Maoliang Li

Zihao Zheng, Zhihao Mao, Sicheng Tian et al.

Vision-Language-Action (VLA) Models have become the mainstream solution for robot control, but suffer from slow inference speeds. Speculative Decoding (SD) is a promising acceleration method which can be divided into two categories: drafter-based SD and retrieval-based SD. Existing methods fail to analyze the advantages and disadvantages of these two types of SD in VLA models, leading to their sole application or optimization. In this paper, we analyze the trajectory patterns of robots controlled by the VLA model and derive a key insight: the two types of SD should be used in a hybrid manner. However, achieving hybrid SD in VLA models poses several challenges: (1) draft rejection and persistent errors in retrieval-based SD; (2) difficulty in determining the hybrid boundary. To address these, we propose the HeiSD framework. We propose a retrieval-based SD optimization method in HeiSD,which contains a verify-skip mechanism and a sequence-wise relaxed acceptance strategy. Moreover, we proposed a kinematic-based fused metric in HeiSD to automatically determine the hybrid boundary. Experimental results demonstrate that HeiSD attains a speedup of up to 2.45x in simulation benchmarks and 2.06x~2.41x in real-world scenarios, while sustaining a high task success rate.

Zihao Zheng, Zhihao Mao, Maoliang Li et al.

Vision-Language-Action (VLA) models build a token-domain robot control paradigm, yet suffer from low speed. Speculative Decoding (SD) is an optimization strategy that can boost inference speed. Two key issues emerge when integrating VLA and SD: first, SD relies on re-inference to address token errors, which is computationally expensive; second, to mitigate token errors, the acceptance threshold in SD requires careful adjustment. Existing works fail to address the above two issues effectively. Meanwhile, as the bridge between AI and the physical world, existing embodied intelligence has overlooked the application of robotic kinematics. To address these issues, we innovatively combine token-domain VLA models with kinematic-domain prediction for SD, proposing a kinematic-rectified SD framework named KERV. We employ a kinematics-based Kalman Filter to predict actions and compensate for SD errors, avoiding costly re-inference. Moreover, we design a kinematics-based adjustment strategy to dynamically rectify the acceptance threshold, addressing the difficulty of threshold determination. Experimental results across diverse tasks and environments demonstrate that KERV achieves 27%~37% acceleration with nearly no Success Rate loss.

Ke Li, Maoliang Li, Jialiang Chen et al.

Video mashup creation represents a complex video editing paradigm that recomposes existing footage to craft engaging audio-visual experiences, demanding intricate orchestration across semantic, visual, and auditory dimensions and multiple levels. However, existing automated editing frameworks often overlook the cross-level multimodal orchestration to achieve professional-grade fluidity, resulting in disjointed sequences with abrupt visual transitions and musical misalignment. To address this, we formulate video mashup creation as a Multimodal Coherency Satisfaction Problem (MMCSP) and propose the DIRECT framework. Simulating a professional production pipeline, our hierarchical multi-agent framework decomposes the challenge into three cascade levels: the Screenwriter for source-aware global structural anchoring, the Director for instantiating adaptive editing intent and guidance, and the Editor for intent-guided shot sequence editing with fine-grained optimization. We further introduce Mashup-Bench, a comprehensive benchmark with tailored metrics for visual continuity and auditory alignment. Extensive experiments demonstrate that DIRECT significantly outperforms state-of-the-art baselines in both objective metrics and human subjective evaluation. Project page and code: https://github.com/AK-DREAM/DIRECT

Jiayu Chen, Junbei Tang, Wenbiao Zhao et al.

Autoregressive video generation enables streaming and open-ended long video synthesis, but still suffers from long-term degradation caused by accumulated errors. Existing KVCache strategies usually apply unified historical-frame retention, implicitly assuming homogeneous historical dependencies across attention heads. We revisit historical-frame attention and reveal three distinct head types: Anchor Heads require broad long-range context, Wave Heads exhibit periodic temporal dependencies, and Veil Heads focus on initial and adjacent frames. Based on this finding, we propose Pyramid Forcing, a head-aware pyramidal KVCache framework that identifies head types offline, assigns behavior-specific cache policies, and supports heterogeneous cache lengths via efficient ragged-cache attention. Experiments on Self Forcing and Causal Forcing show that Pyramid Forcing consistently improves long-horizon generation quality on VBench-Long, increasing the 60-second Self Forcing score from 77.87 to 81.21 while enhancing motion dynamics, visual fidelity, and semantic consistency. Project: https://if-lab-pku.github.io/Pyramid-Forcing/.

