arXiv subjects

Miao Yin

Publications and source records attributed to Miao Yin.

At least 19 recordsLinked to original sources

arXiv2026

Allocation Before Ranking: Decoupled Token Compression for OmniLLMs

Token compression in OmniLLMs is typically posed as a single saliency-ranking problem: score each multimodal token, keep the top-K. We argue this abstraction is mis-specified. The same attention score simultaneously decides two things: how much retained capacity each modality receives, and which tokens within a modality are kept. A shared top-K rule therefore inherits this audio-favoring allocation prior, spending retained capacity on audio before video tokens have a chance to compete. We propose Macer, a training-free compressor that first assigns explicit audio and video budgets, then performs allocation-normalized ranking within each modality at modality-specific shallow layers. Macer significantly reduces token cost while preserving accuracy across audio-grounded, audio--video joint, visual-dominant, and video-centric benchmarks. At 25 % retention, Macer preserves 98.7 % of full-token performance on Qwen2.5-Omni-7B and 97.3 % on Qwen2.5-Omni-3B. On Qwen2.5-Omni-7B, this 25 % setting reaches OmniZip-level performance at 45 % retention while using lower FLOPs. On OmniVinci-9B, the same allocation-before-ranking principle improves over shared top-K ranking by up to 12.9 points.

cs.AI↗

arXiv2026

Splaxel: Efficient Distributed Training of 3D Gaussian Splatting for Large-scale Scene Reconstruction via Pixel-level Communication

3D Gaussian Splatting (3DGS) enables high-fidelity and real-time 3D scene reconstruction, but scaling training to large-scale scenes requires optimizing hundreds of millions of Gaussians across multiple GPUs. Existing distributed approaches either partition scenes into isolated regions, causing global inconsistency, or rely on global Gaussian-level exchanges, which lead to substantial growth in inter-GPU communication and quickly dominate iteration time. We propose Splaxel, a communication-efficient distributed 3DGS training framework based on pixel-level local rendering and global composition. Instead of synchronizing Gaussians, each GPU renders its local subset and exchanges only partial pixel values, maintaining mathematical consistency while keeping communication cost stable as the scene size increases. Splaxel further reduces pixel-level redundancy through geometric and transmittance visibility prediction and improves GPU utilization via conflict-free camera-view consolidation. Evaluated on large-scale datasets with up to 120M Gaussians, Splaxel achieves up to 7.6$\times$ speedup over the state-of-the-art distributed 3DGS framework while preserving high reconstruction quality.

cs.DC↗

arXiv2026

Gaussians on a Diet: High-Quality Memory-Bounded 3D Gaussian Splatting Training

3D Gaussian Splatting (3DGS) has revolutionized novel view synthesis with high-quality rendering through continuous aggregations of millions of 3D Gaussian primitives. However, it suffers from a substantial memory footprint, particularly during training due to uncontrolled densification, posing a critical bottleneck for deployment on memory-constrained edge devices. While existing methods prune redundant Gaussians post-training, they fail to address the peak memory spikes caused by the abrupt growth of Gaussians early in the training process. To solve the training memory consumption problem, we propose a systematic memory-bounded training framework that dynamically optimizes Gaussians through iterative growth and pruning. In other words, the proposed framework alternates between incremental pruning of low-impact Gaussians and strategic growing of new primitives with an adaptive Gaussian compensation, maintaining a near-constant low memory usage while progressively refining rendering fidelity. We comprehensively evaluate the proposed training framework on various real-world datasets under strict memory constraints, showing significant improvements over existing state-of-the-art methods. Particularly, our proposed method practically enables memory-efficient 3DGS training on NVIDIA Jetson AGX Xavier, achieving similar visual quality with up to 80% lower peak training memory consumption than the original 3DGS.

cs.CV↗

arXiv2026

FlashMem: Supporting Modern DNN Workloads on Mobile with GPU Memory Hierarchy Optimizations

