Two-Stage Neural Contextual Bandits for Personalised News Recommendation
This work addresses the challenge of long-term recommendation quality for news platforms by efficiently exploring large item spaces, though it is incremental in building on existing contextual bandits methods.
The paper tackles the problem of biased feedback loops in personalised news recommendation by introducing a two-stage hierarchical topic-news deep contextual bandits framework to balance exploitation and exploration, showing that it outperforms baseline bandit policies on a large-scale dataset.
We consider the problem of personalised news recommendation where each user consumes news in a sequential fashion. Existing personalised news recommendation methods focus on exploiting user interests and ignores exploration in recommendation, which leads to biased feedback loops and hurt recommendation quality in the long term. We build on contextual bandits recommendation strategies which naturally address the exploitation-exploration trade-off. The main challenges are the computational efficiency for exploring the large-scale item space and utilising the deep representations with uncertainty. We propose a two-stage hierarchical topic-news deep contextual bandits framework to efficiently learn user preferences when there are many news items. We use deep learning representations for users and news, and generalise the neural upper confidence bound (UCB) policies to generalised additive UCB and bilinear UCB. Empirical results on a large-scale news recommendation dataset show that our proposed policies are efficient and outperform the baseline bandit policies.