Medication recommendation based on Electronic Health Records (EHRs) is an important research direction, which aims to make prescription recommendations according to EHRs of patients. Most existing methods either only make recommendation through EHRs of the current admission while ignoring the patient’s historical records, or fail to fully consider the correlations among the clinical events from every single admission. These methods have shown their limitations in dealing with the complex structural correlations and temporal dependencies of clinical events in EHRs, which results in the defect of recommendation quality as well as the lack of temporal prediction ability. For that, a novel graph-attention augmented temporal neural network is proposed to model both the structural and temporal information simultaneously. For each admission record, a co-occurrence graph is constructed to establish the correlations among clinical events, and then a graph-attention augmented mechanism is used to learn the structural correlations on the graphs to obtain better representation of this admission. Then a temporal updating module based on the gated recurrent units is further proposed to learn the temporal dependencies between multiple admissions of each patient. Furthermore, our proposed model is also constrained by the co-occurrence graph, which can capture the internal correlations of clinical events and provide better modeling capability when the training data is sparse. Experiments illustrate that our model is superior to the state-of-the-art methods on a real-world dataset MIMIC-III in all effectiveness measures.

Binghao Tang, Boda Lin, Haolong Yan, Si Li. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Undesirable reflections contained in photos taken in front of glass windows or doors often degrade visual quality of the image. Separating two layers apart benefits both human and machine perception. The polarization status of the light changes after refraction or reflection, providing more observations of the scene, which can benefit the reflection separation. Different from previous works that take three or more polarization images as input, we propose to exploit physical constraints from a pair of unpolarized and polarized images to separate reflection and transmission layers in this paper. Due to the simplified capturing setup, the system is more under-determined compared to the existing polarization-based works. In order to solve this problem, we propose to estimate the semi-reflector orientation first to make the physical image formation well-posed, and then learn to reliably separate two layers using additional networks based on both physical and numerical analysis. In addition, a motion estimation network is introduced to handle the misalignment of paired input. Quantitative and qualitative experimental results show our approach performs favorably over existing polarization and single image based solutions.

Gang Zhao, Guanting Dong, Yidong Shi, Haolong Yan, Weiran Xu, Si Li. Findings of the Association for Computational Linguistics: EMNLP 2022. 2022.