Bringing GRACE to Recommendation: Fine-Tuning for Sustainable and Accurate Personalization

2026-07-24

AI Summary

This work addresses the challenge that existing green recommender systems often incur high energy consumption due to full retraining or suffer from inference latency caused by post-hoc reranking, making it difficult to jointly achieve sustainability and personalized accuracy. To overcome this, the authors propose GRACE, a framework that fine-tunes pretrained recommendation models by transforming discrete, non-differentiable sustainability signals—such as ecological scores—into differentiable approximations. GRACE further introduces a preference-anchored gradient projection mechanism that steers optimization toward sustainability objectives while effectively preserving recommendation accuracy. Experiments on real-world datasets demonstrate that GRACE significantly enhances the sustainability of recommendations without requiring reranking or full retraining, thereby enabling efficient and controllable multi-objective balancing.

Abstract

Growing concern about environmental sustainability (e.g., reducing carbon emissions and resource use) and public health has motivated ``green'' recommender systems that steer users toward more eco-friendly and healthier choices. However, many existing green recommendation approaches require training new models from scratch, incurring substantial computational and energy costs. Reranking-based methods, meanwhile, introduce an additional sorting stage at inference, increasing latency and computational cost. In this work, we propose GRACE (Green Recommendation via Adaptive Conflict-rEsolution), a fine-tuning framework that integrates item-level sustainability signals (e.g., eco-scores or health indices) into pretrained recommendation models. Since these green values are usually discrete and non-differentiable, existing methods often rely on pairwise comparisons to promote greener items. GRACE instead introduces a differentiable approximation that enables direct optimization of the green criterion. To balance sustainability and personalization quality, GRACE further employs a gradient projection mechanism to mitigate conflicts between the green objective and the accuracy objective during fine-tuning. Experiments on real-world datasets demonstrate that GRACE improves sustainability-oriented recommendation outcomes while generally preserving recommendation accuracy through a controllable preference-anchored update mechanism.

Problem

• green recommendation
• sustainability
• personalization
• computational cost
• recommendation accuracy

Innovation

• green recommendation
• differentiable approximation
• gradient projection
• fine-tuning
• sustainability-aware personalization