adobe editagent 利用 AI 編輯圖像媲美專業人士

adobe research and uc san diego built editagent, an ai system that breaks image editing tasks into sub-steps and pairs a general-purpose editor with specialized tools. editagent generates candidates with different methods in parallel and uses a standard language model to compare them, producing more accurate results than previous ai agents. the system uses verification records with close-ups and measured image statistics to help the language model pick the best result. editagent outperformed evaluation models specifically trained for the job, with qwen3.8-27b making a usable pick 81.3 percent of the time. however, editagent needs about ten times the compute.