Now showing 1 - 10 of 25
  • Publication
    Topic Extraction from Online Reviews for Classification and Recommendation
    Automatically identifying informative reviews is increasingly important given the rapid growth of user generated reviews on sites like Amazon and TripAdvisor. In this paper, we describe and evaluate techniques for identifying and recommending helpful product reviews using a combination of review features, including topical and sentiment information, mined from a review corpus.
  • Publication
    A Distributed Asynchronous Deep Reinforcement Learning Framework for Recommender Systems
    In this paper we propose DADRL, a distributed, asynchronous reinforcement learning recommender system based on the asynchronous advantage actor-critic model (A3C), which combines ideas from A3C and federated learning (FL). The proposed algorithm keeps the user preferences or interactions on local devices and uses a combination of on-device, local recommendation models and a complementary global model. The global model is trained only by the loss gradients of the local models, rather than directly using user preferences or interactions data. We demonstrate, using well-known datasets and benchmark algorithms, how this approach can deliver performance that is comparable with the current state-of-the-art while enhancing user privacy.
  • Publication
    Personalised Opinion-based Recommendation
    (Springer, 2016-11-02) ;
    E-commerce recommender systems seek out matches betweencustomers and items in order to help customers discover more relevantand satisfying products and to increase the conversion rate of browsers tobuyers. To do this, a recommender system must learn about the likes anddislikes of customers/users as well as the advantages and disadvantages(pros and cons) of products. Recently, the explosion of user-generatedcontent, especially customer reviews, and other forms of opinionated expression,has provided a new source of user and product insights. Theinterests of a user can be mined from the reviews that they write andthe pros and cons of products can be mined from the reviews writtenabout them. In this paper, we build on recent work in this area to generateuser and product proles from user-generated reviews. We furtherdescribe how this information can be used in various recommendationtasks to suggest high-quality and relevant items to users based on eitheran explicit query or their prole. We evaluate these ideas using alarge dataset of TripAdvisor reviews. The results show the benets ofcombining sentiment and similarity in both query-based and user-basedrecommendation scenarios, and also disclose the eect of the number ofreviews written by a user on recommendation performance.
      339Scopus© Citations 6
  • Publication
    Overlap Training to Mitigate Inconsistencies Caused by Image Tiling in CNNs
    (Springer, 2020-12-08) ; ; ;
    This paper focuses on the problem of inconsistent predictions of modern convolutional neural networks (CNN) at patch (i.e. sub-image) boundaries. Limited by the graphics processing unit (GPU) resources, image tiling and stitching countermeasure have been applied for most megapixel images, that is, cutting images into overlapping tiles as CNN input, and then stitching CNN outputs together. However, we found that stitched (i.e. recovered) predictions have discontinuous grid-like noise. We propose a simple yet efficient overlap training framework to mitigate the inconsistent prediction at patch boundaries without changing the model architecture while improving the stability, robustness of the model. We have applied our solution to various CNNs (such as U-Net, DeepLab, RCF) and tested them on two real-world datasets. Extensive experiments suggest that the new framework is sufficient in reducing inconsistency and outperform these countermeasures. The source code and coloured figures are made publicly available online at:
      219Scopus© Citations 3
  • Publication
    UP-TreeRec: Building dynamic user profiles tree for news recommendation
    Online News recommendation systemsaim to address the information explosion of news and make personalized recommendation for users. The key problem of personalized news recommendation is to model users’ interests and track their changes. A common way to deal with usermodeling problem is to build user profiles from observed behavior. However, the majority of existing methods make static representation to user profiles and little research has focused on the effective user modeling that could dynamically capture user interests in news topics.To address this problem, in this paper, we propose UP-TreeRec, a news recommendation framework based on user profile tree (UP-Tree), which is a novel framework combining content-based and collaborative filtering techniques. First, by exploitinga novel topic model namely UI-LDA, we obtain the representationvectorsfor news content in topic spaceas the fundamentalbridgeto associate user interests with news topics. Next, we design a decision tree with a dynamically changeable structure to construct a user interest profile from his feedback. Furthermore, we present a clustering based multidimensional similarity computation method to select the nearest neighbor of UP-Treeefficiently. We also provide a Map-Reduce framework based implementation that enables scaling our solution to real-world news recommendation problems.We conducted several experiments compared to the state-of-the-art approaches on real-world datasets and the experimental results demonstrate that our approach significantly improves the accuracy and effectiveness in news recommendation.
      86Scopus© Citations 5
  • Publication
    Opinionated Product Recommendation
    In this paper we describe a novel approach to case-based product recommendation. It is novel because it does not leverage the usual static, feature-based, purely similarity-driven approaches of traditional case-based recommenders. Instead we harness experiential cases, which are automatically mined from user generated reviews, and we use these as the basis for a form of recommendation that emphasises similarity and sentiment. We test our approach in a realistic product recommendation setting by using live-product data and user reviews.
      623Scopus© Citations 38
  • Publication
    The Demonstration of the Reviewer's Assistant
    User generated reviews are now a familiar and valuable part of most e-commerce sites since high quality reviews are known to influence purchasing decisions. In this demonstration we describe work on the Reviewer's Assistant (RA), which is a recommendation system that is designed to help users to write better quality reviews. It does this by suggesting relevant topics that they may wish to discuss based on the product they are reviewing and the content of their review so far.
  • Publication
    From More-Like-This to Better-Than-This: Hotel Recommendations from User Generated Reviews
    (ACM, 2016-07-17) ;
    To help users discover relevant products and items recommender systems must learn about the likes and dislikes of users and the pros and cons of items. In this paper, we present a novel approach to building rich feature-based user profiles and item descriptions by mining user-generated reviews. We show how this information can be integrated into recommender systems to deliver better recommendations and an improved user experience.
      182Scopus© Citations 8
  • Publication
    Coevolutionary Recommendation Model: Mutual Learning between Ratings and Reviews
    Collaborative filtering (CF) is a common recommendation approach that relies on user-item ratings. However, the natural sparsity of user-item rating data can be problematic in many domains and settings, limiting the ability to generate accurate predictions and effective recommendations. Moreover, in some CF approaches latent features are often used to represent users and items, which can lead to a lack of recommendation transparency and explainability. User-generated, customer reviews are now commonplace on many websites, providing users with an opportunity to convey their experiences and opinions of products and services. As such, these reviews have the potential to serve as a useful source of recommendation data, through capturing valuable sentiment information about particular product features. In this paper, we present a novel deep learning recommendation model, which co-learns user and item information from ratings and customer reviews, by optimizing matrix factorization and an attention-based GRU network. Using real-world datasets we show a significant improvement in recommendation performance, compared to a variety of alternatives. Furthermore, the approach is useful when it comes to assigning intuitive meanings to latent features to improve the transparency and explainability of recommender systems.
      841Scopus© Citations 114
  • Publication
    Multi-level Attention-Based Neural Networks for Distant Supervised Relation Extraction
    We propose a multi-level attention-based neural network forrelation extraction based on the work of Lin et al. to alleviate the problemof wrong labelling in distant supervision. In this paper, we first adoptgated recurrent units to represent the semantic information. Then, weintroduce a customized multi-level attention mechanism, which is expectedto reduce the weights of noisy words and sentences. Experimentalresults on a real-world dataset show that our model achieves significantimprovement on relation extraction tasks compared to both traditionalfeature-based models and existing neural network-based methods