TaleBrush: Visual Sketching of Story Generation with Pretrained Language Models скачать в хорошем качестве

TaleBrush: Visual Sketching of Story Generation with Pretrained Language Models

TaleBrush: Visual Sketching of Story Generation with Pretrained Language Models John Joon Young Chung, Wooseok Kim, Kang Min Yoo, Hwaran Lee, Eytan Adar, Minsuk Chang CHI'22: ACM Conference on Human Factors in Computing Systems Session: Interactivity Abstract Human-AI story co-creation became more plausible with advancing text generation algorithms (e.g., GPT-3). Simple turn-taking has often been the main mode of co-creation, but it is yet difficult for writers to intuitively understand the AI’s generation or guide the iterative generation. We introduce the demo of TaleBrush, a GPT-based generative story ideation tool that leverages line sketching to allow controlling and sensemaking the protagonist’s fortune in co-created stories. We describe how sketching interaction can be designed for easy and expressive control. Then, we explain how TaleBrush adopts sketching interaction to facilitate control and sensemaking of story co-creation. We hope this demonstration to provoke the HCI community to contribute to more designs of control and sensemaking interactions for generative human-AI co-creation systems. WEB:: https://chi2022.acm.org/ Pre-recorded presentations of CHI 2022