Zihao Zheng, Hangyu Cao, Jiayu Chen et al.

Vision-Language-Action (VLA) models are mainstream in embodied intelligence but face high inference costs. Edge-Cloud Collaborative (ECC) deployment offers an effective fix by easing edge-device computing pressure to meet real-time needs. However, existing ECC frameworks are suboptimal for VLA models due to two challenges: (1) Diverse model structures hinder optimal ECC segmentation point identification; (2) Even if the optimal split point is determined, changes in network bandwidth can cause performance drift. To address these issues, we propose a novel ECC deployment framework for various VLA models, termed RoboECC. Specifically, we propose a model-hardware co-aware segmentation strategy to help find the optimal segmentation point for various VLA models. Moreover, we propose a network-aware deployment adjustment approach to adapt to the network fluctuations for maintaining optimal performance. Experiments demonstrate that RoboECC achieves a speedup of up to 3.28x with only 2.55x~2.62x overhead.

Jiayu Chen, Ruoyu Lin, Zihao Zheng et al.

Visual Autoregressive(VAR) models enhance generation quality but face a critical efficiency bottleneck in later stages. In this paper, we present a novel optimization framework for VAR models that fundamentally differs from prior approaches such as FastVAR and SkipVAR. Instead of relying on heuristic skipping strategies, our method leverages attention entropy to characterize the semantic projections across different dimensions of the model architecture. This enables precise identification of parameter dynamics under varying token granularity levels, semantic scopes, and generation scales. Building on this analysis, we further uncover sparsity patterns along three critical dimensions-token, layer, and scale-and propose a set of fine-grained optimization strategies tailored to these patterns. Extensive evaluation demonstrates that our approach achieves aggressive acceleration of the generation process while significantly preserving semantic fidelity and fine details, outperforming traditional methods in both efficiency and quality. Experiments on Infinity-2B and Infinity-8B models demonstrate that ToProVAR achieves up to 3.4x acceleration with minimal quality loss, effectively mitigating the issues found in prior work. Our code will be made publicly available.

Zihao Zheng, Hangyu Cao, Sicheng Tian et al.

Vision-Language-Action (VLA) models are dominant in embodied intelligence but are constrained by inference overheads. While model quantization alleviates these bottlenecks for edge deployment, static quantization approaches remain suboptimal for VLAs due to two critical challenges: (1) Temporal-dynamic sensitivity, where fixed precision wastes resources by ignoring stage-varying error tolerances; and (2) Real-time allocation, where identifying real-time sensitivity to guide bit allocation remains unsolved. To address these challenges, we propose DyQ-VLA, a dynamic quantization framework for VLAs. Specifically, a sensitivity-aware switching strategy leverages real-time kinematic proxies to trigger the bit-width switch, while a kinematic-guided module dynamically allocates the optimal bit-width. Experiments show that DyQ-VLA requires only 30.9% of the original memory footprint while maintaining 99.5% of its original performance, achieving 1.49x simulation and up to 1.43x real-world speedups.

Zihao Zheng, Sicheng Tian, Hangyu Cao et al.

Vision Language Action (VLA) models are mainstream in embodied intelligence but face high inference costs. Edge-Cloud Collaborative (ECC) inference offers an effective fix by easing edge-device computing pressure to meet real-time needs. However, existing ECC frameworks are suboptimal for VLA models due to two challenges: (1) Mainstream environment-oriented edge-cloud partitioning methods are susceptible to interference from visual noise; (2) Existing edge-cloud partitioning methods overlook the step-wise redundancy unique to embodied tasks, thereby disrupting the physical continuity of motion. To address these issues, we propose a novel ECC inference framework, termed RAPID. Specifically, we developed an implementation tailored to the proposed framework. Experiments demonstrate this achieves a speedup of up to 1.73x with only 5%~7% overhead.

