Zhiyuan Yuan

With the eruption of big data, practical recommendation schemes are now very important in various fields, including e-commerce, social networks, and a number of web-based services. Nowadays, there exist many personalized movie recommendation schemes utilizing publicly available movie datasets (e.g., MovieLens and Netflix), and returning improved performance metrics (e.g., Root-Mean-Square Error (RMSE)). However, two fundamental issues faced by movie recommendation systems are still neglected: first, scalability, and second, practical usage feedback and verification based on real implementation. In particular, Collaborative Filtering (CF) is one of the major prevailing techniques for implementing recommendation systems. However, traditional CF schemes suffer from a time complexity problem, which makes them bad candidates for real-world recommendation systems. In this paper, we address these two issues. Firstly, a simple but high-efficient recommendation algorithm is proposed, which exploits users' profile attributes to partition them into several clusters. For each cluster, a virtual opinion leader is conceived to represent the whole cluster, such that the dimension of the original user-item matrix can be significantly reduced, then a Weighted Slope One-VU method is designed and applied to the virtual opinion leader-item matrix to obtain the recommendation results. Compared to traditional clustering-based CF recommendation schemes, our method can significantly reduce the time complexity, while achieving comparable recommendation performance. Furthermore, we have constructed a real personalized web-based movie recommendation system, MovieWatch, opened it to the public, collected user feedback on recommendations, and evaluated the feasibility and accuracy of our system based on this real-world data.

Spatially resolved transcriptomics provides the opportunity to investigate the gene expression profiles and the spatial context of cells in naive state, but at low transcript detection sensitivity or with limited gene throughput. Comprehensive annotating of cell types in spatially resolved transcriptomics to understand biological processes at the single cell level remains challenging. Here we propose Spatial-ID, a supervision-based cell typing method, that combines the existing knowledge of reference single-cell RNA-seq data and the spatial information of spatially resolved transcriptomics data. We present a series of benchmarking analyses on publicly available spatially resolved transcriptomics datasets, that demonstrate the superiority of Spatial-ID compared with state-of-the-art methods. Besides, we apply Spatial-ID on a self-collected mouse brain hemisphere dataset measured by Stereo-seq, that shows the scalability of Spatial-ID to three-dimensional large field tissues with subcellular spatial resolution.