Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces DCEO (Direct Causal Effect Optimization), a novel framework designed to optimize item-level proxy scores in e-commerce search systems by directly aligning them with long-term user objectives like cumulative purchases. By employing an actor-critic approach, DCEO dynamically generates context-dependent fusion weights that enhance the personalization of search rankings, addressing the limitations of traditional multi-objective fusion methods that rely on manually tuned weights. The framework was validated through extensive offline experiments and a large-scale A/B test, demonstrating a 0.36% improvement in gross merchandise value (GMV) over conventional methods.
DCEO outperforms traditional methods by directly optimizing item-level scores for long-term user value, achieving a measurable increase in GMV in real-world applications.
Industrial e-commerce search systems ultimately aim to optimize the user-level long-term objective, such as n-day cumulative purchases or gross merchandise value (GMV) per user. However, such objectives are defined at the user level, whereas search ranking is based on item-level scores within each request. Existing methods typically bridge this granularity gap through manually designed multi-objective fusion, where predictions of multiple item-level objectives, such as clicks, carts, purchases, and transaction value, are combined into a ranking score that serves as a proxy for the ultimate objective. Such hand-crafted fusion schemes rely on a small set of manually tuned weights, limiting fine-grained personalization and leading to suboptimal alignment with the ultimate objective. In this paper, we propose DCEO (Direct Causal Effect Optimization), a data-driven framework for learning item-level proxy scores that are better aligned with the ultimate objective. We first aggregate the item-level proxy scores into a user-level proxy metric and quantify its alignment with the ultimate objective using a relative causal effect. We then develop an actor-critic framework, where the critic estimates the ultimate objective for a given user-level proxy metric, and the actor dynamically generates context-dependent fusion weights over multiple objectives to construct the item-level proxy scores and is trained to directly optimize the relative causal effect. Extensive offline experiments and analyses demonstrate the effectiveness and interpretability of DCEO. In addition, DCEO has been deployed in a large-scale industrial e-commerce search system, outperforming the conventional GMV proxy by 0.36% in GMV in a 41-day online A/B test.