Author name cluster

Yixiao Wang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

8 papers

2 author rows

AAAI Conference 2026 Conference Paper

FlashSVD: Memory-Efficient Inference with Streaming for Low-Rank Models

Zishan Shao
Yixiao Wang
Qinsi Wang
Ting Jiang
Zhixu Du
Hancheng Ye
Danyang Zhuo
Yiran Chen

Singular Value Decomposition (SVD) has recently gained traction as an effective compression technique for large language models (LLMs), with many studies reporting 20-80% parameter reduction at minimal accuracy cost. However, despite reducing weight memory, existing SVD-based approaches still rely on standard dense CUDA kernels during inference, which incur substantial-and ultimately unnecessary-activation memory overhead. Our analysis reveals that this kernel-induced cost, which grows with sequence length and hidden size, in worst case prevents any real reduction in peak inference memory, limiting the practical impact of SVD compression for on-device deployment. To address this bottleneck, we propose FlashSVD, an end-to-end, rank-aware streaming inference framework for SVD-compressed LLMs. FlashSVD integrates seamlessly with any SVD-based model and directly fuses low-rank projection kernels into self-attention and feed-forward pipelines. This design avoids materializing large activation buffers by streaming small tiles of truncated factors through on-chip SRAM, performing on-the-fly multiplication and reduction, and immediately evicting results–thus preserving high GPU occupancy without introducing latency. On standard benchmarks (e.g., BERT-Base), FlashSVD reduces peak activation memory by up to 70.2% and transient memory by 75%, with zero accuracy loss against low-rank baselines, enabling truly memory-efficient deployment of low-rank LLMs.

PDF Details DOI

ICRA Conference 2025 Conference Paper

Adaptive Energy Regularization for Autonomous Gait Transition and Energy-Efficient Quadruped Locomotion

Boyuan Liang
Lingfeng Sun
Xinghao Zhu
Bike Zhang
Ziyin Xiong
Yixiao Wang
Chenran Li
Koushil Sreenath

In reinforcement learning for legged robot locomotion, crafting effective reward strategies is crucial. Predefined gait patterns and complex reward systems are widely used to stabilize policy training. Drawing from the natural locomotion behaviors of humans and animals, which adapt their gaits to minimize energy consumption, we investigate the impact of incorporating an energy-efficient reward term that prioritizes distance-averaged energy consumption into the reinforcement learning framework. Our findings demonstrate that this simple addition enables quadruped robots to autonomously select appropriate gaits-such as four-beat walking at lower speeds and trotting at higher speeds-without the need for explicit gait regularizations. Furthermore, we provide a guideline for tuning the weight of this energy-efficient reward, facilitating its application in real-world scenarios. The effectiveness of our approach is validated through simulations and on a real Unitree Gol robot. This research highlights the potential of energy-centric reward functions to simplify and enhance the learning of adaptive and efficient locomotion in quadruped robots. Videos and more details are at https://sites.google.com/berkeley.edu/efficient-locomotion

Details

NeurIPS Conference 2025 Conference Paper

Enhanced Cyclic Coordinate Descent Methods for Elastic Net Penalized Linear Models

Yixiao Wang
Zishan Shao
Ting Jiang
Aditya Devarakonda

We present a novel enhanced cyclic coordinate descent (ECCD) framework for solving generalized linear models with elastic net constraints that reduces training time in comparison to existing state-of-the-art methods. We redesign the CD method by performing a Taylor expansion around the current iterate to avoid nonlinear operations arising in the gradient computation. By introducing this approximation we are able to unroll the vector recurrences occurring in the CD method and reformulate the resulting computations into more efficient batched computations. We show empirically that the recurrence can be unrolled by a tunable integer parameter, $s$, such that $s > 1$ yields performance improvements without affecting convergence, whereas $s = 1$ yields the original CD method. A key advantage of ECCD is that it avoids the convergence delay and numerical instability exhibited by block coordinate descent. Finally, we implement our proposed method in C++ using Eigen to accelerate linear algebra computations. Comparison of our method against existing state-of-the-art solvers show consistent performance improvements of $3\times$ in average for regularization path variant on diverse benchmark datasets. Our implementation is available at [https: //github. com/Yixiao-Wang-Stats/ECCD](https: //github. com/Yixiao-Wang-Stats/ECCD).

PDF Details

ICRA Conference 2025 Conference Paper

Physics-Aware Robotic Palletization With Online Masking Inference

Tianqi Zhang
Zheng Wu 0002
Yuxin Chen
Yixiao Wang
Boyuan Liang
Scott Moura
Masayoshi Tomizuka
Mingyu Ding

The efficient planning of stacking boxes, especially in the online setting where the sequence of item arrivals is unpredictable, remains a critical challenge in modern warehouse and logistics management. Existing solutions often address box size variations, but overlook their intrinsic and physical properties, such as density and rigidity, which are crucial for real-world applications. We use reinforcement learning (RL) to solve this problem by employing action space masking to direct the RL policy toward valid actions. Unlike previous methods that rely on heuristic stability assessments which are difficult to assess in physical scenarios, our framework utilizes online learning to dynamically train the action space mask, eliminating the need for manual heuristic design. Extensive experiments demonstrate that our proposed method outperforms existing state-of-the-arts. Furthermore, we deploy our learned task planner in a real-world robotic palletizer, validating its practical applicability in operational settings. The code is available at https://github.com/tianqi-zh/palletization.

