Intelligent Systems Research (United States)

This article focuses on nonconvex federated composite optimization (FCO) problem, where the loss function is nonconvex and contains a nonsmooth regularizer. To resolve this problem, we propose FedRREF, a novel federated learning algorithm that integrates error feedback (EF) with the efficient random reshuffling (RR) technique, resulting in lower computation and communication costs. To the best of our knowledge, FedRREF is the first algorithm to consider RR and biased compression simultaneously in federated learning, especially in nonsmooth and nonconvex settings, and it is shown to have a $\mathcal {O}(1/\sqrt {T})$ convergence rate, where $T$ is the number of communication rounds. Finally, the numerical experiments illustrate the validity of the proposed algorithm.

Lung cancer is among the malignancies with the highest incidence and mortality rates worldwide, and it poses a serious threat to human health. Increasing the accuracy of pulmonary nodule detection in CT images is essential for the early diagnosis and treatment of lung cancer. However, the grayscale characteristics of lung CT images, together with the variability in the sizes and morphologies of nodules, make the existing detection models prone to false positives and false negatives, posing challenges for achieving accurate detection. To address these problems, an improved WTAM-YOLO model based on YOLOv11 is proposed in this study. The model features four main improvements: a wavelet convolution approach to expand the receptive field, a lightweight convolutional block attention module (CBAM) to enhance the key feature representations, a hierarchical residual attention mixer (HRAMi) module to improve the multiscale detection performance of the model, and an improved exponential moving average (iEMA) module to strengthen the detail capture ability of the model and reduce the number of false positives. Experiments are conducted with a pulmonary nodule dataset acquired from the Roboflow platform and the LUNA16 dataset. Compared with those of YOLOv11, the proposed model improves the mAP@50 values by 3.4% and 2.5%, the mAP@75 values by 9.5% and 7.0%, the precision values by 4.4% and 0.7%, and the recall values by 2.3% and 4.6% based on the Roboflow and LUNA16 datasets, respectively.

Rectified Flow (RF) has been widely used as an effective generative model. Although RF is primarily based on probability flow Ordinary Differential Equations (ODE), recent studies have shown that injecting noise through reverse-time Stochastic Differential Equations (SDE) for sampling can achieve superior generative performance. Inspired by Positive-incentive Noise (Pi-noise), we propose an innovative generative algorithm to train Pi-noise generators, namely Rectified Noise (RN), which improves the generative performance by injecting Pi-noise into the velocity field of pre-trained RF models. After introducing the Rectified Noise pipeline, pre-trained RF models can be efficiently transformed into Pi-noise generators. We validate Rectified Noise by conducting extensive experiments across various model architectures on different datasets. Notably, we find that: (1) RF models using Rectified Noise reduce FID from10.16 to 9.05 on ImageNet-1k. (2) The models of Pi-noise generators achieve improved performance with only 0.39% additional training parameters.

Scene flow estimation aims to predict 3D motion from consecutive point cloud frames, which is of great interest in autonomous driving field. Existing methods face challenges such as insufficient spatio-temporal modeling and inherent loss of fine-grained feature during voxelization. However, the success of Mamba, a representative state space model (SSM) that enables global modeling with linear complexity, provides a promising solution. In this paper, we propose MambaFlow, a novel scene flow estimation network with a mamba-based decoder. It enables deep interaction and coupling of spatio-temporal features using a well-designed backbone. Innovatively, we steer the global attention modeling of voxel-based features with point offset information using an efficient Mamba-based decoder, learning voxel-to-point patterns that are used to devoxelize shared voxel representations into point-wise features. To further enhance the model's generalization capabilities across diverse scenarios, we propose a novel scene-adaptive loss function that automatically adapts to different motion patterns. Extensive experiments on the Argoverse 2 benchmark demonstrate that MambaFlow achieves state-of-the-art performance with real-time inference speed among existing works, enabling accurate flow estimation in real-world urban scenarios. The code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/SCNU-RISLAB/MambaFlow</uri>.

ABSTRACT Small object detection, a critical technique for recognizing and localizing diminutive targets in visual data, plays a vital role in applications ranging from remote sensing and unmanned aerial vehicle (UAV) vision to autonomous driving. Current methodologies, however, face substantial challenges, including detection accuracy limitations, low‐resolution image processing difficulties, background noise interference, and target occlusion issues. To address these challenges, we propose FMF‐DETR, an innovative small object detection framework featuring frequency‐domain feature optimization through three key components: (1) High‐Low Frequency Fusion Model (HLFM), (2) Focused Diffusion Feature Pyramid Network (FDFPN), and (3) BiPathNet (BPNet). Specifically, the HLFM module enhances multiscale feature representation by emphasizing high‐frequency details while suppressing low‐frequency background noise. The FDFPN architecture improves detection performance in complex scenarios through multiscale feature fusion and saliency‐aware diffusion. BPNet introduces a dual‐path feature extraction mechanism that simultaneously enhances feature discriminability and reduces computational overhead. Through the synergistic integration of these components, the proposed framework enhances both detection accuracy and operational efficiency. Comprehensive evaluations on the VisDrone dataset demonstrate FMF‐DETR's superior performance, achieving a 2.2% accuracy improvement while reducing model parameters by 13.03M and computational complexity by 97.2G FLOPs compared to baseline methods. These results validate both the effectiveness and efficiency of our proposed framework.

Spiking Neural Networks (SNNs) process information through discrete, time-dependent spikes, closely 2 mimicking the dynamics of biological neurons. This temporal coding enables SNNs to capture rich spatio-3 temporal patterns and exploit event-driven sensing and computation, offering sophisticated information 4 processing capabilities and practical efficiency at the same time.The key strengths of SNNs reside in their biological plausibility and energy efficiency. SNNs can operate 6 with significantly reduced power consumption, compared to conventional neural networks and deep learning 7 models. Advances in learning algorithms, and theoretical and software frameworks have accelerated 8 the development of SNNs, bringing neuroscience-inspired models to practical machine intelligence in 9 engineering systems, while also reducing the performance gap from deep neural networks (DNNs).

Safety-critical systems often operate in partially observable environments, where assessing the safety of the underlying policy remains a fundamental challenge. This study focuses on evaluating policies by identifying regions of the belief-space that can lead the system’s policy to an undesirable state with a non-negligible probability. In this paper, we introduce Backward Monte Carlo Tree Search, the first Monte Carlo tree search framework that expands backward in time within the belief-space. The tree search begins from an undesired terminal belief and recursively explores its possible predecessors, constructing a tree of belief transitions that could lead to an unsafe outcome within a given horizon. Evaluations in gridworld and autonomous driving domains show that identifying beliefs from which failures may occur enables runtime risk forecasting and targeted policy retraining, marking a conceptual shift in how safety is validated under uncertainty.