Learning-to-Rank(LTR) is widely used in many Information Retrieval(IR) scenarios, including web search and Location Based Services(LBS) search. However, most existing LTR techniques mainly focus on homogeneous ranking. Taking QAC in Dianping search as an example, heterogeneous documents including suggested queries (SQ) and Point-of-Interests(POI) need to be ranked and presented to enhance user experience. New challenges are faced when conducting heterogeneous ranking, including inconsistent feature space and more serious position bias caused by distinct representation spaces. Therefore, we propose Deep Debiasing Experts Network (DDEN), a novel heterogeneous LTR approach based on Mixture-of-Experts architecture and gating network, to deal with the inconsistent feature space of documents in ranking system. Furthermore, DDEN mitigates the position bias by adopting adversarial-debiasing framework embedded with heterogeneous LTR techniques. We conduct reproducible experiments on industrial datasets from Dianping, one of the largest local life platforms, and deploy DDEN in online application. Results show that DDEN substantially improves ranking performance in offline evaluation and boost the overall click-through rate in online A/B test by 2.1%.

MOTIVATION: The increasing amount of time-series single-cell RNA sequencing (scRNA-seq) data raises the key issue of connecting cell states (i.e. cell clusters or cell types) to obtain the continuous temporal dynamics of transcription, which can highlight the unified biological mechanisms involved in cell state transitions. However, most existing trajectory methods are specifically designed for individual cells, so they can hardly meet the needs of accurately inferring the trajectory topology of the cell state, which usually contains cells assigned to different branches. RESULTS: Here, we present CStreet, a computed Cell State trajectory inference method for time-series scRNA-seq data. It uses time-series information to construct the k-nearest neighbor connections between cells within each time point and between adjacent time points. Then, CStreet estimates the connection probabilities of the cell states and visualizes the trajectory, which may include multiple starting points and paths, using a force-directed graph. By comparing the performance of CStreet with that of six commonly used cell state trajectory reconstruction methods on simulated data and real data, we demonstrate the high accuracy and high tolerance of CStreet. AVAILABILITY AND IMPLEMENTATION: CStreet is written in Python and freely available on the web at https://github.com/TongjiZhanglab/CStreet and https://doi.org/10.5281/zenodo.4483205. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.