arXiv subjects

Yuan Du

Publications and source records attributed to Yuan Du.

At least 19 recordsLinked to original sources

arXiv2026

LiftQuant: Continuous Bit-Width LLM via Dimensional Lifting and Projection

Existing quantization methods are fundamentally limited by rigid, integer-based bit-widths (e.g., 2, 3-bit), resulting in a ``deployment gap" where Large Language Models cannot be optimally fitted to specific memory budgets. To bridge this gap, we introduce LiftQuant, a novel framework that enables continuous bit-width control for true Pareto-optimal deployment. The core innovation is a ``lift-then-project" mechanism which approximates low-dimensional weight vectors by projecting a simple 1-bit lattice from a higher-dimensional ``lifted" space. Crucially, the effective bit-width is determined simply by the ratio of the lifted dimension to the original dimension, which allows the bit-width to be tuned quasi-continuous as the dimension is a flexible structural parameter. This projection generates a structured yet non-uniform codebook, capturing the expressive power of Vector Quantization (VQ). While beneficial over VQ, LiftQuant's decoding path relies solely on linear transformations and 1-bit uniform quantizers, retaining hardware-friendly nature. This flexibility is transformative: LiftQuant enables a 70B LLM to be compressed to 2.4 bits to precisely fit a 24GB GPU, where its performance significantly surpasses state-of-the-art 2-bit models fitted on the same device. Our code and ckpt is available at https://github.com/Heliulu/LiftQuant.

cs.LG↗

arXiv2026

MoASE++: Mixture of Activation Sparsity Experts with Domain-Adaptive On-policy Distillation for Continual Test Time Adaptation

Continual test-time adaptation adapts a source-pretrained model to non-stationary, unlabeled target streams while retaining past competence, yet texture-biased backbones risk error accumulation and catastrophic forgetting. Drawing inspiration from the process of decoupling shape and texture in the human visual system, we introduce MoASE, a plug-in mixture-of-experts that disentangles domain-agnostic structure from domain-specific texture using Activation Sparsity Experts with Spatial Differentiable Dropout, forming complementary high- and low-activation pathways, while high- and low-rank bottlenecks diversify representations. The Activation Sparsity Gate produces input-adaptive SDD thresholds for precise token selection, and the Domain-Aware Router assigns per-sample expert weights using texture-sensitive cues. To curb confirmation bias on unlabeled streams and stabilize supervision, we then introduce Domain-Adaptive On-Policy Distillation to constitute MoASE++, with an EMA-anchored on-policy reverse KL distillation and an augmentation policy conditioned on entropy and confidence that aligns predictions across the same views and improves the robustness-plasticity balance. Extensive experiments on classification (CIFAR-10/100-C, ImageNet-C) and semantic segmentation (Cityscapes->ACDC) demonstrate consistent state-of-the-art performance, offering a principled, controllable approach to continual adaptation in dynamic visual environments.

cs.CV↗

arXiv2026

Memory Efficient Full-gradient Attacks (MEFA) Framework for Adversarial Defense Evaluations

This work studies the robust evaluation of iterative stochastic purification defenses under white-box adversarial attacks. Our key technical insight is that gradient checkpointing makes exact end-to-end gradient computation through long purification trajectories practical by trading additional recomputation for substantially lower memory usage. This enables full-gradient adaptive attacks against diffusion- and Langevin-based purification defenses, where prior evaluations often resort to approximate backpropagation due to memory constraints. These approximations can weaken the attack signal and risk overestimating robustness. In parallel, stochasticity in iterative purification is frequently under-controlled, even though different purification trajectories can substantially change reported robustness metrics. Building on this insight, we introduce a memory-efficient full-gradient evaluation framework for stochastic purification defenses. The framework combines checkpointed backpropagation with evaluation protocols that control stochastic variability, thereby reducing memory bottlenecks while preserving exact gradients. We evaluate diffusion-based purification and Langevin sampling with Energy-Based Models (EBMs), demonstrating that full-gradient attacks uncover vulnerabilities missed by approximate-gradient evaluations. Our framework yields stronger state-of-the-art $\ell_{\infty}$ and $\ell_{2}$ white-box attacks and further supports probing out-of-distribution robustness. Overall, our results show that exact-gradient evaluation is essential for reliable benchmarking of iterative stochastic defenses.

