Abstract：A Deep Neural Architecture for Decision-Aware Meta-Review Generation Automatically generating meta-reviews from peer-reviews is a new and challenging taskAlthough closethe task is not precisely summarizing the peer-reviewsUsuallya conference chair or a journal editor writes a meta-review after going through the rev Consulting ServiceCan assist librarians or researchers in quickly understanding and summarizing literature reviews or feedbackThis study focuses on using deep learning technologyparticularly a transformer-based multi-encoder architectureto automatically generate meta-reviews and predict decisions in the peer review processMeta-reviews synthesize the opinions of peer reviewers and summarize the review processs outcomeAutomatically generating meta-reviews is a challenging new task in NLPand this study attempts to address this issue