[NetEase Smart News, August 25th] Recently, Disney's research team has been exploring some fascinating new technologies, including real-time voice animation, shared augmented reality experiences, and some impressive live performance innovations. Currently, researchers from Disney and the University of Massachusetts Boston are collaborating on neural networks capable of evaluating short stories. While these artificial intelligence systems haven't reached the level of professional literary critics in analyzing narratives, they attempt to predict which stories might resonate most with audiences.
"Our neural network has shown promising results in predicting the popularity of a story," said Disney research scientist Boyang "Albert" Li in a recent statement. "While you can't rely on them to choose winners for your local writing contest just yet, they do offer valuable insights that could guide future studies."
To develop this technology, researchers utilized a vast database from Quora, a popular social networking platform where many responses often take the form of personal narratives. These stories' popularity is gauged by reader upvotes, providing a tangible metric for narrative quality. The team gathered approximately 55,000 answers, categorizing over 28,000 as stories, averaging around 369 words per entry. They then built multiple neural networks: one focused on examining various segments of each story and another designed to grasp the overall meaning of the narrative. Each AI predicts the relative popularity of a given story.
Both neural networks proved more effective than traditional text-based assessments in determining a story's appeal. However, the comprehensive network outperformed the segment-focused model by about 18%.
It's easy to envision how film studios might adopt future iterations of this technology to help select scripts for upcoming projects. Nevertheless, this field remains in its early stages. Researchers are keen to refine this tool to evaluate narrative quality beyond mere popularity. After all, no one desires another Transformers-style movie.
(Selected from Engadget, translated by NetEase Smart Editor Robot, reviewed by Li Muzhen)
Mccb Breaker,Mccb Circuit Breaker,Overload Protection Mccb,Moulded Case Circuit Breaker
ZHEJIANG QIANNA ELECTRIC CO.,LTD , https://www.traner-elec.com