Beyond One-Size-Fits-All: Adaptive Test-Time Augmentation for Sequential Recommendation

2026-04-17

AI Summary

This work addresses the limitation of existing test-time augmentation (TTA) methods in sequential recommendation, which employ uniform strategies and overlook the heterogeneity of user behaviors, thereby constraining performance. To overcome this, the authors propose AdaTTA, a reinforcement learning–based adaptive inference framework that, for the first time, formulates the selection of augmentation operators as a Markov decision process. AdaTTA leverages an Actor-Critic architecture with hybrid state representations and a joint macro-ranking reward to dynamically determine the optimal augmentation strategy for each user. The framework introduces a plug-and-play adaptive mechanism that transcends the conventional one-size-fits-all paradigm. Extensive experiments demonstrate that AdaTTA significantly outperforms the best fixed-strategy baselines across four real-world datasets and two backbone models, achieving a relative improvement of up to 26.31% on the Home dataset while maintaining manageable computational overhead.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Reasoning under Uncertainty: Sequential Decision Making
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Test-time augmentation (TTA) has become a promising approach for mitigating data sparsity in sequential recommendation by improving inference accuracy without requiring costly model retraining. However, existing TTA methods typically rely on uniform, user-agnostic augmentation strategies. We show that this "one-size-fits-all" design is inherently suboptimal, as it neglects substantial behavioral heterogeneity across users, and empirically demonstrate that the optimal augmentation operators vary significantly across user sequences with different characteristics for the first time. To address this limitation, we propose AdaTTA, a plug-and-play reinforcement learning-based adaptive inference framework that learns to select sequence-specific augmentation operators on a per-sequence basis. We formulate augmentation selection as a Markov Decision Process and introduce an Actor-Critic policy network with hybrid state representations and a joint macro-rank reward design to dynamically determine the optimal operator for each input user sequence. Extensive experiments on four real-world datasets and two recommendation backbones demonstrate that AdaTTA consistently outperforms the best fixed-strategy baselines, achieving up to 26.31% relative improvement on the Home dataset while incurring only moderate computational overhead

Problem

• test-time augmentation
• sequential recommendation
• user heterogeneity
• adaptive inference
• data sparsity

Innovation

• Adaptive Test-Time Augmentation
• Sequential Recommendation
• Reinforcement Learning
• User Heterogeneity
• Plug-and-Play Framework