ICMRec: Item Cluster-Wise Multi-Objective Optimization for Unbiased Recommendation

The traditional observed data used to train the recommender model suffers from severe bias issues (e.g., exposure bias, popularity bias). Interactions of a small fraction of head items account for almost the whole training data. The normal training paradigm from such biased data tends to repetitively generate recommendations from the head items, which further exacerbates the biases and affects the exploration of potentially interesting items from the niche set. In this work, distinct from existing methods, we innovatively explore the central theme of unbiased recommendation from an item cluster-wise multi-objective optimization perspective. Aiming to balance the learning on various item clusters that differ in popularity during the training process, we characterize the recommendation task as an item cluster-wise multi-objective optimization problem. To this end, we propose a model-agnostic framework namely Item Cluster-Wise Multi-Objective Recommendation (ICMRec) for unbiased recommendation. In detail, we define our item cluster-wise optimization target that the recommender model should balance all item clusters that differ in popularity. Thus we set the model learning on each item cluster as a unique optimization objective. To achieve this goal, we first explore items' popularity levels from a novel causal reasoning perspective. Then, we devise popularity discrepancy-based bisecting clustering to separate the discriminated item clusters. Next, we adaptively find the overall harmonious gradient direction for multiple item cluster-wise optimization objectives from a Pareto-efficient solver. Finally, in the prediction stage, we perform counterfactual inference to further eliminate the impact of user conformity. Extensive experimental results demonstrate the superiorities of ICMRec on overall recommendation performance and biases elimination. Codes will be open-source upon acceptance.