Xinming Wei, Jiahao Zhang, Haoran Li et al.

The proliferation of agentic Large Language Models (LLMs) on personal devices introduces a new class of workloads characterized by a dichotomy of objectives. Reactive tasks, initiated by users, demand immediate, low-latency responses, while proactive tasks operate invisibly and prioritize throughput. Existing on-device LLM engines, designed for isolated inferences, fail to efficiently manage these concurrent and conflicting requests on consumer-grade heterogeneous SoCs with CPU, integrated GPU, and NPU. This paper introduces Agent.xpu, an efficient serving system for agentic LLM workloads on memory-unified heterogeneous SoCs. With dedicated offline profiling, Agent.xpu first constructs a heterogeneous execution graph, which fuses and chunks model kernels for affinity-guided, elastic accelerator mapping with predictive kernel annotation. At runtime, its online scheduler enables fine-grained, kernel-level preemption to guarantee the responsiveness of reactive tasks. To maximize SoC utilization, it adopts slack-aware kernel backfill to opportunistically append proactive tasks, and mitigates NPU-iGPU contention via bandwidth-aware dispatch. Evaluation on an Intel Core Ultra SoC shows that Agent.xpu achieves 4.6$\times$ lower latency for reactive tasks and sustains 1.6$\times$-6.8$\times$ higher throughput for proactive tasks compared to state-of-the-art inference engines.

Tao Xie, David Harel, Dezhi Ran et al.

Sustainable AI is a subfield of AI for concerning developing and using AI systems in ways of aiming to reduce environmental impact and achieve sustainability. Sustainable AI is increasingly important given that training of and inference with AI models such as large langrage models are consuming a large amount of computing power. In this article, we discuss current issues, opportunities and example solutions for addressing these issues, and future challenges to tackle, from the data and system perspectives, related to data acquisition, data processing, and AI model training and inference.

Zihao Zheng, Zhihao Mao, Xingyue Zhou et al.

Vision-and-Language Navigation (VLN) increasingly relies on large vision-language models, but their inference cost conflicts with real-time deployment. Token caching is a promising training-free strategy that avoids redundant computation by reusing stable visual tokens across frames. However, existing methods assume a static camera and fixed semantic focus, assumptions that VLN fundamentally violates. We identify two failure modes: (1) visual dynamics, where viewpoint shift displaces token positions across frames, causing position-wise matching to pair misaligned content; (2) semantic dynamics, where token relevance shifts across task stages as navigation progresses, making cached states stale. We propose VLN-Cache, a visual-dynamic-aware and semantic-dynamic-aware caching framework that introduces view-aligned remapping to recover geometric correspondences and a task-relevance saliency filter to veto reuse at semantic transitions. A layer-adaptive entropy policy further balances the per-layer reuse budget. Experiments on the R2R-CE simulation benchmark show up to 1.52x speedup while maintaining competitive navigation success rates.

Zihao Zheng, Xiuping Cui, Size Zheng et al.

With the advances in artificial intelligence, Mix-of-Experts (MoE) has become the main form of Large Language Models (LLMs), and its demand for model compression is increasing. Quantization is an effective method that not only compresses the models but also significantly accelerates their performance. Existing quantization methods have gradually shifted the focus from parameter scaling to the analysis of data distributions. However, their analysis is designed for dense LLMs, which are suboptimal for MoE quantization, due to MoEs' complex data-model distribution. To address this problem, we decouple the complexity of MoEs' data-model distribution into a multi-stage analysis and reveal MoEs' inherent dynamics. The analysis results show that the expert performance of MoE varies dynamically both within and across data distributions. Based on these, we design two quantization strategies with data-model distribution awareness and integrate them into an end-to-end framework for MoE quantization, which is named MoQa. MoQa uses an expert-level mix-precision base quantization with distribution awareness. Moreover, MoQa uses a channel-level quantization adjustment to dynamically adjust expert performance to adapt to novel distributions. Experiments show that MoQa's base quantization achieves a 0.49~8.51 PPL decrease on known distributions. With the adjustments, MoQa achieves a 2.74~6.44 PPL decrease and 1.85%~3.77% average accuracy improvements on novel distributions. We believe MoQa will play a role in future MoE construction, optimization, and compression.

