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%.

Figure question answering (FQA) is a new multimodal task for visual question answering (VQA). Given a scientific-style Figure and a related question, the machine must answer the question through reasoning. The Relation Networks (RN), the earliest proposed approach for FQA, computes the representation of relations between objects within images to infer the answers. However, RN generates numerous relation features, which makes the reasoning process more complicated and restricts the performance. To better solve this problem, we introduce a novel framework, which consists of a deconvolutional network, an LSTM network and an affinitydriven relation network. Specifically, the deconvolutional network enhances the feature fusion by combining low-level and high-level features of images. The affinity-driven relation network efficiently represents the intra-relation within images and the inter-relation between images and questions, and makes the reasoning process more effective. The experimental results show that our approach outperforms most state-of the-art methods in FQA tasks.

On-demand cinemas are a new type of offline entertainment venues which have shown the rapid expansion in the recent years. Recommending movies of interest to the potential audiences in on-demand cinemas is keen but challenging because the recommendation scenario is totally different from all the existing recommendation applications including online video recommendation, offline item recommendation and group recommendation. In this paper, we propose a novel spatio-temporal approach called Pegasus. Because of the specific characteristics of on-demand cinema recommendation, Pegasus exploits the POI (Point of Interest) information around cinemas and the content descriptions of movies, apart from the historical movie consumption records of cinemas. Pegasus explores the temporal dynamics and spatial influences rooted in audience behaviors, and captures the similarities between cinemas, the changes of audience crowds, time-varying features and regional disparities of movie popularity. It offers an effective and explainable way to recommend movies to on-demand cinemas. The corresponding Pegasus system has been deployed in some pilot on-demand cinemas. Based on the real-world data from on-demand cinemas, extensive experiments as well as pilot tests are conducted. Both experimental results and post-deployment feedback show that Pegasus is effective.