김현수 석사 졸업생, 문재완 박사과정, 박성민 석박통합과정 KDD 2026 국제 학술대회 논문 채택

24 Nov 2025

DIAL 연구실 소속 인공지능학과 김현수(석사 졸업생, 공동 제1저자) 학생, 문재완(박사과정, 공동 제1저자) 학생, 박성민(석박통합과정, 제3저자) 학생, 이종욱(교신저자) 교수가 참여한 논문 "MergeRec: Model Merging for Data-Isolated Cross-Domain Sequential Recommendation"이 데이터 마이닝 분야 최우수 국제 학술대회인 The 32nd International ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD)에 게재 승인되었으며 2026년 8월에 대한민국 제주에서 발표될 예정입니다. 

Abstract

Modern recommender systems trained on domain-specific data often struggle to generalize across multiple domains. In this work, we explore a novel approach called model merging, which integrates the parameters of multiple fine-tuned models into a unified model without requiring access to raw user interaction data. We propose MergeRec, a new framework for cross-domain sequential recommendation. MergeRec consists of three key components: (1) merging initialization, (2)pseudo-user data construction, and (3) collaborative merging optimization. First, we initialize a merged model using training-free merging techniques. Next, we construct pseudo-user data by utilizing a single item as a virtual sequence from each domain, enabling the synthesis of meaningful training samples without relying on real user interactions. Finally, we optimize domain-specific merging weights through a joint objective that combines a recommendation loss, which encourages the merged model to identify relevant items, and a distillation loss, which transfers collaborative filtering signals from the fine-tuned source models. Extensive experiments demonstrate that MergeRec not only preserves the strengths of the original models but also significantly enhances generalizability to unseen domains. Compared to conventional model merging methods, MergeRec consistently achieves superior performance, with average improvements of up to 17.21% on Recall@10, highlighting the potential of model merging as a scalable and effective solution for building universal recommender systems. The source code is available at https://github.com/jaewan7599/MergeRec_KDD2026.