cs.LG↗

arXiv2026

A Switch-Centric In-Network Architecture for Accelerating LLM Inference in Shared-Memory Network

Tensor parallelism (TP) has become a key technique for latency-sensitive LLM inference, but it introduces frequent, tightly synchronized All-Reduce operations that lie directly on the inference critical path, making inter-GPU communication a major performance bottleneck. In-network computing offers a promising approach to mitigating this bottleneck. However, we observe that the existing solution, NVLink SHARP (NVLS), is fundamentally constrained by its GPU-driven execution model that exposes in-network computing as GPU-issued, element-granular memory operations rather than a switch-orchestrated collective. This design forces redundant switch-GPU-switch data movement and cannot perform scale-aware quantized All-Reduce due to its lack of cross-element coordination. To address these limitations, we propose SCIN, the first switch-centric in-network computing architecture for multi-accelerator shared-memory networks. SCIN replaces GPU-driven collective orchestration with an in-switch accelerator (ISA) that directly initiates memory transactions and orchestrates the complete All-Reduce operation. This switch-driven execution eliminates redundant data movement inherent in NVLS, shortens synchronization paths, and fully offloads collective execution from GPUs. Furthermore, centralized ISA control enables cross-element coordination between quantized data and their scale factors. SCIN exploits this capability through a pipelined dequantization-reduction-requantization datapath, enabling scale-aware in-network quantized All-Reduce and nearly halving communication volume with negligible accuracy degradation. We implement SCIN on a multi-FPGA prototype to validate its feasibility. For a simulated 8-GPU system, SCIN accelerates small- and large-message All-Reduce by up to 1.8x and 2.6x over NVLS, respectively, translating into up to 1.12x TPOT and 1.42x TTFT speedups.

cs.AR↗

arXiv2026

Key-Embedded Privacy for Decentralized AI in Biomedical Omics

The rapid adoption of data-driven methods in biomedicine has intensified concerns over privacy, governance, and regulation, limiting raw data sharing and hindering the assembly of representative cohorts for clinically relevant AI. This landscape necessitates practical, efficient privacy solutions, as cryptographic defenses often impose heavy overhead and differential privacy can degrade performance, leading to sub-optimal outcomes in real-world settings. Here, we present a lightweight federated learning method, INFL, based on Implicit Neural Representations that addresses these challenges. Our approach integrates plug-and-play, coordinate-conditioned modules into client models, embeds a secret key directly into the architecture, and supports seamless aggregation across heterogeneous sites. Across diverse biomedical omics tasks, including cohort-scale classification in bulk proteomics, regression for perturbation prediction in single-cell transcriptomics, and clustering in spatial transcriptomics and multi-omics with both public and private data, we demonstrate that INFL achieves strong, controllable privacy while maintaining utility, preserving the performance necessary for downstream scientific and clinical applications.

cs.LG↗

arXiv2026

A Scheduling Framework for Efficient MoE Inference on Edge GPU-NDP Systems

Mixture-of-Experts (MoE) models facilitate edge deployment by decoupling model capacity from active computation, yet their large memory footprint drives the need for GPU systems with near-data processing (NDP) capabilities that offload experts to dedicated processing units. However, deploying MoE models on such edge-based GPU-NDP systems faces three critical challenges: 1) severe load imbalance across NDP units due to non-uniform expert selection and expert parallelism, 2) insufficient GPU utilization during expert computation within NDP units, and 3) extensive data pre-profiling necessitated by unpredictable expert activation patterns for pre-fetching. To address these challenges, this paper proposes an efficient inference framework featuring three key optimizations. First, the underexplored tensor parallelism in MoE inference is exploited to partition and compute large expert parameters across multiple NDP units simultaneously towards edge low-batch scenarios. Second, a load-balancing-aware scheduling algorithm distributes expert computations across NDP units and GPU to maximize resource utilization. Third, a dataset-free pre-fetching strategy proactively loads frequently accessed experts to minimize activation delays. Experimental results show that our framework enables GPU-NDP systems to achieve 2.41x on average and up to 2.56x speedup in end-to-end latency compared to state-of-the-art approaches, significantly enhancing MoE inference efficiency in resource-constrained environments.

