Weird realistic-looking child robot can mimic facial expressions

Daily Mail - Science & tech

An eerie robot with the face of a small child can make realistic-looking facial expressions. Creepy footage shows Affetto, an android with just a head and no body mimic human expressions like smiling and frowning. The robot was made by researchers from Osaka University in Japan who say it could open the door for androids to have'deeper interactions with humans'. Affetto, who has flesh-coloured skin on its face, can mimic a range of human expressions with incredible accuracy. An eerie robot with the face of a small child can make realistic-looking facial expressions.

Masked Julie fish can tell individual faces apart

Daily Mail - Science & tech

The ability to recognise faces was thought to be too complex for a fish, but a new study suggests that this might not be the case. Scientists have found that the tiny striped Masked Julie, which lives among the rocks of Lake Tanganyika in East Africa, can distinguish a friend from a foe. The species detects patterns around the eyes to distinguish individual faces, a skill thought to be limited to mammals and birds. Scientists have found that the tiny striped Masked Julie (pictured) can distinguish a friend from a foe by detecting unfamiliar patterns around the eyes. Eight adult male Masked Julie fish were placed them in a tank by researchers at Osaka City University.

No hiding at the back! Teacher uses facial recognition technology to see if students are BORED

Daily Mail - Science & tech

The days of dozing off at the back of a classroom may soon be coming to an end. A Chinese university lecturer is using facial-recognition technology on his students to check if they're bored – and he says it could be used in wider education. Professor Wei Xiaoyong, who lectures in computer science at Sichuan University in China, developed the'face reader' to identify the emotions of his students. A Chinese university lecturer is using facial-recognition technology on his students to check if they're bored. The reader produces a'curve' for each student, showing whether they are happy or not, and giving indications of whether they are bored.

Companies use OKCupid photos, social media to train face recognition

Daily Mail - Science & tech

With images aggregated from social media platforms, dating sites, or even CCTV footage of a trip to the local coffee shop, companies could be using your face to train a sophisticated facial recognition software. As reported by the New York Times, among the sometimes massive data sets that researchers use to teach artificially intelligent software to recognize faces is a database collected by Stanford researchers called Brainwash. More than 10,000 images of customers at a cafe in San Francisco were collected in 2014 without their knowledge. OKCupid and photo-sharing platforms like Flickr are among for researchers looking to load their databases up with images that help train facial recognition software. That same database was then made available to other academics, including some in China at the National University of Defense Technology.

Supporting Feedback and Assessment of Digital Ink Answers to In-Class Exercises

AAAI Conferences

Effective teaching involves treating the presentation of new material and the assessment of students' mastery of this material as part of a seamless and continuous feedback cycle. We have developed a computer system, called Classroom Learning Partner (CLP), that supports this methodology, and we have used it in teaching an introductory computer science course at MIT over the past year. Through evaluation of controlled classroom experiments, we have demonstrated that this approach reaches students who would have otherwise been left behind, and that it leads to greater attentiveness in class, greater student satisfaction, and better interactions between the instructor and student. The current CLP system consists of a network of Tablet PCs, and software for posing questions to students, interpreting their handwritten answers, and aggregating those answers into equivalence classes, each of which represents a particular level of understanding or misconception of the material. The current system supports a useful set of recognizers for specific types of answers, and employs AI techniques in the knowledge representation and reasoning necessary to support interpretation and aggregation of digital ink answers.