Open-domain Keyphrase extraction (KPE) onthe Web is a fundamental yet complex NLP task with a wide range of practical applications within the field of Information Retrieval.In contrast to other document types, web page designs are intended for easy navigation and information finding.Effective designs encode within the layout and formatting signals that point to where the important information can be found.In this work, we propose a modeling approach that leverages these multi-modal signals to aid in the KPE task.In particular, we leverage both lexical and visual features (e.g., size, font, position) at the micro-level to enable effective strategy induction, and metalevel features that describe pages at a macrolevel to aid in strategy selection.Our evaluation demonstrates that a combination of effective strategy induction and strategy selection within this approach for the KPE task outperforms state-of-the-art models.A qualitative post-hoc analysis illustrates how these features function within the model.

Finding diagrams that contain a specific part or a similar part is important in many engineering tasks. In this search task, the query part is expected to match only a small region in a complex image. This paper investigates several local matching networks that explicitly model local region-to-region similarities. Deep convolutional neural networks extract local features and model local matching patterns. Spatial convolution is employed to cross-match local regions at different scale levels, addressing cases where the target part appears at a different scale, position, and/or angle. A gating network automatically learns region importance, removing noise from sparse areas and visual metadata in engineering diagrams.

Data is of vital importance in the development of machine learning technologies. Recently, within the information retrieval field, a number of neural ranking frameworks have been proposed to address the ad-hoc search. These models usually need a large amount of query-document relevance judgments for training. However, obtaining this kind of relevance judgments needs a lot of money and manual effort. To shed light on this problem, researchers seek to use implicit feedback from users of search engines to improve the ranking performance. In this paper, we present a new dataset, Sogou-QCL, which contains 537,366 queries and five kinds of weak relevance labels for over 12 million query-document pairs. We apply Sogou-QCL dataset to train recent neural ranking models and show its potential to serve as weak supervision for ranking. We believe that Sogou-QCL will have a broad impact on corresponding areas.