cs.DC↗

arXiv2025

RepCaM++: Exploring Transparent Visual Prompt With Inference-Time Re-Parameterization for Neural Video Delivery

Recently, content-aware methods have been employed to reduce bandwidth and enhance the quality of Internet video delivery. These methods involve training distinct content-aware super-resolution (SR) models for each video chunk on the server, subsequently streaming the low-resolution (LR) video chunks with the SR models to the client. Prior research has incorporated additional partial parameters to customize the models for individual video chunks. However, this leads to parameter accumulation and can fail to adapt appropriately as video lengths increase, resulting in increased delivery costs and reduced performance. In this paper, we introduce RepCaM++, an innovative framework based on a novel Re-parameterization Content-aware Modulation (RepCaM) module that uniformly modulates video chunks. The RepCaM framework integrates extra parallel-cascade parameters during training to accommodate multiple chunks, subsequently eliminating these additional parameters through re-parameterization during inference. Furthermore, to enhance RepCaM's performance, we propose the Transparent Visual Prompt (TVP), which includes a minimal set of zero-initialized image-level parameters (e.g., less than 0.1%) to capture fine details within video chunks. We conduct extensive experiments on the VSD4K dataset, encompassing six different video scenes, and achieve state-of-the-art results in video restoration quality and delivery bandwidth compression.

cs.NI↗

arXiv2025

BEVUDA++: Geometric-aware Unsupervised Domain Adaptation for Multi-View 3D Object Detection

Vision-centric Bird's Eye View (BEV) perception holds considerable promise for autonomous driving. Recent studies have prioritized efficiency or accuracy enhancements, yet the issue of domain shift has been overlooked, leading to substantial performance degradation upon transfer. We identify major domain gaps in real-world cross-domain scenarios and initiate the first effort to address the Domain Adaptation (DA) challenge in multi-view 3D object detection for BEV perception. Given the complexity of BEV perception approaches with their multiple components, domain shift accumulation across multi-geometric spaces (e.g., 2D, 3D Voxel, BEV) poses a significant challenge for BEV domain adaptation. In this paper, we introduce an innovative geometric-aware teacher-student framework, BEVUDA++, to diminish this issue, comprising a Reliable Depth Teacher (RDT) and a Geometric Consistent Student (GCS) model. Specifically, RDT effectively blends target LiDAR with dependable depth predictions to generate depth-aware information based on uncertainty estimation, enhancing the extraction of Voxel and BEV features that are essential for understanding the target domain. To collaboratively reduce the domain shift, GCS maps features from multiple spaces into a unified geometric embedding space, thereby narrowing the gap in data distribution between the two domains. Additionally, we introduce a novel Uncertainty-guided Exponential Moving Average (UEMA) to further reduce error accumulation due to domain shifts informed by previously obtained uncertainty guidance. To demonstrate the superiority of our proposed method, we execute comprehensive experiments in four cross-domain scenarios, securing state-of-the-art performance in BEV 3D object detection tasks, e.g., 12.9\% NDS and 9.5\% mAP enhancement on Day-Night adaptation.

cs.CV↗

arXiv2025

SQAP-VLA: A Synergistic Quantization-Aware Pruning Framework for High-Performance Vision-Language-Action Models

Vision-Language-Action (VLA) models exhibit unprecedented capabilities for embodied intelligence. However, their extensive computational and memory costs hinder their practical deployment. Existing VLA compression and acceleration approaches conduct quantization or token pruning in an ad-hoc manner but fail to enable both for a holistic efficiency improvement due to an observed incompatibility. This work introduces SQAP-VLA, the first structured, training-free VLA inference acceleration framework that simultaneously enables state-of-the-art quantization and token pruning. We overcome the incompatibility by co-designing the quantization and token pruning pipeline, where we propose new quantization-aware token pruning criteria that work on an aggressively quantized model while improving the quantizer design to enhance pruning effectiveness. When applied to standard VLA models, SQAP-VLA yields significant gains in computational efficiency and inference speed while successfully preserving core model performance, achieving a $\times$1.93 speedup and up to a 4.5\% average success rate enhancement compared to the original model.

