Multi-turn GUI agents enable complex task completion through sequential decision-making, but suffer from severe context inflation as interaction history accumulates. Existing strategies either sacr...
<p><strong>Tác giả:</strong> Yurun Song, Jiong Yin, Rongjunchen Zhang, Ian G. Harris</p>
<p>Multi-turn GUI agents enable complex task completion through sequential decision-making, but suffer from severe context inflation as interaction history accumulates. Existing strategies either sacrifice long-term context via truncation or compromise spatial structure through token pruning. In this paper, we propose Coordinate Compression Policy Optimization (CCPO), an efficient policy optimization framework that couples visual compression with policy optimization for multi-turn GUI agents. CC...</p>
<p>📄 DOI: <a href="https://doi.org/https://doi.org/10.1145/3767308.3835478" target="_blank" rel="noopener">https://doi.org/10.1145/3767308.3835478</a></p>
<p>📖 <a href="https://doi.org/https://doi.org/10.1145/3767308.3835478" target="_blank" rel="noopener">Xem bài báo đầy đủ</a></p>
Bài viết được biên soạn lại bởi LIFVietnam. Nguồn tham khảo: OpenAlex
