2018-03-20

“Alphago prevail over human Go champion.” ”Self-driving cars will replace drivers.”
Artificial Intelligence is sweeping over the globe! To embrace AI, Big data and the era of IOT, NTHU is not only the investing target of MOST, but with the entering of AI system center and the excellent soft/hardware environment of NTHU EE.

Our professors run many sorts of AI projects and offer a great teaching quality. All of these efforts are aiming for providing the best learning environment for creating future technology. Encouraging our students to explore, to experiment, to learn and to study. There is no limit of the depth and the width of knowledge. Crossing among medicine, education, psychology, humanities, mechanical and information. The following is the introduction of the AI project of our department:

Machine can learn to identify emotions - Integrated enhancing model of cross-language emotions

Nowadays the development and mastering of machine learning and AI is improving. Hence, the realizing of smart home is becoming possible. Emotions, especially for the expression of voice, are the critical point for the machines to truly understand human beings. Prof. Chi-Chun Lee’s AI study is mainly researching on the emotion models of multi-language and further integrating as a completed and precisely prediction system for emotion recognition.

Recognizing Autism with AI- using facial movement units to construct multi-modes recognition system.

Furthermore, in the past, clinical diagnosis of Autism was basically relying on the short period of interaction between the doctor and the children. During the interaction, the children’s different behaviors were observed and documented according to ADOS (Autism Diagnostic Observation Schedule). This process is not only time consuming, but also containing other uncertain factors such as subjectivity.

In order to solve the drawbacks of the lack of objectivity, AI experiments is a best choice since computer is a pure third-party observer. We catch the dynamic facial expression during human interactions and coding them as action units which were further used for machine modeling and recognizing. In the case of Autism children, by using their facial expression, we could analyze whether do they have the tendency to Autism.

Fast recognition of Emergency triage to save the golden treatment time

On the other hands, machine learning could be applied on the fast recognition of emergency triage which could speed up the emergency procedure, saving the golden treatment time. Emergency triage could be achieved by using machine learning to catch patients’ facial expression in the emergency room and recognize the facial action units and further tell the pain score.