cs.CV↗

arXiv2025

AnalogTester: A Large Language Model-Based Framework for Automatic Testbench Generation in Analog Circuit Design

Recent advancements have demonstrated the significant potential of large language models (LLMs) in analog circuit design. Nevertheless, testbench construction for analog circuits remains manual, creating a critical bottleneck in achieving fully automated design processes. Particularly when replicating circuit designs from academic papers, manual Testbench construction demands time-intensive implementation and frequent adjustments, which fails to address the dynamic diversity and flexibility requirements for automation. AnalogTester tackles automated analog design challenges through an LLM-powered pipeline: a) domain-knowledge integration, b) paper information extraction, c) simulation scheme synthesis, and d) testbench code generation with Tsinghua Electronic Design (TED). AnalogTester has demonstrated automated Testbench generation capabilities for three fundamental analog circuit types: operational amplifiers (op-amps), bandgap references (BGRs), and low-dropout regulators (LDOs), while maintaining a scalable framework for adaptation to broader circuit topologies. Furthermore, AnalogTester can generate circuit knowledge data and TED code corpus, establishing fundamental training datasets for LLM specialization in analog circuit design automation.

cs.MA↗

arXiv2025

DiffCkt: A Diffusion Model-Based Hybrid Neural Network Framework for Automatic Transistor-Level Generation of Analog Circuits

Analog circuit design consists of the pre-layout and layout phases. Among them, the pre-layout phase directly decides the final circuit performance, but heavily depends on experienced engineers to do manual design according to specific application scenarios. To overcome these challenges and automate the analog circuit pre-layout design phase, we introduce DiffCkt: a diffusion model-based hybrid neural network framework for the automatic transistor-level generation of analog circuits, which can directly generate corresponding circuit structures and device parameters tailored to specific performance requirements. To more accurately quantify the efficiency of circuits generated by DiffCkt, we introduce the Circuit Generation Efficiency Index (CGEI), which is determined by both the figure of merit (FOM) of a single generated circuit and the time consumed. Compared with relative research, DiffCkt has improved CGEI by a factor of $2.21 \sim 8365\times$, reaching a state-of-the-art (SOTA) level. In conclusion, this work shows that the diffusion model has the remarkable ability to learn and generate analog circuit structures and device parameters, providing a revolutionary method for automating the pre-layout design of analog circuits. The circuit dataset will be open source, its preview version is available at https://github.com/CjLiu-NJU/DiffCkt.

cs.ET↗

arXiv2025

Image2Net: Datasets, Benchmark and Hybrid Framework to Convert Analog Circuit Diagrams into Netlists

Large Language Model (LLM) exhibits great potential in designing of analog integrated circuits (IC) because of its excellence in abstraction and generalization for knowledge. However, further development of LLM-based analog ICs heavily relies on textual description of analog ICs, while existing analog ICs are mostly illustrated in image-based circuit diagrams rather than text-based netlists. Converting circuit diagrams to netlists help LLMs to enrich the knowledge of analog IC. Nevertheless, previously proposed conversion frameworks face challenges in further application because of limited support of image styles and circuit elements. Up to now, it still remains a challenging task to effectively convert complex circuit diagrams into netlists. To this end, this paper constructs and opensources a new dataset with rich styles of circuit diagrams as well as balanced distribution of simple and complex analog ICs. And a hybrid framework, named Image2Net, is proposed for practical conversion from circuit diagrams to netlists. The netlist edit distance (NED) is also introduced to precisely assess the difference between the converted netlists and ground truth. Based on our benchmark, Image2Net achieves 80.77% successful rate, which is 34.62%-45.19% higher than previous works. Specifically, the proposed work shows 0.116 averaged NED, which is 62.1%-69.6% lower than state-of-the-arts. Our datasets and benchmark are available at https://github.com/LAD021/ci2n_datasets.

