LibRec 精选：如何评估交互式推荐系统？

2019 年 5 月 5 日 LibRec智能推荐

LibRec 精选

LibRec智能推荐第 31 期（至2019.5.4），更新 11 篇精选内容。

为什么我们需要陪伴？因为陪伴很温暖，它意味着这个世界上，有人愿意把最美好的东西给你，那就是时间。

【PPT】A Short Introduction to NLP，链接: https://docs.google.com/presentation/d/1MGDtkMnp8fG5BzllugNf2PAlg8BdYhGBm4X8mw-emLE/edit#slide=id.g5651326165_0_42

【视频】两个MIT博士学生做了两个视频，简单地解释什么是机器学习？，链接：https://www.youtube.com/watch?v=ZmBUnJ7lGvQ&feature=youtu.be，机器学习的两大分类：监督学习和无监督学习，链接：https://www.youtube.com/watch?v=wy-m6sd1BOA&feature=youtu.be

【论文及源码】场景文本识别（Scene Text Recognition, STR）任务的多个方法的实验对比。

+ 论文：https://arxiv.org/pdf/1904.01906.pdf

+ 源码：https://github.com/clovaai/deep-text-recognition-benchmark

近期热点论文

导读：交互式推荐系统使用户可以通过提供实时反馈等方式参与及影响推荐生成的过程。那么，该如何有效地评估交互式推荐系统的性能呢？本期第一篇论文为此提出了一个评估指标和框架。

1. An Evaluation Framework for Interactive Recommender System

Oznur Alkan, Elizabeth M. Daly, Adi Botea

https://arxiv.org/abs/1904.07765v1

Interactive recommender systems present an opportunity to engage the user in the process by allowing them to interact with the recommendations, provide feedback and impact the results in real-time. As a result, we present an evaluation framework which aims to simulate the users interacting with the recommender. We formulate metrics to evaluate the quality of the interactive recommenders which are outputted by the framework once simulation is completed.

2. Multi-Interest Network with Dynamic Routing for Recommendation at Tmall

Chao Li, Zhiyuan Liu, Mengmeng Wu, Yuchi Xu, Pipei Huang, Huan Zhao, Guoliang Kang, Qiwei Chen, Wei Li, Dik Lun Lee

https://arxiv.org/abs/1904.08030v1

The matching stage retrieves candidate items relevant to user interests, while the ranking stage sorts candidate items by user interests. Thus, the most critical ability is to model and represent user interests for either stage. We propose the Multi-Interest Network with Dynamic routing (MIND) for dealing with user's diverse interests in the matching stage.

3. PL-NMF: Parallel Locality-Optimized Non-negative Matrix Factorization

Gordon E. Moon, Aravind Sukumaran-Rajam, Srinivasan Parthasarathy, P. Sadayappan

https://arxiv.org/abs/1904.07935v1

Non-negative Matrix Factorization (NMF) is a key kernel for unsupervised dimension reduction used in a wide range of applications, including topic modeling, recommender systems and bioinformatics. Due to the compute-intensive nature of applications that must perform repeated NMF, several parallel implementations have been developed in the past. Efficient realizations of the algorithm on multi-core CPUs and GPUs are developed, demonstrating significant performance improvement over existing state-of-the-art parallel NMF algorithms.

4. Dynamic Learning with Frequent New Product Launches: A Sequential Multinomial Logit Bandit Problem

Junyu Cao, Wei Sun

https://arxiv.org/abs/1904.12445v1

We propose a sequential multinomial logit (SMNL) model to characterize customers' behavior when product recommendations are presented in tiers. For the offline version with known customers' preferences, we propose a polynomial-time algorithm and characterize the properties of the optimal tiered product recommendation. For the online problem, we propose a learning algorithm and quantify its regret bound.

5. Hierarchical Context enabled Recurrent Neural Network for Recommendation

Kyungwoo Song, Mingi Ji, Sungrae Park, Il-Chul Moon

https://arxiv.org/abs/1904.12674v1

The analyses on the user history require the robust sequential model to anticipate the transitions and the decays of user interests. To resolve these challenges, we suggest HCRNN with three hierarchical contexts of the global, the local, and the temporary interests. As we suggest a new RNN structure, we support HCRNN with a complementary \textit{bi-channel attention} structure to utilize hierarchical context.

6. Relational Collaborative Filtering: Modeling Multiple Item Relations for Recommendation

Xin Xin, Xiangnan He, Yongfeng Zhang, Yongdong Zhang, Joemon Jose

https://arxiv.org/abs/1904.12796v1

However, how to incorporate multiple item relations is less explored in recommendation research. In this work, we propose Relational Collaborative Filtering (RCF), a general framework to exploit multiple relations between items in recommender system. Furthermore, we also conduct qualitative analyses to show the benefits of explanations brought by the modeling of multiple item relations.

7. Inductive Graph Pattern Learning for Recommender Systems Based on a Graph Neural Network

Muhan Zhang, Yixin Chen

https://arxiv.org/abs/1904.12058v1

Most modern successful recommender systems are based on matrix factorization techniques, i.e., learning a latent embedding for each user and each item from the given rating matrix and use the embeddings to complete the matrix. However, these learned latent embeddings are inherently transductive and are not designed to generalize to unseen users/items or new tasks. In this paper, we aim to learn an inductive model for recommender systems based on the local graph patterns around user-item pairs.

8. Collaborative Filtering via High-Dimensional Regression

Harald Steck

https://arxiv.org/abs/1904.13033v1

For this reason, we focus in this paper on variants of high-dimensional regression problems that have closed-form solutions. Moreover, we motivate a re-scaling rather than a re-weighting approach for dealing with biases regarding item-popularities in the data. In experiments on three publicly available data sets, we observed not only extremely reduced training times, but also significantly improved ranking accuracy compared to SLIM.