Gigaom Minute – June 16

Transcript

• IEEE Spectrum reports researchers at UNC and the Washington University School of Medicine have applied machine learning to successfully predict Autism in infants. Of 59 6-month-old infants, the algorithm correctly predicted 9 of 11 who would later receive a positive diagnosis for autism. Author of the subsequent UNC report, Robert Emerson, notes the team’s work could lead to the creation of “a highly sensitive and accurate early diagnostic test for autism.” He adds, “AI is going to be key to making that happen.” Read more.

• Researchers from the KTH Royal Institute of Technology are hoping AI can help translate the language and thoughts of dolphins. Help may come from Gavagai AB, a Swedish language technology company which has already mastered 40 human languages with their AI-based language analysis software. Now the team will tackle the sounds of bottlenose dolphins in an effort to compile a dictionary of dolphin language. Read more.

• While AI systems are great at recognizing objects in images, AI still struggles to do the same with videos. However, DeepMind is addressing that problem with its Kinetics dataset to help improve AI’s capability to recognize human actions in motion. Research on the data set, consisting of 300,000 video clips and 400 human action classes, promises to improve AI computer vision for video. Read more.

• Google Assistant now features a new App Directory with more than 40 voice apps offering users a multitude of new abilities. Among those is the “This Day in History” action, a new voice app from the History Channel. Not to be confused with an Alexa skill by the same name, the app delivers a synopsis of historical events for any day of the year. Read more.