Generating synthetic Computed Tomography (CT) images from Cone Beam Computed Tomography (CBCT) is desirable for improving the image quality of CBCT. Existing synthetic CT (sCT) generation methods using Convolutional Neural Networks (CNN) and Transformers often face difficulties in effectively capturing both global and local features and contrasts for high-quality sCT generation. In this work, we propose a Global-Local Feature and Contrast learning (GLFC) framework for sCT generation. First, a Mamba-Enhanced UNet (MEUNet) is introduced by integrating Mamba blocks into the skip connections of a high-resolution UNet for effective global and local feature learning. Second, we propose a Multiple Contrast Loss (MCL) that calculates synthetic loss at different intensity windows to improve quality for both soft tissues and bone regions. Experiments on the SynthRAD2023 dataset demonstrate that GLFC improved the SSIM of sCT from 77.91% to 91.50% compared with the original CBCT, and significantly outperformed several existing methods for sCT generation. The code is available at https://github.com/HiLab-git/GLFC

Distributed training techniques have been widely deployed in large-scale deep neural networks (DNNs) training on dense-GPU clusters. However, on public cloud clusters, due to the moderate inter-connection bandwidth between instances, traditional state-of-the-art distributed training systems cannot scale well in training large-scale models. In this paper, we propose a new computing and communication efficient top-k sparsification communication library for distributed training. To further improve the system scalability, we optimize I/O by proposing a simple yet efficient multi-level data caching mechanism and optimize the update operation by introducing a novel parallel tensor operator. Experimental results on a 16-node Tencent Cloud cluster (each node with 8 Nvidia Tesla V100 GPUs) show that our system achieves 25%-40% faster than existing state-of-the-art systems on CNNs and Transformer. We finally break the record on DAWNBench on training ResNet-50 to 93% top-5 accuracy on ImageNet.

This paper presents a project awarded by MGB HK to handle their inspection scheduling problem. MGB HK is the buying office of one of the largest retailers in the world, Metro Group. MGB HK handles all product procurement of Metro Group out of Europe. The inspection process is one of their critical processes along their entire procurement exercise. The objective of this project is to provide an effective scheduling engine so that in-house inspectors can handle as many inspections as possible using the least amount of time and costs. Meanwhile, we also help the company overcome their difficulties of data collection and maintenance as a result of the system we developed. Our engine will be deployed and integrated into the company’s IMS. The engine recorded an improvement in the scheduling of their inspections and initial prognosis indicates that delayed inspections have been greatly reduced by compared with previous schedule. The system can effectively schedule inspections by urgency, shipment value, and supplier’s historical performance. Other than the schedule, the AI engine can also generate solutions based on different strategies and criteria, which facilitate the decision-making process for the scheduling team and management at MGB HK.