Lei Miao

Lei Miao, Jianfeng Mao, Christos G. Cassandras

It has been shown that using appropriate channel coding schemes in wireless environments, transmission energy can be significantly reduced by controlling the packet transmission rate. This paper seeks optimal solutions for downlink transmission control problems, motivated by this observation and by the need to minimize energy consumption in real-time wireless networks. Our problem formulation deals with a more general setting than the paper authored by Gamal et. al., in which the MoveRight algorithm is proposed. The MoveRight algorithm is an iterative algorithm that converges to the optimal solution. We show that even under the more general setting, the optimal solution can be efficiently obtained through an approach decomposing the optimal sample path through certain "critical tasks" which in turn can be efficiently identified. We include simulation results showing that our algorithm is significantly faster than the MoveRight algorithm. We also discuss how to utilize our results and receding horizon control to perform on-line transmission scheduling where future task information is unknown.

Lei Miao, Lijian Xu

Today's fixed-cycle traffic signaling is highly suboptimal and aggravates traffic congestion and waste of energy in urban areas. In addition, it offers no quality-of-service guarantee and makes travel time prediction extremely hard. While existing traffic light control research primarily focuses on improving the average wait time of cars, we study in this paper how traffic light scheduling affects the worst-case wait time. In particular, we derive the time a car spends at an intersection in the best-case and the worst-case, respectively. Using the theoretical results, we propose a simple but effective controller and run simulation to verify its performance. The result shows that it works much better than fixed-cycle controllers in both light and heavy traffic scenarios.

Jiebin Fang, Zidi Yan, Churu Mao et al.

Nuclear magnetic resonance (NMR) spectroscopy provides an experimental readout of local chemical environments, but its use in molecular representation learning has been constrained by heterogeneous data and incomplete atom-level assignments. Here we construct complementary high-fidelity experimental and computational 13C NMR resources, which reveal a recurrent form of representational collapse: atoms that are equivalent in molecular topology can remain experimentally distinct in their real chemical environments, whereas explicit 3D descriptions are further limited by static conformations in dynamic regimes. To alleviate this bottleneck, we develop CLAIM (Contrastive Learning for Atom-to-molecule Inference of Molecular NMR), a framework that aligns efficient topological molecular inputs with atom-resolved NMR observables. Through hierarchical chemical priors and cross-level contrastive learning, CLAIM restores lost chemical resolution and markedly improves atom-level molecule-spectrum retrieval. CLAIM remains robust in flexible and tautomeric systems for 13C NMR prediction, improves stereoisomer discrimination without explicit 3D modelling, and transfers to broader molecular property tasks including ADMET prediction and fluorescence estimation. These results establish physically grounded spectral alignment as an effective strategy for alleviating chemical-environment collapse and for guiding experimentally grounded molecular representation learning.

Hao Yang, Youzhi Jin, Ziyin Li et al.

Numerous researches have proved that deep neural networks (DNNs) can fit everything in the end even given data with noisy labels, and result in poor generalization performance. However, recent studies suggest that DNNs tend to gradually memorize the data, moving from correct data to mislabeled data. Inspired by this finding, we propose a novel method named Dynamic Loss Thresholding (DLT). During the training process, DLT records the loss value of each sample and calculates dynamic loss thresholds. Specifically, DLT compares the loss value of each sample with the current loss threshold. Samples with smaller losses can be considered as clean samples with higher probability and vice versa. Then, DLT discards the potentially corrupted labels and further leverages supervised learning techniques. Experiments on CIFAR-10/100 and Clothing1M demonstrate substantial improvements over recent state-of-the-art methods. In addition, we investigate two real-world problems for the first time. Firstly, we propose a novel approach to estimate the noise rates of datasets based on the loss difference between the early and late training stages of DNNs. Secondly, we explore the effect of hard samples (which are difficult to be distinguished) on the process of learning from noisy labels.