine-tuning (PEFT) method, low-rank adaptation (LoRA) can save significant costs in storage and computing, but its strong adaptability to a single task is often accompanied by insufficient cross-task generalization capabilities. To improve this, existing work combines LoRA with mixture-of-experts (MoE) to enhance the model's adaptability through expert modules and routing mechanisms. However, existing LoRA-MoE methods lack unified standards in models, datasets, hyperparameters, and evaluation methods, making it difficult to conduct fair comparisons between different methods. To this end, we proposed a unified benchmark named LoRALib. Specifically, we standardized datasets from $40$ downstream tasks into a unified format, fine-tuned them using the same hyperparameters and obtained $680$ LoRA modules across $17$ model architectures. Based on this LoRA library, we conduct large-scale experiments on $3$ representative LoRA-MoE methods and different LoRA selection mechanisms using the open-sourced testing tool OpenCompass. Extensive experiments show that LoRAMoE performs best, and that prioritizing LoRAs relevant to the target task can further improve the performance of MoE. We hope these findings will inspire future work. Our datasets and LoRA library are available at https://huggingface.co/datasets/YaoLuzjut/LoRAOcean_dataset and https://huggingface.co/YaoLuzjut/models.

cs.LG↗

arXiv2025

AMMKD: Adaptive Multimodal Multi-teacher Distillation for Lightweight Vision-Language Models

The success of large-scale visual language pretraining (VLP) models has driven widespread adoption of image-text retrieval tasks. However, their deployment on mobile devices remains limited due to large model sizes and computational complexity. We propose Adaptive Multi-Modal Multi-Teacher Knowledge Distillation (AMMKD), a novel framework that integrates multi-modal feature fusion, multi-teacher distillation, and adaptive optimization to deliver lightweight yet effective retrieval models. Specifically, our method begins with a feature fusion network that extracts and merges discriminative features from both the image and text modalities. To reduce model parameters and further improve performance, we design a multi-teacher knowledge distillation framework to pre-train two CLIP teacher models. We decouple modalities by pre-computing and storing text features as class vectors via the teacher text encoder to enhance efficiency. To better align teacher and student outputs, we apply KL scatter for probability distribution matching. Finally, we design an adaptive dynamic weighting scheme that treats multi-teacher distillation as a multi-objective optimization problem. By leveraging gradient space diversity, we dynamically adjust the influence of each teacher, reducing conflicts and guiding the student toward more optimal learning directions. Extensive experiments on three benchmark datasets demonstrate that AMMKD achieves superior performance while significantly reducing model complexity, validating its effectiveness and flexibility.

cs.CV↗

arXiv2025

Scaling Learned Image Compression Models up to 1 Billion

Recent advances in large language models (LLMs) highlight a strong connection between intelligence and compression. Learned image compression, a fundamental task in modern data compression, has made significant progress in recent years. However, current models remain limited in scale, restricting their representation capacity, and how scaling model size influences compression performance remains unexplored. In this work, we present a pioneering study on scaling up learned image compression models and revealing the performance trends through scaling laws. Using the recent state-of-the-art HPCM model as baseline, we scale model parameters from 68.5 millions to 1 billion and fit power-law relations between test loss and key scaling variables, including model size and optimal training compute. The results reveal a scaling trend, enabling extrapolation to larger scale models. Experimental results demonstrate that the scaled-up HPCM-1B model achieves state-of-the-art rate-distortion performance. We hope this work inspires future exploration of large-scale compression models and deeper investigations into the connection between compression and intelligence.