The increasing size and complexity of modern deep neural networks (DNNs) pose significant challenges for on-device inference on mobile GPUs, with limited memory and computational resources. Existing DNN acceleration frameworks primarily deploy a weight preloading strategy, where all model parameters are loaded into memory before execution on mobile GPUs. We posit that this approach is not adequate for modern DNN workloads that comprise very large model(s) and possibly execution of several distinct models in succession. In this work, we introduce FlashMem, a memory streaming framework designed to efficiently execute large-scale modern DNNs and multi-DNN workloads while minimizing memory consumption and reducing inference latency. Instead of fully preloading weights, FlashMem statically determines model loading schedules and dynamically streams them on demand, leveraging 2.5D texture memory to minimize data transformations and improve execution efficiency. Experimental results on 11 models demonstrate that FlashMem achieves 2.0x to 8.4x memory reduction and 1.7x to 75.0x speedup compared to existing frameworks, enabling efficient execution of large-scale models and multi-DNN support on resource-constrained mobile GPUs.

cs.DC↗

arXiv2026

CoSA: Compressed Sensing-Based Adaptation of Large Language Models

Parameter-Efficient Fine-Tuning (PEFT) has emerged as a practical paradigm for adapting large language models (LLMs) without updating all parameters. Most existing approaches, such as LoRA and PiSSA, rely on low-rank decompositions of weight updates. However, the low-rank assumption may restrict expressivity, particularly in task-specific adaptation scenarios where singular values are distributed relatively uniformly. To address this limitation, we propose CoSA (Compressed Sensing-Based Adaptation), a new PEFT method extended from compressed sensing theory. Instead of constraining weight updates to a low-rank subspace, CoSA expresses them through fixed random projection matrices and a compact learnable core. We provide a formal theoretical analysis of CoSA as a synthesis process, proving that weight updates can be compactly encoded into a low-dimensional space and mapped back through random projections. Extensive experimental results show that CoSA provides a principled perspective for efficient and expressive multi-scale model adaptation. Specifically, we evaluate CoSA on 10 diverse tasks, including natural language understanding and generation, employing 5 models of different scales from RoBERTa, Llama, and Qwen families. Across these settings, CoSA consistently matches or outperforms state-of-the-art PEFT methods.

cs.LG↗

arXiv2026

Turbo4DGen: Ultra-Fast Acceleration for 4D Generation

4D generation, or dynamic 3D content generation, integrates spatial, temporal, and view dimensions to model realistic dynamic scenes, playing a foundational role in advancing world models and physical AI. However, maintaining long-chain consistency across both frames and viewpoints through the unique spatio-camera-motion (SCM) attention mechanism introduces substantial computational and memory overhead, often leading to out-of-memory (OOM) failures and prohibitive generation times. To address these challenges, we propose Turbo4DGen, an ultra-fast acceleration framework for diffusion-based multi-view 4D content generation. Turbo4DGen introduces a spatiotemporal cache mechanism that persistently reuses intermediate attention across denoising steps, combined with dynamically semantic-aware attention pruning and an adaptive SCM chain bypass scheduler, to drastically reduce redundant SCM attention computation. Our experimental results show that Turbo4DGen achieves an average 9.7$\times$ speedup without quality degradation on the ObjaverseDy and Consistent4D datasets. To the best of our knowledge, Turbo4DGen is the first dedicated acceleration framework for 4D generation.

cs.GR↗

arXiv2026

Event-VStream: Event-Driven Real-Time Understanding for Long Video Streams

Real-time understanding of long video streams remains challenging for multimodal large language models (VLMs) due to redundant frame processing and rapid forgetting of past context. Existing streaming systems rely on fixed-interval decoding or cache pruning, which either produce repetitive outputs or discard crucial temporal information. We introduce Event-VStream, an event-aware framework that represents continuous video as a sequence of discrete, semantically coherent events. Our system detects meaningful state transitions by integrating motion, semantic, and predictive cues, and triggers language generation only at those boundaries. Each event embedding is consolidated into a persistent memory bank, enabling long-horizon reasoning while maintaining low latency. Across OVOBench-Realtime, and long-form Ego4D evaluations, Event-VStream achieves competitive performance. It improves over a VideoLLM-Online-8B baseline by +10.4 points on OVOBench-Realtime, achieves performance close to Flash-VStream-7B despite using only a general-purpose LLaMA-3-8B text backbone, and maintains around 70% GPT-5 win rate on 2-hour Ego4D streams.

cs.CV↗

arXiv2025

AdaRing: Towards Ultra-Light Vision-Language Adaptation via Cross-Layer Tensor Ring Decomposition

