• Corpus ID: 54446463

EENMF: An End-to-End Neural Matching Framework for E-Commerce Sponsored Search

  title={EENMF: An End-to-End Neural Matching Framework for E-Commerce Sponsored Search},
  author={Wenjin Wu and Guojun Liu and Hui Ye and Chenshuang Zhang and Tianshu Wu and Daorui Xiao and Wei Lin and Xiaoyu Zhu},
E-commerce sponsored search contributes an important part of revenue for the e-commerce company. In consideration of effectiveness and efficiency, a large-scale sponsored search system commonly adopts a multi-stage architecture. We name these stages as ad retrieval, ad pre-ranking and ad ranking. Ad retrieval and ad pre-ranking are collectively referred to as ad matching in this paper. We propose an end-to-end neural matching framework (EENMF) to model two tasks---vector-based ad retrieval and… 

Figures and Tables from this paper

Towards a Better Tradeoff between Effectiveness and Efficiency in Pre-Ranking: A Learnable Feature Selection based Approach

A novel pre-ranking approach is proposed which supports complicated models with interaction-focused architecture and achieves a better tradeoff between effectiveness and efficiency by utilizing the proposed learnable Feature Selection method based on feature Complexity and variational Dropout.

COLD: Towards the Next Generation of Pre-Ranking System

This paper designs a new pre- ranking system by joint optimization of both the pre-ranking model and the computing power it costs, and names it COLD (Computing power cost-aware Online and Lightweight Deep Pre-ranking system).

Optimizing AD Pruning of Sponsored Search with Reinforcement Learning

The approach considers the downstream as a black-box environment, and the agent sequentially selects items and finally feeds into the downstream, where revenue would be estimated and used as a reward to improve the selection policy.

Production Ranking Systems: A Review

This paper presents an overview of several production quality ranking systems, showing that due to conflicting goals of employing the most effective machine learning models and responding to users in real time, ranking systems have evolved into a system of systems, where each subsystem can be viewed as a component layer.

Making Rewards More Rewarding: Sequential Learnable Environments for Deep Reinforcement Learning-based Sponsored Ranking

By augmenting a learnable simulated environment based on intuitive design principles, this work can significantly improve RL performances and boost the explainability of the model and paves the way to a wide range of Reinforcement Learning applications to the sponsored ranking problem.



Beyond Keywords and Relevance: A Personalized Ad Retrieval Framework in E-Commerce Sponsored Search

A novel ad retrieval framework beyond keywords and relevance in e-commerce sponsored search is proposed, in which personalized information is introduced and a model on top of the hierarchical network representing signals, keys and ads is trained.

Cascade Ranking for Operational E-commerce Search

A novel cascade ranking model is designed and deployed in an operational e-commerce search application and an extensive set of experiments demonstrate the advantage of the proposed work to address multiple factors of effectiveness, efficiency and user experience in the real-world application.

Learning deep structured semantic models for web search using clickthrough data

A series of new latent semantic models with a deep structure that project queries and documents into a common low-dimensional space where the relevance of a document given a query is readily computed as the distance between them are developed.

Multi-task learning for learning to rank in web search

A boosting framework for learning to rank in the multi- task learning context to learn non-parametric common structures adaptively from multiple tasks in a stage-wise way and demonstrates that multi-task learning methods bring significant relevance improvements over existing baseline methods.

Embedding-based News Recommendation for Millions of Users

An embedding-based method to use distributed representations in a three step end-to-end manner that performed well in an experimental offline evaluation using past access data on Yahoo! JAPAN's homepage and compared its online performance with a method that was conventionally incorporated into the system.

Context- and Content-aware Embeddings for Query Rewriting in Sponsored Search

This work proposes rewriting method based on a novel query embedding algorithm, which jointly models query content as well as its context within a search session, and shows the proposed approach significantly outperformed existing state-of-the-art, strongly indicating its benefits and the monetization potential.

Deep Interest Network for Click-Through Rate Prediction

A novel model: Deep Interest Network (DIN) is proposed which tackles this challenge by designing a local activation unit to adaptively learn the representation of user interests from historical behaviors with respect to a certain ad.

Scalable Semantic Matching of Queries to Ads in Sponsored Search Advertising

This study presents a novel advance match approach based on the idea of semantic embeddings of queries and ads that significantly outperforms baselines in terms of relevance, coverage and incremental revenue.

Personalized click prediction in sponsored search

This paper develops user-specific and demographic-based features that reflect the click behavior of individuals and groups in sponsored search and demonstrates that the personalized models significantly improve the accuracy of click prediction.

Learning to Match using Local and Distributed Representations of Text for Web Search

This work proposes a novel document ranking model composed of two separate deep neural networks, one that matches the query and the document using a local representation, and another that Matching with distributed representations complements matching with traditional local representations.