cs.AR↗

arXiv2025

A Large Language Model-based Multi-Agent Framework for Analog Circuits' Sizing Relationships Extraction

In the design process of the analog circuit pre-layout phase, device sizing is an important step in determining whether an analog circuit can meet the required performance metrics. Many existing techniques extract the circuit sizing task as a mathematical optimization problem to solve and continuously improve the optimization efficiency from a mathematical perspective. But they ignore the automatic introduction of prior knowledge, fail to achieve effective pruning of the search space, which thereby leads to a considerable compression margin remaining in the search space. To alleviate this problem, we propose a large language model (LLM)-based multi-agent framework for analog circuits' sizing relationships extraction from academic papers. The search space in the sizing process can be effectively pruned based on the sizing relationship extracted by this framework. Eventually, we conducted tests on 3 types of circuits, and the optimization efficiency was improved by $2.32 \sim 26.6 \times$. This work demonstrates that the LLM can effectively prune the search space for analog circuit sizing, providing a new solution for the combination of LLMs and conventional analog circuit design automation methods.

cs.AI↗

arXiv2025

BASE-Q: Bias and Asymmetric Scaling Enhanced Rotational Quantization for Large Language Models

Rotations have become essential to state-of-the-art quantization pipelines for large language models (LLMs) by effectively smoothing outliers in weights and activations. However, further optimizing the rotation parameters offers only limited performance gains and introduces significant training overhead: due to rotation parameter sharing, full-model must be loaded simultaneously to enable backpropagation, resulting in substantial memory consumption and limited practical utility. In this work, we identify two fundamental limitations of current rotational quantization methods: (i) rotation fails to align channel means, resulting in wider quantization bounds and increased rounding errors; and (ii) rotation makes the activation distribution more Gaussian-like, increasing energy loss caused by clipping errors. To address these issues, we introduce \textbf{BASE-Q}, a simple yet powerful approach that combines bias correction and asymmetric scaling to effectively reduce rounding and clipping errors. Furthermore, BASE-Q enables blockwise optimization, eliminating the need for memory-intensive full-model backpropagation. Extensive experiments on various LLMs and benchmarks demonstrate the effectiveness of BASE-Q, narrowing the accuracy gap to full-precision models by 50.5\%, 42.9\%, and 29.2\% compared to QuaRot, SpinQuant, and OSTQuant, respectively. The code will be released soon.

cs.LG↗

arXiv2025

MoLe-VLA: Dynamic Layer-skipping Vision Language Action Model via Mixture-of-Layers for Efficient Robot Manipulation

Multimodal Large Language Models (MLLMs) excel in understanding complex language and visual data, enabling generalist robotic systems to interpret instructions and perform embodied tasks. Nevertheless, their real-world deployment is hindered by substantial computational and storage demands. Recent insights into the homogeneous patterns in the LLM layer have inspired sparsification techniques to address these challenges, such as early exit and token pruning. However, these methods often neglect the critical role of the final layers that encode the semantic information most relevant to downstream robotic tasks. Aligning with the recent breakthrough of the Shallow Brain Hypothesis (SBH) in neuroscience and the mixture of experts in model sparsification, we conceptualize each LLM layer as an expert and propose a Mixture-of-Layers Vision-Language-Action model (MoLe-VLA, or simply MoLe) architecture for dynamic LLM layer activation. We introduce a Spatial-Temporal Aware Router (STAR) for MoLe to selectively activate only parts of the layers based on the robot's current state, mimicking the brain's distinct signal pathways specialized for cognition and causal reasoning. Additionally, to compensate for the cognitive ability of LLMs lost in MoLe, we devise a Cognition Self-Knowledge Distillation (CogKD) framework. CogKD enhances the understanding of task demands and improves the generation of task-relevant action sequences by leveraging cognitive features. Extensive experiments conducted in both RLBench simulation and real-world environments demonstrate the superiority of MoLe-VLA in both efficiency and performance. Specifically, MoLe-VLA achieves an 8% improvement in the mean success rate across ten tasks while reducing computational costs by up to x5.6 compared to standard LLMs.

