Khai Le-Duc, Tuyen Tran, Bach Phan Tat, Nguyen Kim Hai Bui, Quan Dang Anh, Hung-Phong Tran, Thanh Thuy Nguyen, Ly Nguyen, Tuan Minh Phan, Thi Thu Phuong Tran, Chris Ngo, Khanh Xuan Nguyen, Thanh Nguyen-Tang. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Accurate lung nodule segmentation in computed tomography (CT) scans is a critical step in early lung cancer detection. While deep learning approaches have shown considerable success, challenges persist due to the small size, varied shape, and blurred boundaries of pulmonary nodules. To address these difficulties, we propose a learnable Partial Differential Equation (PDE) module that performs anisotropic diffusion as a trainable feature refinement layer within neural architectures. This module is fully differentiable and jointly optimized with the network, enabling adaptive spatial smoothing based on local gradient structures. By enhancing intra-region consistency and preserving important anatomical edges, the PDE module improves the quality of feature representation prior to prediction. Integrated into an enhanced PCAM-ResUNet — a modified variant of ResUNet++ with spatial and channel attention — the proposed approach achieves a Dice Similarity Coefficient of 90.22% on the LUNA16 dataset and 89.59% on its noise-augmented variant (LUNA-Noise), along with a Miss Rate of 9.69%, highlighting its strong detection sensitivity and robustness.

An allocation of resources to a virtual machine specifies the maximum amount of each individual element of each resource type that will be utilized, as well as the aggregate amount of each resource of each type. An allocation is thus represented by two vectors, a maximum elementary allocation vector and an aggregate allocation vector. There are more general types of resource allocation problems than those we consider here. In this paper, we present an approach for improving parallel deadlock detection algorithm, to schedule the policies of resource which supply for resource allocation in heterogeneous distributed platform. Parallel deadlock detection algorithm has a run time complexity of O(min(m,n)), where m is the number of resources and n is the number of processes. We propose the algorithm for allocating multiple resources to competing services running in virtual machines on a heterogeneous distributed platform. The experiments also compare the performance of the proposed approach with other related work.