AI News, Story highlights by Amir Ziai artificial intelligence

This AI Watched 100 Films to Learn How to Recognize a Kiss

Post Syndicated from Jeremy Hsu original https://spectrum.ieee.org/tech-talk/robotics/artificial-intelligence/ai-learns-how-to-pucker-up-from-hollywood The study of AI-based kiss detection came fromAmir Ziai,a senior data scientist at Netflix, as he was completingcoursework to obtain anAI graduate certificate at Stanford University.

Lest anyone get the wrong impression, it’s still unclear whether or not the kiss detection methodworks with more sexual scenes that go beyond kissing.“In my training set, I’ve stayed away from overly sexual scenes to make sure that the model is not confusing kissing and sex,” Ziai says.

Ziai’s demonstration of kiss detection technology that works with videoshints at future applications that couldautomatically categorize video content, create personalized video recommendations for viewers, and possibly even screenout certain videos as part of online content moderation.

“This is a good example of how modern computer vision techniques make it fairly easy to develop specific ‘sense and respond’ software, cued to qualitative/unstructured things (like the presence of kissing in a scene),” said Jack Clark, strategy and communications director at OpenAI, inhisImport AI newsletter, whichrecently highlighted the kiss detection study.“I think this is one of the most under-hyped aspects of how AI is changing the scope of individual software development.” When it came time to visually identify kissing scenes, the deep learning modelthat proved most successful was ResNet-18, an image classification algorithm that was already pre-trained on more than one million images from the popular ImageNet database.

Thattwo-pronged approach of training AI to processboth images and audio of kissing helped the overallmodelachieveafairly impressive F1 score of 0.95—a measure that represents the weighted average of the algorithm’s accuracy regarding both false positives and false negatives.

This AI Watched 100 Films to Learn How to Recognize a Kiss

Like someone who hasnever been kissed, AI began learning the basics bybinge-watching romanticfilm clipsto seehow Hollywood stars lock lips.

By training deep learning algorithms that have already proven adept at recognizing faces and objectsto also recognize steamykissing scenes dramatized by professional actors, a data scientist has shown how AI systemscould gain greater insight into the mostintimate human activities.

Lest anyone get the wrong impression, it’s still unclear whether or not the kiss detection methodworks with more sexual scenes that go beyond kissing.“In my training set, I’ve stayed away from overly sexual scenes to make sure that the model is not confusing kissing and sex,” Ziai says.

Ziai’s demonstration of kiss detection technology that works with videoshints at future applications that couldautomatically categorize video content, create personalized video recommendations for viewers, and possibly even screenout certain videos as part of online content moderation.

“This is a good example of how modern computer vision techniques make it fairly easy to develop specific ‘sense and respond’ software, cued to qualitative/unstructured things (like the presence of kissing in a scene),” said Jack Clark, strategy and communications director at OpenAI, inhisImport AI newsletter, whichrecently highlighted the kiss detection study.“I think this is one of the most under-hyped aspects of how AI is changing the scope of individual software development.” When it came time to visually identify kissing scenes, the deep learning modelthat proved most successful was ResNet-18, an image classification algorithm that was already pre-trained on more than one million images from the popular ImageNet database.