cs.CV↗

arXiv2025

Matrix-3D: Omnidirectional Explorable 3D World Generation

Explorable 3D world generation from a single image or text prompt forms a cornerstone of spatial intelligence. Recent works utilize video model to achieve wide-scope and generalizable 3D world generation. However, existing approaches often suffer from a limited scope in the generated scenes. In this work, we propose Matrix-3D, a framework that utilize panoramic representation for wide-coverage omnidirectional explorable 3D world generation that combines conditional video generation and panoramic 3D reconstruction. We first train a trajectory-guided panoramic video diffusion model that employs scene mesh renders as condition, to enable high-quality and geometrically consistent scene video generation. To lift the panorama scene video to 3D world, we propose two separate methods: (1) a feed-forward large panorama reconstruction model for rapid 3D scene reconstruction and (2) an optimization-based pipeline for accurate and detailed 3D scene reconstruction. To facilitate effective training, we also introduce the Matrix-Pano dataset, the first large-scale synthetic collection comprising 116K high-quality static panoramic video sequences with depth and trajectory annotations. Extensive experiments demonstrate that our proposed framework achieves state-of-the-art performance in panoramic video generation and 3D world generation. See more in https://matrix-3d.github.io.

cs.CV↗

arXiv2025

Learned Image Compression with Hierarchical Progressive Context Modeling

Context modeling is essential in learned image compression for accurately estimating the distribution of latents. While recent advanced methods have expanded context modeling capacity, they still struggle to efficiently exploit long-range dependency and diverse context information across different coding steps. In this paper, we introduce a novel Hierarchical Progressive Context Model (HPCM) for more efficient context information acquisition. Specifically, HPCM employs a hierarchical coding schedule to sequentially model the contextual dependencies among latents at multiple scales, which enables more efficient long-range context modeling. Furthermore, we propose a progressive context fusion mechanism that incorporates contextual information from previous coding steps into the current step, effectively exploiting diverse contextual information. Experimental results demonstrate that our method achieves state-of-the-art rate-distortion performance and strikes a better balance between compression performance and computational complexity. The code is available at https://github.com/lyq133/LIC-HPCM.

eess.IV↗

arXiv2025

Frequency-Aligned Knowledge Distillation for Lightweight Spatiotemporal Forecasting

Spatiotemporal forecasting tasks, such as traffic flow, combustion dynamics, and weather forecasting, often require complex models that suffer from low training efficiency and high memory consumption. This paper proposes a lightweight framework, Spectral Decoupled Knowledge Distillation (termed SDKD), which transfers the multi-scale spatiotemporal representations from a complex teacher model to a more efficient lightweight student network. The teacher model follows an encoder-latent evolution-decoder architecture, where its latent evolution module decouples high-frequency details and low-frequency trends using convolution and Transformer (global low-frequency modeler). However, the multi-layer convolution and deconvolution structures result in slow training and high memory usage. To address these issues, we propose a frequency-aligned knowledge distillation strategy, which extracts multi-scale spectral features from the teacher's latent space, including both high and low frequency components, to guide the lightweight student model in capturing both local fine-grained variations and global evolution patterns. Experimental results show that SDKD significantly improves performance, achieving reductions of up to 81.3% in MSE and in MAE 52.3% on the Navier-Stokes equation dataset. The framework effectively captures both high-frequency variations and long-term trends while reducing computational complexity. Our codes are available at https://github.com/itsnotacie/SDKD

cs.LG↗

arXiv2025

Cross Knowledge Distillation between Artificial and Spiking Neural Networks

Recently, Spiking Neural Networks (SNNs) have demonstrated rich potential in computer vision domain due to their high biological plausibility, event-driven characteristic and energy-saving efficiency. Still, limited annotated event-based datasets and immature SNN architectures result in their performance inferior to that of Artificial Neural Networks (ANNs). To enhance the performance of SNNs on their optimal data format, DVS data, we explore using RGB data and well-performing ANNs to implement knowledge distillation. In this case, solving cross-modality and cross-architecture challenges is necessary. In this paper, we propose cross knowledge distillation (CKD), which not only leverages semantic similarity and sliding replacement to mitigate the cross-modality challenge, but also uses an indirect phased knowledge distillation to mitigate the cross-architecture challenge. We validated our method on main-stream neuromorphic datasets, including N-Caltech101 and CEP-DVS. The experimental results show that our method outperforms current State-of-the-Art methods. The code will be available at https://github.com/ShawnYE618/CKD

cs.CV↗

arXiv2025

MPQ-DMv2: Flexible Residual Mixed Precision Quantization for Low-Bit Diffusion Models with Temporal Distillation