cs.RO↗

arXiv2025

FBQuant: FeedBack Quantization for Large Language Models

Deploying Large Language Models (LLMs) on edge devices is increasingly important, as it eliminates reliance on network connections, reduces expensive API calls, and enhances user privacy. However, on-device deployment is challenging due to the limited computational resources of edge devices. In particular, the key bottleneck stems from memory bandwidth constraints related to weight loading. Weight-only quantization effectively reduces memory access, yet often induces significant accuracy degradation. Recent efforts to incorporate sub-branches have shown promise for mitigating quantization errors, but these methods either lack robust optimization strategies or rely on suboptimal objectives. To address these gaps, we propose FeedBack Quantization (FBQuant), a novel approach inspired by negative feedback mechanisms in automatic control. FBQuant inherently ensures that the reconstructed weights remain bounded by the quantization process, thereby reducing the risk of overfitting. To further offset the additional latency introduced by sub-branches, we develop an efficient CUDA kernel that decreases 60% of extra inference time. Comprehensive experiments demonstrate the efficiency and effectiveness of FBQuant across various LLMs. Notably, for 3-bit Llama2-7B, FBQuant improves zero-shot accuracy by 1.2%.

cs.LG↗

arXiv2024

AmpAgent: An LLM-based Multi-Agent System for Multi-stage Amplifier Schematic Design from Literature for Process and Performance Porting

Multi-stage amplifiers are widely applied in analog circuits. However, their large number of components, complex transfer functions, and intricate pole-zero distributions necessitate extensive manpower for derivation and param sizing to ensure their stability. In order to achieve efficient derivation of the transfer function and simplify the difficulty of circuit design, we propose AmpAgent: a multi-agent system based on large language models (LLMs) for efficiently designing such complex amplifiers from literature with process and performance porting. AmpAgent is composed of three agents: Literature Analysis Agent, Mathematics Reasoning Agent and Device Sizing Agent. They are separately responsible for retrieving key information (e.g. formulas and transfer functions) from the literature, decompose the whole circuit's design problem by deriving the key formulas, and address the decomposed problem iteratively. AmpAgent was employed in the schematic design of seven types of multi-stage amplifiers with different compensation techniques. In terms of design efficiency, AmpAgent has reduced the number of iterations by 1.32$ \sim $4${\times}$ and execution time by 1.19$ \sim $2.99${\times}$ compared to conventional optimization algorithms, with a success rate increased by 1.03$ \sim $6.79${\times}$. In terms of circuit performance, it has improved by 1.63$ \sim $27.25${\times}$ compared to the original literature. The findings suggest that LLMs could play a crucial role in the field of complex analog circuit schematic design, as well as process and performance porting.

cs.ET↗

arXiv2024

PAT: Pruning-Aware Tuning for Large Language Models

Large language models (LLMs) excel in language tasks, especially with supervised fine-tuning after pre-training. However, their substantial memory and computational requirements hinder practical applications. Structural pruning, which reduces less significant weight dimensions, is one solution. Yet, traditional post-hoc pruning often leads to significant performance loss, with limited recovery from further fine-tuning due to reduced capacity. Since the model fine-tuning refines the general and chaotic knowledge in pre-trained models, we aim to incorporate structural pruning with the fine-tuning, and propose the Pruning-Aware Tuning (PAT) paradigm to eliminate model redundancy while preserving the model performance to the maximum extend. Specifically, we insert the innovative Hybrid Sparsification Modules (HSMs) between the Attention and FFN components to accordingly sparsify the upstream and downstream linear modules. The HSM comprises a lightweight operator and a globally shared trainable mask. The lightweight operator maintains a training overhead comparable to that of LoRA, while the trainable mask unifies the channels to be sparsified, ensuring structural pruning. Additionally, we propose the Identity Loss which decouples the transformation and scaling properties of the HSMs to enhance training robustness. Extensive experiments demonstrate that PAT excels in both performance and efficiency. For example, our Llama2-7b model with a 25\% pruning ratio achieves 1.33$\times$ speedup while outperforming the LoRA-finetuned model by up to 1.26\% in accuracy with a similar training cost. Code: https://github.com/kriskrisliu/PAT_Pruning-Aware-Tuning

cs.LG↗