Pairwise learning-to-rank algorithms have been shown to allow recommender systems to leverage unary user feedback. We propose Multi-feedback Bayesian Personalized Ranking (MF-BPR), a pairwise method that exploits different types of feedback with an extended sampling method. The feedback types are drawn from different "channels", in which users interact with items (e.g., clicks, likes, listens, follows, and purchases). We build on the insight that different kinds of feedback, e.g., a click versus a like, reflect different levels of commitment or preference. Our approach differs from previous work in that it exploits multiple sources of feedback simultaneously during the training process. The novelty of MF-BPR is an extended sampling method that equates feedback sources with "levels" that reflect the expected contribution of the signal. We demonstrate the effectiveness of our approach with a series of experiments carried out on three datasets containing multiple types of feedback. Our experimental results demonstrate that with a right sampling method, MF-BPR outperforms BPR in terms of accuracy. We find that the advantage of MF-BPR lies in its ability to leverage level information when sampling negative items.

Abstract. Factorization machines offer an advantage over other existing collaborative filtering approaches to recommendation. They make it possible to work with any auxiliary information that can be encoded as a real-valued feature vector as a supplement to the information in the user-item matrix. We build on the assumption that different patterns characterize the way that users interact with (i.e., rate or download) items of a certain type (e.g., movies or books). We view interactions with a specific type of item as constituting a particular domain and allow interaction information from an auxiliary domain to inform recommendation in a target domain. Our proposed approach is tested on a data set from Amazon and compared with a state-of-the-art approach that has been proposed for Cross-Domain Collaborative Filtering. Experimental results demonstrate that our approach, which has a lower computational complexity, is able to achieve performance improvements.

No abstract

Abstract. We developed a learning-based question classifier for question answering systems. A question classifier tries to predict the entity type of the possible answers to a given question written in natural language. We extracted several lexical, syntactic and semantic features and examined their usefulness for question classification. Furthermore we developed a weighting approach to combine features based on their importance. Our result on the well-known TREC questions dataset is competitive with the state-of-the-art on this task.

Abstract. This study aims to develop a recommender system for social learning platforms that combine traditional learning management systems with commercial social networks like Facebook. We therefore take into account social interactions of users to make recommendations on learning resources. We propose to make use of graph-walking methods for improving performance of the wellknown baseline algorithms. We evaluate the proposed graph-based approach in terms of their F1 score, which is an effective combination of precision and recall as two fundamental metrics used in recommender systems area. The results show that the graph-based approach can help to improve performance of the baseline recommenders; particularly for rather sparse educational datasets used in this study.

The 2014 ACM Recommender Systems Challenge invited researchers and practitioners to work towards a common goal, this goal being the prediction of users engagement in movie ratings expressed on Twitter. More than 200 participants sought to join the challenge and work on the new dataset released in its scope. The participants were asked to develop new algorithms to predict user engagement and evaluate them in a common setting, ensuring that the comparison was objective and unbiased in the setting of the challenge.

This paper states the case for the principle of minimal necessary data: If two recommender algorithms achieve the same effectiveness, the better algorithm is the one that requires less user data. Applying this principle involves carrying out training data requirements analysis, which we argue should be adopted as best practice for the development and evaluation of recommender algorithms. We take the position that responsible recommendation is recommendation that serves the people whose data it uses. To minimize the imposition on users' privacy, it is important that a recommender system does not collect or store more user information than it absolutely needs. Further, algorithms using minimal necessary data reduce training time and address the cold start problem. To illustrate the trade-off between training data volume and accuracy, we carry out a set of classic recommender system experiments. We conclude that consistently applying training data requirements analysis would represent a relatively small change in researchers' current practices, but a large step towards more responsible recommender systems. ACM Reference format:

User interactions can be considered to constitute different feedback channels, for example, view, click, like or follow, that provide implicit information on users’ preferences. Each implicit feedback channel typically carries a unary, positive-only signal that can be exploited by collaborative filtering models to generate lists of personalized recommendations. This article investigates how a learning-to-rank recommender system can best take advantage of implicit feedback signals from multiple channels. We focus on Factorization Machines (FMs) with Bayesian Personalized Ranking (BPR), a pairwise learning-to-rank method, that allows us to experiment with different forms of exploitation. We perform extensive experiments on three datasets with multiple types of feedback to arrive at a series of insights. We compare conventional, direct integration of feedback types with our proposed method, which exploits multiple feedback channels during the sampling process of training. We refer to our method as multi-channel sampling. Our results show that multi-channel sampling outperforms conventional integration, and that sampling with the relative “level” of feedback is always superior to a level-blind sampling approach. We evaluate our method experimentally on three datasets in different domains and observe that with our multi-channel sampler the accuracy of recommendations can be improved considerably compared to the state-of-the-art models. Further experiments reveal that the appropriate sampling method depends on particular properties of datasets such as popularity skewness.