Diffusion models have demonstrated remarkable performance on vision generation tasks. However, the high computational complexity hinders its wide application on edge devices. Quantization has emerged as a promising technique for inference acceleration and memory reduction. However, existing quantization methods do not generalize well under extremely low-bit (2-4 bit) quantization. Directly applying these methods will cause severe performance degradation. We identify that the existing quantization framework suffers from the outlier-unfriendly quantizer design, suboptimal initialization, and optimization strategy. We present MPQ-DMv2, an improved \textbf{M}ixed \textbf{P}recision \textbf{Q}uantization framework for extremely low-bit \textbf{D}iffusion \textbf{M}odels. For the quantization perspective, the imbalanced distribution caused by salient outliers is quantization-unfriendly for uniform quantizer. We propose \textit{Flexible Z-Order Residual Mixed Quantization} that utilizes an efficient binary residual branch for flexible quant steps to handle salient error. For the optimization framework, we theoretically analyzed the convergence and optimality of the LoRA module and propose \textit{Object-Oriented Low-Rank Initialization} to use prior quantization error for informative initialization. We then propose \textit{Memory-based Temporal Relation Distillation} to construct an online time-aware pixel queue for long-term denoising temporal information distillation, which ensures the overall temporal consistency between quantized and full-precision model. Comprehensive experiments on various generation tasks show that our MPQ-DMv2 surpasses current SOTA methods by a great margin on different architectures, especially under extremely low-bit widths.

cs.CV↗

arXiv2025

Can tidal evolution lead to close-in planetary bodies around white dwarfs II: volcanism and transits

Planetary material accreted by white dwarfs provides unique insights regarding exoplanetary composition. The evolutionary pathways of planetary bodies around white dwarfs are crucial to understanding the presence of close-in planetary material, observed in the form of pollutants in the atmospheres of white dwarfs and planetary material transiting white dwarfs. Periodic transits around white dwarfs potentially reveal the existence of close-in planetary bodies undergoing dust production. Tidal interactions can bring planetesimals that have been gravitationally perturbed onto long-period highly eccentric orbits around white dwarfs towards shorter orbital periods and smaller eccentricities. Tidal interactions may also induce melting and volcanism in these planetesimals, potentially being a mechanism for dust and debris production, the result of which may be seen in transit. Tidally induced volcanism may be triggered in a wide parameter space: for a 100km-sized rocky planetesimals perturbed to a pericentre distance $\lesssim$ 0.01AU ($\gtrsim$ twice its Roche limit), both on long-period (~ 100day) highly eccentric orbits and short-period (~ 10hr) near circular orbits. We comment on the potential link between the resultant volcanic ejecta and observed optical transits.

astro-ph.EP↗

arXiv2025

Can tidal evolution lead to close-in planetary bodies around white dwarfs I: Orbital period distribution

The evolution of planetary systems around white dwarfs is crucial to understanding the presence of planetary material in the atmospheres of white dwarfs. These systems uniquely probe exoplanetary compositions. Periodic signals in the photometry of a handful of white dwarfs suggest material blocking the star, potentially from disintegrating planetesimals. Tidal evolution followed by scattering can bring planetesimals onto close-in orbits that would have been within the envelope of the white dwarf progenitor. The orbital period distribution of planetesimals undergoing tidal evolution will peak at short-period (nearly) circularized orbits (~ 10 hour-1 day), with a rising tail towards long-period highly eccentric orbits (~ 100 day). This prediction is generally consistent with the observed white dwarf transiting systems. In order for the planetesimal on the 4.5 hour period around WD 1145+017 to be explained by the tidal evolution of a planetesimal, that planetesimal must have an ultimate tensile strength comparable to that of iron meteorites.

astro-ph.EP↗

arXiv2025

Measurements of three exo-planetesimal compositions: a planetary core, a chondritic body, and an icy Kuiper belt analogue