Adapter-based fine-tuning has gained remarkable attention in adapting large pre-trained vision language models (VLMs) for a wide range of downstream tasks efficiently. In this paradigm, only the inserted adapters are fine-tuned, without the need for training the original VLM backbone. Existing works scale adapters by integrating them into every layer of VLMs to increase the capacity of adapters. However, these methods face two primary limitations: 1) limited compression rate due to ignoring cross-layer redundancy, and 2) limited representational capacity across homogeneous adapters. In this paper, we propose a novel vision-language fine-tuning framework based on cross-layer tensor ring decomposition (TRD) with the integration and collaboration of diverse adapters, called AdaRing, achieving ultra-light parameter-efficient adaptation of VLMs on various tasks. To remove the high redundancy that exists among adapters across layers, we exploit the tensor-level low-rankness to formulate adapters as layer-shared tensor cores and layer-specific slices. Moreover, guided by generalization-aware fine-tuning, diverse rank-driven adapters cooperate to handle tasks that require different representations. Our experiments show that the proposed AdaRing achieves the state-of-the-art performance while reducing average training parameters by 90%.

cs.CV↗

arXiv2025

FLARE: A Dataflow-Aware and Scalable Hardware Architecture for Neural-Hybrid Scientific Lossy Compression

Scientific simulation leveraging high-performance computing (HPC) systems is crucial for modeling complex systems and phenomena in fields such as astrophysics, climate science, and fluid dynamics, generating massive datasets that often reach petabyte to exabyte scales. However, managing these vast data volumes introduces significant I/O and network bottlenecks, limiting practical performance and scalability. While cutting-edge lossy compression frameworks powered by deep neural networks (DNNs) have demonstrated superior compression ratios by capturing complex data correlations, their integration into HPC workflows poses substantial challenges due to the hybrid non-neural and neural computation patterns, causing excessive memory access overhead, large sequential stalls, and limited adaptability to varying data sizes and workloads in existing hardware platforms. To overcome these challenges and push the limit of high-performance scientific computing, we for the first time propose FLARE, a dataflow-aware and scalable hardware architecture for neural-hybrid scientific lossy compression. FLARE minimizes off-chip data access, reduces bubble overhead through efficient dataflow, and adopts a modular design that provides both scalability and flexibility, significantly enhancing throughput and energy efficiency on modern HPC systems. Particularly, the proposed FLARE achieves runtime speedups ranging from $3.50 \times$ to $96.07 \times$, and energy efficiency improvements ranging from $24.51 \times$ to $520.68 \times$, across various datasets and hardware platforms.

cs.DC↗

arXiv2024

AdaCM$^2$: On Understanding Extremely Long-Term Video with Adaptive Cross-Modality Memory Reduction

The advancements in large language models (LLMs) have propelled the improvement of video understanding tasks by incorporating LLMs with visual models. However, most existing LLM-based models (e.g., VideoLLaMA, VideoChat) are constrained to processing short-duration videos. Recent attempts to understand long-term videos by extracting and compressing visual features into a fixed memory size. Nevertheless, those methods leverage only visual modality to merge video tokens and overlook the correlation between visual and textual queries, leading to difficulties in effectively handling complex question-answering tasks. To address the challenges of long videos and complex prompts, we propose AdaCM$^2$, which, for the first time, introduces an adaptive cross-modality memory reduction approach to video-text alignment in an auto-regressive manner on video streams. Our extensive experiments on various video understanding tasks, such as video captioning, video question answering, and video classification, demonstrate that AdaCM$^2$ achieves state-of-the-art performance across multiple datasets while significantly reducing memory usage. Notably, it achieves a 4.5% improvement across multiple tasks in the LVU dataset with a GPU memory consumption reduction of up to 65%.

cs.CV↗

arXiv2024

GaussianSpa: An "Optimizing-Sparsifying" Simplification Framework for Compact and High-Quality 3D Gaussian Splatting

