Local SGD (FedAvg) reduces federated communication by running multiple local steps between server aggregations, but it suffers from “client drift” under non-IID data. Methods such as SCAFFOLD mitigate this drift by transmitting control variates, which roughly doubles the per-round payload. We propose Momentum-Buffered Local SGD (MB-LSGD), an algorithm that corrects client drift with no additional payload: each client maintains a local drift buffer updated from the standard model aggregate, and a half-step bias present in naive drift estimators is eliminated by an additive update rule. For smooth non-convex objectives, MB-LSGD attains an \(\mathcal {O}(1/\sqrt {M K R})\) convergence rate, matching centralized mini-batch SGD up to constants. The bound explicitly couples the buffer rate α with the participation rate p, providing concrete guidance for hyperparameter selection under partial client sampling. With 100 clients under Dir(0.1) partitioning, MB-LSGD reaches a peak accuracy of \(48.37\%\) on TinyImageNet, exceeding SCAFFOLD (\(47.49\%\)) at half the bandwidth, and a final accuracy of \(61.67\%\) on CIFAR-10, the highest among all stable baselines. MB-LSGD is also the only method whose final accuracy on CIFAR-10 matches its peak, indicating sustained convergence without late-stage degradation, and its halved per-round payload relative to SCAFFOLD translates to close to a 2 × reduction in end-to-end completion time in communication-bound deployments.

Motion planning is a crucial component in autonomous driving. State-of-the-art motion planners are trained on meticulously curated datasets, which are not only expensive to annotate but also insufficient in capturing rarely seen critical scenarios. Failing to account for such scenarios poses a significant risk to motion planners and may lead to incidents during testing. An intuitive solution is to manually compose such scenarios by programming and executing a simulator (e.g., CARLA). However, this approach incurs substantial human costs. Motivated by this, we propose an inexpensive method for generating diverse critical traffic scenarios to train more robust motion planners. First, we represent traffic scenarios as scripts, which are then used by the simulator to generate traffic scenarios. Next, we develop a method that accepts user-specified text descriptions, which a Large Language Model translates into scripts using in-context learning. The output scripts are sent to the simulator that produces the corresponding traffic scenarios. As our method can generate abundant safety-critical traffic scenarios, we use them as synthetic training data for motion planners. To demonstrate the value of generated scenarios, we train existing motion planners on our synthetic data, real-world datasets, and a combination of both. Our experiments show that motion planners trained with our data significantly outperform those trained solely on real-world data, showing the usefulness of our synthetic data and the effectiveness of our data generation method.

Depth completion has played a vital role in enhancing depth perception in various applications such as autonomous driving, robotics, and 3D reconstruction. It is a critical task that involves generating a dense depth map from an RGB image and an aligned sparse depth map. However, many existing studies simply assume the depth map to be readily available, overlooking the fact that depth maps are often missing or subject to perturbations in real-world scenarios. The absence of these data can mislead the network into overfitting incorrect patterns, ultimately diminishing the performance of depth completion models. To tackle these challenges, we propose G-Depth, a method that innovatively involves Graph Neural Networks (GNNs) into a two-branch backbone. By integrating the graph convolutional mechanism into the deep layers of the network, G-Depth achieves an expanded receptive field, enabling a more accurate understanding and capture of spatial relationships between objects. To enhance robustness, we also propose a novel Enhanced Data Distillation(EDD) method to address the issue of modality absence, mitigating its negative effects, achieving a 73.2% average improvement in RMSE without sparse depth input. Comprehensive experiments on KITTI dataset demonstrate the effectiveness and robustness of the proposed G-Depth.