The study of planetesimal debris accreted by white dwarfs offers unique insights into the composition of exoplanets. Using far-ultraviolet and optical spectroscopy, we have analysed the composition of planetesimals accreted by three metal enriched H-dominated white dwarfs with effective temperatures of T_eff = 20 000 K. WD 0059+257 is accreting an object composed of 71.8 +/- 7.9 per cent Fe and Ni by mass, indicating a large core mass fraction of 69 per cent, similar to that of Mercury. We model this planetesimal as having a differentiated Earth-like composition with 65 per cent of its mantle stripped, and we find this mass loss can be caused by vaporisation of the planetesimal's mantle during post-main sequence evolution. The tentative S detection in WD 0059+257 is a possible clue to the nature of the light element in planetary cores, including that of the Earth. The volatile-rich composition of WD 1943+163 is consistent with accretion of a carbonaceous chondrite-like object, but with an extreme Si depletion. WD 1953-715 accretes a planetesimal which contains 64 +/- 21 per cent of O in the form of ices, likely H2O. This body therefore requires an initial orbit at formation beyond a radial distance of > 100 au for ice survival into the white dwarf phase. These three planetary enriched white dwarfs provide evidence of differing core fractions, volatile budgets, and initial orbital separations of the accreted planetesimals, all of which help us understand their formation and evolutionary history.

astro-ph.EP↗

arXiv2025

Physics-informed mixture of experts network for interpretable battery degradation trajectory computation amid second-life complexities

Retired electric vehicle batteries offer immense potential to support low-carbon energy systems, but uncertainties in their degradation behavior and data inaccessibilities under second-life use pose major barriers to safe and scalable deployment. This work proposes a Physics-Informed Mixture of Experts (PIMOE) network that computes battery degradation trajectories using partial, field-accessible signals in a single cycle. PIMOE leverages an adaptive multi-degradation prediction module to classify degradation modes using expert weight synthesis underpinned by capacity-voltage and relaxation data, producing latent degradation trend embeddings. These are input to a use-dependent recurrent network for long-term trajectory prediction. Validated on 207 batteries across 77 use conditions and 67,902 cycles, PIMOE achieves an average mean absolute percentage (MAPE) errors of 0.88% with a 0.43 ms inference time. Compared to the state-of-the-art Informer and PatchTST, it reduces computational time and MAPE by 50%, respectively. Compatible with random state of charge region sampling, PIMOE supports 150-cycle forecasts with 1.50% average and 6.26% maximum MAPE, and operates effectively even with pruned 5MB training data. Broadly, PIMOE framework offers a deployable, history-free solution for battery degradation trajectory computation, redefining how second-life energy storage systems are assessed, optimized, and integrated into the sustainable energy landscape.

cs.LG↗

arXiv2025

SRKD: Towards Efficient 3D Point Cloud Segmentation via Structure- and Relation-aware Knowledge Distillation

3D point cloud segmentation faces practical challenges due to the computational complexity and deployment limitations of large-scale transformer-based models. To address this, we propose a novel Structure- and Relation-aware Knowledge Distillation framework, named SRKD, that transfers rich geometric and semantic knowledge from a large frozen teacher model (>100M) to a lightweight student model (<15M). Specifically, we propose an affinity matrix-based relation alignment module, which distills structural dependencies from the teacher to the student through point-wise similarity matching, enhancing the student's capability to learn contextual interactions. Meanwhile, we introduce a cross-sample mini-batch construction strategy that enables the student to perceive stable and generalized geometric structure. This aligns across diverse point cloud instances of the teacher, rather than within a single sample. Additionally, KL divergence is applied to align semantic distributions, and ground-truth supervision further reinforces accurate segmentation. Our method achieves state of the art performance with significantly reduced model complexity, demonstrating its effectiveness and efficiency in real-world deployment scenarios. Our Code is available at https://github.com/itsnotacie/SRKD.

cs.CV↗