3D Gaussian Splatting (3DGS) has emerged as a mainstream for novel view synthesis, leveraging continuous aggregations of Gaussian functions to model scene geometry. However, 3DGS suffers from substantial memory requirements to store the multitude of Gaussians, hindering its practicality. To address this challenge, we introduce GaussianSpa, an optimization-based simplification framework for compact and high-quality 3DGS. Specifically, we formulate the simplification as an optimization problem associated with the 3DGS training. Correspondingly, we propose an efficient "optimizing-sparsifying" solution that alternately solves two independent sub-problems, gradually imposing strong sparsity onto the Gaussians in the training process. Our comprehensive evaluations on various datasets show the superiority of GaussianSpa over existing state-of-the-art approaches. Notably, GaussianSpa achieves an average PSNR improvement of 0.9 dB on the real-world Deep Blending dataset with 10$\times$ fewer Gaussians compared to the vanilla 3DGS. Our project page is available at https://noodle-lab.github.io/gaussianspa/.

cs.CV↗

arXiv2024

MoE-I$^2$: Compressing Mixture of Experts Models through Inter-Expert Pruning and Intra-Expert Low-Rank Decomposition

The emergence of Mixture of Experts (MoE) LLMs has significantly advanced the development of language models. Compared to traditional LLMs, MoE LLMs outperform traditional LLMs by achieving higher performance with considerably fewer activated parameters. Despite this efficiency, their enormous parameter size still leads to high deployment costs. In this paper, we introduce a two-stage compression method tailored for MoE to reduce the model size and decrease the computational cost. First, in the inter-expert pruning stage, we analyze the importance of each layer and propose the Layer-wise Genetic Search and Block-wise KT-Reception Field with the non-uniform pruning ratio to prune the individual expert. Second, in the intra-expert decomposition stage, we apply the low-rank decomposition to further compress the parameters within the remaining experts. Extensive experiments on Qwen1.5-MoE-A2.7B, DeepSeek-V2-Lite, and Mixtral-8$\times$7B demonstrate that our proposed methods can both reduce the model size and enhance inference efficiency while maintaining performance in various zero-shot tasks. The code will be available at \url{https://github.com/xiaochengsky/MoEI-2.git}

cs.LG↗

arXiv2024

Enhancing Lossy Compression Through Cross-Field Information for Scientific Applications

Lossy compression is one of the most effective methods for reducing the size of scientific data containing multiple data fields. It reduces information density through prediction or transformation techniques to compress the data. Previous approaches use local information from a single target field when predicting target data points, limiting their potential to achieve higher compression ratios. In this paper, we identified significant cross-field correlations within scientific datasets. We propose a novel hybrid prediction model that utilizes CNN to extract cross-field information and combine it with existing local field information. Our solution enhances the prediction accuracy of lossy compressors, leading to improved compression ratios without compromising data quality. We evaluate our solution on three scientific datasets, demonstrating its ability to improve compression ratios by up to 25% under specific error bounds. Additionally, our solution preserves more data details and reduces artifacts compared to baseline approaches.

cs.LG↗

arXiv2024

NeurLZ: An Online Neural Learning-Based Method to Enhance Scientific Lossy Compression

Large-scale scientific simulations generate massive datasets, posing challenges for storage and I/O. Traditional lossy compression struggles to advance more in balancing compression ratio, data quality, and adaptability to diverse scientific data features. While deep learning-based solutions have been explored, their common practice of relying on large models and offline training limits adaptability to dynamic data characteristics and computational efficiency. To address these challenges, we propose NeurLZ, a neural method designed to enhance lossy compression by integrating online learning, cross-field learning, and robust error regulation. Key innovations of NeurLZ include: (1) compression-time online neural learning with lightweight skipping DNN models, adapting to residual errors without costly offline pertaining, (2) the error-mitigating capability, recovering fine details from compression errors overlooked by conventional compressors, (3) $1\times$ and $2\times$ error-regulation modes, ensuring strict adherence to $1\times$ user-input error bounds strictly or relaxed 2$\times$ bounds for better overall quality, and (4) cross-field learning leveraging inter-field correlations in scientific data to improve conventional methods. Comprehensive evaluations on representative HPC datasets, e.g., Nyx, Miranda, Hurricane, against state-of-the-art compressors show NeurLZ's effectiveness. During the first five learning epochs, NeurLZ achieves an 89% bit rate reduction, with further optimization yielding up to around 94% reduction at equivalent distortion, significantly outperforming existing methods, demonstrating NeurLZ's superior performance in enhancing scientific lossy compression as a scalable and efficient solution.

cs.DC↗

arXiv2024

SmartMem: Layout Transformation Elimination and Adaptation for Efficient DNN Execution on Mobile

