Hongwei Sun

Zi Wang, Xiaotong Yu, Chengyan Wang et al.

Magnetic resonance imaging (MRI) is a widely used radiological modality renowned for its radiation-free, comprehensive insights into the human body, facilitating medical diagnoses. However, the drawback of prolonged scan times hinders its accessibility. The k-space undersampling offers a solution, yet the resultant artifacts necessitate meticulous removal during image reconstruction. Although Deep Learning (DL) has proven effective for fast MRI image reconstruction, its broader applicability across various imaging scenarios has been constrained. Challenges include the high cost and privacy restrictions associated with acquiring large-scale, diverse training data, coupled with the inherent difficulty of addressing mismatches between training and target data in existing DL methodologies. Here, we present a novel Physics-Informed Synthetic data learning framework for Fast MRI, called PISF. PISF marks a breakthrough by enabling generalized DL for multi-scenario MRI reconstruction through a single trained model. Our approach separates the reconstruction of a 2D image into many 1D basic problems, commencing with 1D data synthesis to facilitate generalization. We demonstrate that training DL models on synthetic data, coupled with enhanced learning techniques, yields in vivo MRI reconstructions comparable to or surpassing those of models trained on matched realistic datasets, reducing the reliance on real-world MRI data by up to 96%. Additionally, PISF exhibits remarkable generalizability across multiple vendors and imaging centers. Its adaptability to diverse patient populations has been validated through evaluations by ten experienced medical professionals. PISF presents a feasible and cost-effective way to significantly boost the widespread adoption of DL in various fast MRI applications.

Tianlun Hu, Tiancheng Hu, Shengsheng Litang et al.

Mixture-of-Experts (MoE) inference requires large-scale token exchange across devices, making dispatch and combine major bottlenecks in both prefill and decode. Beyond network transfer, routing-driven layout transformation, temporary relay, and output restoration can add substantial overhead. Existing MoE communication paths are often buffer-centric, using explicit inter-process relay and reordering buffers around collective transfer. This report presents a relay-buffer-free communication design for MoE inference acceleration on Ascend systems. The design reorganizes dispatch and combine around direct placement into destination expert windows and direct reading from remote expert windows. Built on globally pooled high-bandwidth memory and symmetric-memory allocation, it removes most intermediate relay and reordering buffers while retaining only lightweight control state, including counts, offsets, and synchronization metadata. We instantiate the design as two schedules for the main phases of MoE inference: a prefill schedule with richer planning state for throughput-oriented execution, and a compact decode schedule for latency-sensitive execution. Experiments on Ascend-based MoE workloads show reduced dispatch and combine latency in both settings. At the serving level, the implementation improves time to first token (TTFT), preserves competitive time per output token (TPOT), and enlarges the feasible scheduling space under practical latency constraints. These results indicate that, on platforms with globally addressable device memory, reducing intermediate buffering and output restoration around expert execution is an effective direction for accelerating MoE inference.

Haijun Zhang, Jinxiang Wang, Zhenhua Yu et al.

Large language models (LLMs) have made a profound impact across various fields due to their advanced capabilities. However, training these models at unprecedented scales requires extensive AI accelerator clusters and sophisticated parallelism strategies, which pose significant challenges in maintaining system reliability over prolonged training periods. A major concern is the substantial loss of training time caused by inevitable hardware and software failures. To address these challenges, we present FlashRecovery, a fast and low-cost failure recovery system comprising three core modules: (1) Active and real-time failure detection. This module performs continuous training state monitoring, enabling immediate identification of hardware and software failures within seconds, thus ensuring rapid incident response; (2) Scale-independent task restart. By employing different recovery strategies for normal and faulty nodes, combined with an optimized communication group reconstruction protocol, our approach ensures that the recovery time remains nearly constant, regardless of cluster scale; (3) Checkpoint-free recovery within one step. Our novel recovery mechanism enables single-step restoration, completely eliminating dependence on traditional checkpointing methods and their associated overhead. Collectively, these innovations enable FlashRecovery to achieve optimal Recovery Time Objective (RTO) and Recovery Point Objective (RPO), substantially improving the reliability and efficiency of long-duration LLM training. Experimental results demonstrate that FlashRecovery system can achieve training restoration on training cluster with 4, 800 devices in 150 seconds. We also verify that the time required for failure recovery is nearly consistent for different scales of training tasks.

Yan Xia, Hao Feng, Hongwei Sun et al.

Low-rank representation learning has emerged as a powerful tool for recovering missing values in power load data due to its ability to exploit the inherent low-dimensional structures of spatiotemporal measurements. Among various techniques, low-rank factorization models are favoured for their efficiency and interpretability. However, their performance is highly sensitive to the choice of regularization parameters, which are typically fixed or manually tuned, resulting in limited generalization capability or slow convergence in practical scenarios. In this paper, we propose a Regularization-optimized Low-Rank Factorization, which introduces a Proportional-Integral-Derivative controller to adaptively adjust the regularization coefficient. Furthermore, we provide a detailed algorithmic complexity analysis, showing that our method preserves the computational efficiency of stochastic gradient descent while improving adaptivity. Experimental results on real-world power load datasets validate the superiority of our method in both imputation accuracy and training efficiency compared to existing baselines.

Zi Wang, Chen Qian, Di Guo et al.

Deep learning has shown astonishing performance in accelerated magnetic resonance imaging (MRI). Most state-of-the-art deep learning reconstructions adopt the powerful convolutional neural network and perform 2D convolution since many magnetic resonance images or their corresponding k-space are in 2D. In this work, we present a new approach that explores the 1D convolution, making the deep network much easier to be trained and generalized. We further integrate the 1D convolution into the proposed deep network, named as One-dimensional Deep Low-rank and Sparse network (ODLS), which unrolls the iteration procedure of a low-rank and sparse reconstruction model. Extensive results on in vivo knee and brain datasets demonstrate that, the proposed ODLS is very suitable for the case of limited training subjects and provides improved reconstruction performance than state-of-the-art methods both visually and quantitatively. Additionally, ODLS also shows nice robustness to different undersampling scenarios and some mismatches between the training and test data. In summary, our work demonstrates that the 1D deep learning scheme is memory-efficient and robust in fast MRI.

Hongwei Sun, Qiang Wu

Distributed machine learning systems have been receiving increasing attentions for their efficiency to process large scale data. Many distributed frameworks have been proposed for different machine learning tasks. In this paper, we study the distributed kernel regression via the divide and conquer approach. This approach has been proved asymptotically minimax optimal if the kernel is perfectly selected so that the true regression function lies in the associated reproducing kernel Hilbert space. However, this is usually, if not always, impractical because kernels that can only be selected via prior knowledge or a tuning process are hardly perfect. Instead it is more common that the kernel is good enough but imperfect in the sense that the true regression can be well approximated by but does not lie exactly in the kernel space. We show distributed kernel regression can still achieves capacity independent optimal rate in this case. To this end, we first establish a general framework that allows to analyze distributed regression with response weighted base algorithms by bounding the error of such algorithms on a single data set, provided that the error bounds has factored the impact of the unexplained variance of the response variable. Then we perform a leave one out analysis of the kernel ridge regression and bias corrected kernel ridge regression, which in combination with the aforementioned framework allows us to derive sharp error bounds and capacity independent optimal rates for the associated distributed kernel regression algorithms. As a byproduct of the thorough analysis, we also prove the kernel ridge regression can achieve rates faster than $N^{-1}$ (where $N$ is the sample size) in the noise free setting which, to our best knowledge, are first observed and novel in regression learning.