Introducing Context into Recommender Systems

AAAI Conferences

In this paper, we give an overview of our work to investigate the integration of context into different kind of recommender systems. Context adds an additional another dimension to the user-item data model of recommender system and can be utilized in different ways during contentbased or collaborative recommendation processes. We give several application examples we are working on to apply contextual recommenders in real world scenarios.


How to evaluate Recommender Systems – Carlos Pinela – Medium

#artificialintelligence

We have seen a variety of Recommender Systems. But we left an important issue aside: How do we evaluate RecSys? Before answering that question per se, I want to make emphasys on something. Using just one error metric can give us a limited view of how these systems work. We should always try to evaluate with different methods our models, almost as picky as your ex, but prorizing quick iteration with the lowest cost possible.


A Fairness-aware Hybrid Recommender System

arXiv.org Machine Learning

Recommender systems are used in variety of domains affecting people's lives. This has raised concerns about possible biases and discrimination that such systems might exacerbate. There are two primary kinds of biases inherent in recommender systems: observation bias and bias stemming from imbalanced data. Observation bias exists due to a feedback loop which causes the model to learn to only predict recommendations similar to previous ones. Imbalance in data occurs when systematic societal, historical, or other ambient bias is present in the data. In this paper, we address both biases by proposing a hybrid fairness-aware recommender system. Our model provides efficient and accurate recommendations by incorporating multiple user-user and item-item similarity measures, content, and demographic information, while addressing recommendation biases. We implement our model using a powerful and expressive probabilistic programming language called probabilistic soft logic. We experimentally evaluate our approach on a popular movie recommendation dataset, showing that our proposed model can provide more accurate and fairer recommendations, compared to a state-of-the art fair recommender system.


Happy Movie: A Group Recommender Application in Facebook

AAAI Conferences

In this paper we introduce our recommender Happy Movie, a Facebook application for movie recommendation to groups. This system exploits information about the social relationships and behaviour of the users to provide better recommendations. Our previous works have shown that social factors improve the recommendation results. However it required many questionnaires to be filled for obtaining the social information, so we have moved to a social network environment where this information is easily available.


Introduction to Recommender System. Part 2 (Neural Network Approach)

#artificialintelligence

Spotlight is a well-implemented python framework for constructing a recommender system. It contains two major types of models, factorization model and sequence model. The former one makes use of the idea behind SVD, decomposing the utility matrix (the matrix that records the interaction between users and items) into two latent representation of user and item matrices, and feeding them into the network. The latter one is built with time-series model such as Long Short-term Memory (LSTM) and 1-D Convolutional Neural Networks (CNN). Since the backend of Spotlight is PyTorch, make sure you have installed proper version of PyTorch before using it.