Session: Domain-Specific Recommendation

Date: Monday September 19, 4:00 PM – 5:45 PM (PDT)

  • PALearning Recommendations from User Actions in the Item-poor Insurance Domain
    by Simone Borg Bruun (University of Copenhagen, Denmark), Maria Maistro (University of Copenhagen, Denmark), Christina Lioma (University of Copenhagen, Denmark)

    While personalised recommendations are successful in domains like retail, where large volumes of user feedback on items are available, the generation of automatic recommendations in data-sparse domains, like insurance purchasing, is an open problem. The insurance domain is notoriously data-sparse because the number of products is typically low (compared to retail) and they are usually purchased to last for a long time. Also, many users still prefer the telephone over the web for purchasing products, reducing the amount of web-logged user interactions. To address this, we present a recurrent neural network recommendation model that uses past user sessions as signals for learning recommendations. Learning from past user sessions allows dealing with the data scarcity of the insurance domain. Specifically, our model learns from several types of user actions that are not always associated with items, and unlike all prior session-based recommendation models, it models relationships between input sessions and a target action (purchasing insurance) that does not take place within the input sessions. Evaluation on a real-world dataset from the insurance domain (ca. 44K users, 16 items, 54K purchases, and 117K sessions) against several state-of-the-art baselines shows that our model outperforms the baselines notably. Ablation analysis shows that this is mainly due to the learning of dependencies across sessions in our model. We contribute the first ever session-based model for insurance recommendation, and make available our dataset to the research community.

    Full text in ACM Digital Library

  • INReusable Self-Attention Recommender Systems in Fashion Industry Applications
    by Marjan Celikik (Zalando SE, Germany), Ana Peleteiro Ramallo (Zalando SE, Germany), Jacek Wasilewski (Zalando SE, Germany)

    A large number of empirical studies on applying self-attention models in the domain of recommender systems are based on offline evaluation and metrics computed on standardized datasets. Moreover, many of them do not consider side information such as item and customer metadata although deep-learning recommenders live up to their full potential only when numerous features of heterogeneous type are included. Also, normally the model is used only for a single use case. Due to these shortcomings, even if relevant, previous works are not always representative of their actual effectiveness in real-world industry applications. In this talk, we contribute to bridging this gap by presenting live experimental results demonstrating improvements in user retention of up to 30%. Moreover, we share our learnings and challenges from building a re-usable and configurable recommender system for various applications from the fashion industry. In particular, we focus on fashion inspiration use-cases, such as outfit ranking, outfit recommendation and real-time personalized outfit generation.

    Full text in ACM Digital Library

  • PAMulti-Modal Dialog State Tracking for Interactive Fashion Recommendation
    by Yaxiong Wu (University of Glasgow, United Kingdom), Craig Macdonald (University of Glasgow, United Kingdom), Iadh Ounis (University of Glasgow, United Kingdom)

    Multi-modal interactive recommendation is a type of task that allows users to receive visual recommendations and express natural-language feedback about the recommended items across multiple iterations of interactions. However, such multi-modal dialog sequences (i.e. turns consisting of the system’s visual recommendations and the user’s natural-language feedback) make it challenging to correctly incorporate the users’ preferences across multiple turns. Indeed, the existing formulations of interactive recommender systems suffer from their inability to capture the multi-modal sequential dependencies of textual feedback and visual recommendations because of their use of recurrent neural network-based (i.e., RNN-based) or transformer-based models. To alleviate the multi-modal sequential dependency issue, we propose a novel multi-modal recurrent attention network (MMRAN) model to effectively incorporate the users’ preferences over the long visual dialog sequences of the users’ natural-language feedback and the system’s visual recommendations. Specifically, we leverage a gated recurrent network (GRN) with a feedback gate to separately process the textual and visual representations of natural-language feedback and visual recommendations into hidden states (i.e. representations of the past interactions) for multi-modal sequence combination.
    In addition, we apply a multi-head attention network (MAN) to refine the hidden states generated by the GRN and to further enhance the model’s ability in dynamic state tracking. Following previous work, we train and evaluate our MMRAN model by using a vision-language transformer-based user simulator as a surrogate for real human users. Thorough and extensive experiments — conducted on the Fashion IQ Dresses, Shirts, and Tops & Tees datasets — show that our proposed MMRAN model yields significant improvements in comparison to several existing state-of-the-art baseline models.

    Full text in ACM Digital Library

  • INRethinking Personalized Ranking at Pinterest: An End-to-End Approach
    by Jiajing Xu (Pinterest, United States)

    In this work, we present our journey to revolutionize the recommendation ranking engine through end-to-end learning from raw user actions. We encode user’s long-term interest in PinnerFormer, a user embedding optimizing for long-term future actions via a new dense all-action loss, and capture user’s short-term intent by directly learning from the realtime action sequences. We conducted both offline and online experiments to validate the performance of the new model architecture, and also address the challenge of serving such a complex model using mixed CPU/GPU setup in production. The proposed system has been deployed in production at Pinterest and has delivered significant online gains across applications.

    Full text in ACM Digital Library

  • INTranslating the Public Service Media Remit into Metrics and Algorithms
    by Andreas Grün (ZDF, Germany), Xenija Neufeld (Accso – Accelerated Solutions GmbH, Germany)

    After multiple years of providing automated video recommendations in the ZDFmediathek, ZDF has established a solid ground for the usage of recommender systems. Being a Public Service Media (PSM) provider, our most important driver on this journey is our Public Service Media Remit (PSMR). We are committed to cultivate PSM values such as diversity, fairness, and transparency while providing fresh and relevant content. Therefore, it is important for us to not only measure the success of our recommender systems in terms of basic business Key Performance Indicators (KPIs) such as clicks and viewing minutes but also to ensure and to measure the achievement of PSM values. While speaking about PSM values, however, it is important to keep in mind that there is no easy way to directly measure values as such. In order to be able to measure their extent in a recommender system, we need to translate these values into public value metrics. However, not only the final results are essential for the PSMR. Additionally, it is highly important to establish transparency while working towards these results, that is, while defining the data, the algorithms, and the pipelines used in recommender systems. In our talk we will provide a deeper insight into how we approach this task with Model Cards and give an overview of some models, their Model Cards, and metrics that we are currently using for ZDFmediathek.

    Full text in ACM Digital Library

Diamond Supporter
 
Platinum Supporters
 
 
 
Gold Supporters
 
 
 
 
 
 
Challenge Sponsor
 
Special Supporters
In-Cooperation