Chenhui Xu, Fuxun Yu, Maoliang Li et al.

The past neural network design has largely focused on feature representation space dimension and its capacity scaling (e.g., width, depth), but overlooked the feature interaction space scaling. Recent advancements have shown shifted focus towards element-wise multiplication to facilitate higher-dimensional feature interaction space for better information transformation. Despite this progress, multiplications predominantly capture low-order interactions, thus remaining confined to a finite-dimensional interaction space. To transcend this limitation, classic kernel methods emerge as a promising solution to engage features in an infinite-dimensional space. We introduce InfiNet, a model architecture that enables feature interaction within an infinite-dimensional space created by RBF kernel. Our experiments reveal that InfiNet achieves new state-of-the-art, owing to its capability to leverage infinite-dimensional interactions, significantly enhancing model performance.

Maoliang Li, Ke Li, Yaoyang Liu et al.

To effectively leverage user-specific data, retrieval augmented generation (RAG) is employed in multimodal large language model (MLLM) applications. However, conventional retrieval approaches often suffer from limited retrieval accuracy. Recent advances in multi-vector retrieval (MVR) improve accuracy by decomposing queries and matching against segmented images. They still suffer from sub-optimal accuracy and efficiency, overlooking alignment between the query and varying image objects and redundant fine-grained image segments. In this work, we present an efficient scheduling framework for image retrieval - HiMIR. First, we introduce a novel hierarchical paradigm, employing multiple intermediate granularities for varying image objects to enhance alignment. Second, we minimize redundancy in retrieval by leveraging cross-hierarchy similarity consistency and hierarchy sparsity to minimize unnecessary matching computation. Furthermore, we configure parameters for each dataset automatically for practicality across diverse scenarios. Our empirical study shows that, HiMIR not only achieves substantial accuracy improvements but also reduces computation by up to 3.5 times over the existing MVR system.

Feng Jiang, Zihao Zheng, Xiuping Cui et al.

With the development of Embodied Artificial intelligence, the end-to-end control policy such as Vision-Language-Action (VLA) model has become the mainstream. Existing VLA models faces expensive computing/storage cost, which need to be optimized. Quantization is considered as the most effective method which can not only reduce the memory cost but also achieve computation acceleration. However, we find the token alignment of VLA models hinders the application of existing quantization methods. To address this, we proposed an optimized framework called EaqVLA, which apply encoding-aligned quantization to VLA models. Specifically, we propose an complete analysis method to find the misalignment in various granularity. Based on the analysis results, we propose a mixed precision quantization with the awareness of encoding alignment. Experiments shows that the porposed EaqVLA achieves better quantization performance (with the minimal quantization loss for end-to-end action control and xxx times acceleration) than existing quantization methods.

Zihao Zheng, Ziyao Wang, Xiuping Cui et al.

Federated Learning (FL) is a decentralized model training approach that preserves data privacy but struggles with low efficiency. Quantization, a powerful training optimization technique, has been widely explored for integration into FL. However, many studies fail to consider the distinct performance attribution between particular quantization strategies, such as post-training quantization (PTQ) or quantization-aware training (QAT). As a result, existing FL quantization methods rely solely on either PTQ or QAT, optimizing for speed or accuracy while compromising the other. To efficiently accelerate FL and maintain distributed convergence accuracy across various FL settings, this paper proposes a hybrid quantitation approach combining PTQ and QAT for FL systems. We conduct case studies to validate the effectiveness of using hybrid quantization in FL. To solve the difficulty of modeling speed and accuracy caused by device and data heterogeneity, we propose a hardware-related analysis and data-distribution-related analysis to help identify the trade-off boundaries for strategy selection. Based on these, we proposed a novel framework named FedHQ to automatically adopt optimal hybrid strategy allocation for FL systems. Specifically, FedHQ develops a coarse-grained global initialization and fine-grained ML-based adjustment to ensure efficiency and robustness. Experiments show that FedHQ achieves up to 2.47x times training acceleration and up to 11.15% accuracy improvement and negligible extra overhead.