Details

ICML Conference 2025 Conference Paper

SADA: Stability-guided Adaptive Diffusion Acceleration

Ting Jiang
Yixiao Wang
Hancheng Ye
Zishan Shao
Jingwei Sun 0002
Jingyang Zhang
Zekai Chen
Jianyi Zhang

Diffusion models have achieved remarkable success in generative tasks but suffer from high computational costs due to their iterative sampling process and quadratic-attention costs. Existing training-free acceleration strategies that reduce per-step computation cost, while effectively reducing sampling time, demonstrate low faithfulness compared to the original baseline. We hypothesize that this fidelity gap arises because (a) different prompts correspond to varying denoising trajectory, and (b) such methods do not consider the underlying ODE formulation and its numerical solution. In this paper, we propose Stability-guided Adaptive Diffusion Acceleration (SADA), a novel paradigm that unifies step-wise and token-wise sparsity decisions via a single stability criterion to accelerate sampling of ODE-based generative models (Diffusion and Flow-matching). For (a), SADA adaptively allocates sparsity based on the sampling trajectory. For (b), SADA introduces principled approximation schemes that leverage the precise gradient information from the numerical ODE solver. Comprehensive evaluations on SD-2, SDXL, and Flux using both EDM and DPM++ solvers reveal consistent $\ge 1. 8\times$ speedups with minimal fidelity degradation (LPIPS $\leq 0. 10$ and FID $\leq 4. 5$) compared to unmodified baselines, significantly outperforming prior methods. Moreover, SADA adapts seamlessly to other pipelines and modalities: It accelerates ControlNet without any modifications and speeds up MusicLDM by $1. 8\times$ with $\sim 0. 01$ spectrogram LPIPS. Our code is available at: https: //github. com/Ting-Justin-Jiang/sada-icml.

Details

JBHI Journal 2024 Journal Article

DRL-Based URLLC-Constraint and Energy-Efficient Task Offloading for Internet of Health Things

Yixiao Wang
Huaming Wu
Rutvij H. Jhaveri
Youcef Djenouri

Internet of Health Things (IoHT) is a promising e-Health paradigm that involves offloading numerous computational-intensive and delay-sensitive tasks from locally limited IoHT points to edge servers (ESs) with abundant computational resources in close proximity. However, existing computation offloading techniques struggle to meet the burgeoning health demands in ultra-reliable and low-latency communication (URLLC), one of the 5G application scenarios. This article proposes a Multi-Agent Soft-Actor-Critic-discrete based URLLC-constrained task offloading and resource allocation (MASACDUA) scheme to maximize throughput while minimizing power consumption on the remote side, considering the long-term URLLC constraints. The URLLC constraint conditions are formulated using extreme value theory, and Lyapunov optimization is employed to divide the problem into task offloading and computation resource allocation. MASAC-discrete and a queue backlog-aware algorithm are utilized to approach task offloading and computation resource allocation, respectively. Extensive simulation results demonstrate that MASACDUA outperforms traditional DRL algorithms under different IoHT points and data arrival rate intervals and achieves superior performance in delay, bound violation probability, and other characteristics related to URLLC.

Details DOI

IROS Conference 2024 Conference Paper

Joint Pedestrian Trajectory Prediction through Posterior Sampling

Haotian Lin 0003
Yixiao Wang
Mingxiao Huo
Chensheng Peng
Zhiyuan Liu
Masayoshi Tomizuka

Joint pedestrian trajectory prediction has long grappled with the inherent unpredictability of human behaviors. Recent works employing conditional diffusion models in trajectory prediction have exhibited notable success. Nevertheless, the heavy dependence on accurate historical data results in their vulnerability to noise disturbances and data incompleteness. To improve the robustness and reliability, we introduce the Guided Full Trajectory Diffuser (GFTD), a novel diffusion-based framework that translates prediction as the inverse problem of spatial-temporal inpainting and models the full joint trajectory distribution which includes both history and the future. By learning from the full trajectory and leveraging flexible posterior sampling methods, GFTD can produce accurate predictions while improving the robustness that can generalize to scenarios with noise perturbation or incomplete historical data. Moreover, the pre-trained model enables controllable generation without an additional training budget. Through rigorous experimental evaluation, GFTD exhibits superior performance in joint trajectory prediction with different data quality and in controllable generation tasks. See more results at https://sites.google.com/andrew.cmu.edu/posterior-sampling-prediction.

Details

ICRA Conference 2023 Conference Paper

Learning Agile Flight Maneuvers: Deep SE(3) Motion Planning and Control for Quadrotors

Yixiao Wang
Bingheng Wang
Shenning Zhang
Han Wei Sia
Lin Zhao 0009

Agile flights of autonomous quadrotors in clut-tered environments require constrained motion planning and control subject to translational and rotational dynamics. Tra-ditional model-based methods typically demand complicated design and heavy computation. In this paper, we develop a novel deep reinforcement learning-based method that tackles the challenging task of flying through a dynamic narrow gate. We design a model predictive controller with its adaptive tracking references parameterized by a deep neural network (DNN). These references include the traversal time and the quadrotor SE(3) traversal pose that encourage the robot to fly through the gate with maximum safety margins from various initial conditions. To cope with the difficulty of training in highly dynamic environments, we develop a reinforce-imitate learning framework to train the DNN efficiently that generalizes well to diverse settings. Furthermore, we propose a binary search algorithm that allows online adaption of the SE(3) references to dynamic gates in real-time. Finally, through extensive high-fidelity simulations, we show that our approach is adaptive to different gate trajectories, velocities, and orientations.

Details