This work is motivated by recent developments in Deep Neural Networks, particularly the Transformer architectures underlying applications such as ChatGPT, and the need for performing inference on mobile devices. Focusing on emerging transformers (specifically the ones with computationally efficient Swin-like architectures) and large models (e.g., Stable Diffusion and LLMs) based on transformers, we observe that layout transformations between the computational operators cause a significant slowdown in these applications. This paper presents SmartMem, a comprehensive framework for eliminating most layout transformations, with the idea that multiple operators can use the same tensor layout through careful choice of layout and implementation of operations. Our approach is based on classifying the operators into four groups, and considering combinations of producer-consumer edges between the operators. We develop a set of methods for searching such layouts. Another component of our work is developing efficient memory layouts for 2.5 dimensional memory commonly seen in mobile devices. Our experimental results show that SmartMem outperforms 5 state-of-the-art DNN execution frameworks on mobile devices across 18 varied neural networks, including CNNs, Transformers with both local and global attention, as well as LLMs. In particular, compared to DNNFusion, SmartMem achieves an average speedup of 2.8$\times$, and outperforms TVM and MNN with speedups of 6.9$\times$ and 7.9$\times$, respectively, on average.

cs.LG↗

arXiv2024

GWLZ: A Group-wise Learning-based Lossy Compression Framework for Scientific Data

The rapid expansion of computational capabilities and the ever-growing scale of modern HPC systems present formidable challenges in managing exascale scientific data. Faced with such vast datasets, traditional lossless compression techniques prove insufficient in reducing data size to a manageable level while preserving all information intact. In response, researchers have turned to error-bounded lossy compression methods, which offer a balance between data size reduction and information retention. However, despite their utility, these compressors employing conventional techniques struggle with limited reconstruction quality. To address this issue, we draw inspiration from recent advancements in deep learning and propose GWLZ, a novel group-wise learning-based lossy compression framework with multiple lightweight learnable enhancer models. Leveraging a group of neural networks, GWLZ significantly enhances the decompressed data reconstruction quality with negligible impact on the compression efficiency. Experimental results on different fields from the Nyx dataset demonstrate remarkable improvements by GWLZ, achieving up to 20% quality enhancements with negligible overhead as low as 0.0003x.

cs.DC↗

arXiv2024

ELRT: Efficient Low-Rank Training for Compact Convolutional Neural Networks

Low-rank compression, a popular model compression technique that produces compact convolutional neural networks (CNNs) with low rankness, has been well-studied in the literature. On the other hand, low-rank training, as an alternative way to train low-rank CNNs from scratch, has been exploited little yet. Unlike low-rank compression, low-rank training does not need pre-trained full-rank models, and the entire training phase is always performed on the low-rank structure, bringing attractive benefits for practical applications. However, the existing low-rank training solutions still face several challenges, such as a considerable accuracy drop and/or still needing to update full-size models during the training. In this paper, we perform a systematic investigation on low-rank CNN training. By identifying the proper low-rank format and performance-improving strategy, we propose ELRT, an efficient low-rank training solution for high-accuracy, high-compactness, low-rank CNN models. Our extensive evaluation results for training various CNNs on different datasets demonstrate the effectiveness of ELRT.

cs.CV↗

arXiv2023

COMCAT: Towards Efficient Compression and Customization of Attention-Based Vision Models

Attention-based vision models, such as Vision Transformer (ViT) and its variants, have shown promising performance in various computer vision tasks. However, these emerging architectures suffer from large model sizes and high computational costs, calling for efficient model compression solutions. To date, pruning ViTs has been well studied, while other compression strategies that have been widely applied in CNN compression, e.g., model factorization, is little explored in the context of ViT compression. This paper explores an efficient method for compressing vision transformers to enrich the toolset for obtaining compact attention-based vision models. Based on the new insight on the multi-head attention layer, we develop a highly efficient ViT compression solution, which outperforms the state-of-the-art pruning methods. For compressing DeiT-small and DeiT-base models on ImageNet, our proposed approach can achieve 0.45% and 0.76% higher top-1 accuracy even with fewer parameters. Our finding can also be applied to improve the customization efficiency of text-to-image diffusion models, with much faster training (up to $2.6\times$ speedup) and lower extra storage cost (up to $1927.5\times$ reduction) than the existing works.

cs.CV↗