Future Gradient Descent for Adapting the Temporal Shifting Data Distribution in Online Recommendation Systems
This addresses domain generalization for online recommendation systems, but it appears incremental as it builds on existing gradient-based methods.
The paper tackles the problem of temporal domain shift in online recommendation systems by proposing a meta future gradient generator that forecasts future gradient information for training, achieving smaller temporal domain generalization error compared to baselines like Batch Update.
One of the key challenges of learning an online recommendation model is the temporal domain shift, which causes the mismatch between the training and testing data distribution and hence domain generalization error. To overcome, we propose to learn a meta future gradient generator that forecasts the gradient information of the future data distribution for training so that the recommendation model can be trained as if we were able to look ahead at the future of its deployment. Compared with Batch Update, a widely used paradigm, our theory suggests that the proposed algorithm achieves smaller temporal domain generalization error measured by a gradient variation term in a local regret. We demonstrate the empirical advantage by comparing with various representative baselines.