Now, MIT professor Picard has made many achievements. After reading the article, I think this field sounds beautiful but far away. Has the industry started to pay attention to and implement any products? Will the field of affective computing combined with cognitive science become the next hot spot of artificial intelligence?
Is affective computing useful? The answer is yes. At present, there are many application scenarios of emotion computing, such as promoting work efficiency, game feedback and fatigue driving. A positive emotion can strongly promote these scenarios. If a person is in a negative mood for a long time, it is obvious that it is easy to lead to psychological diseases and psychological problems; On the other hand, the performance of emotional computing will also affect the research of human-computer interaction and brain computer interaction, which can promote each other.
The expression of affective computing. Common indicators include facial expression, sound signal, text information, ECG, EEG and some physiological indicators such as skin impedance. In fact, these indicators have their own advantages and disadvantages. Combined with machine learning methods, they all have a certain degree of identification. In recent years, with the development of wearable devices and dry electrode technology, in fact, I think it is like heart computer electrical and physiological indicators.
Where are the difficulties? Let me take EEG as an example. Different from facial expressions, it can not be disguised or fake, but it has low signal-to-noise ratio and is easy to be affected by EMG. In the research of pattern recognition, many extracted features have good results. The most commonly mentioned is differential entropy, but it has not reached the accuracy of practical application. In the past two years, some in-depth learning methods have achieved the most advanced results of many emotional data sets, and there is still much room for progress.
Will AC become a hot spot in artificial intelligence? Personally, I think this field is very interesting and still worth exploring, but it has not reached the heat and research value in the fields of automatic driving and natural language processing.
Product landing. When it comes to the fastest possible applications, I am still optimistic about some emotion recognition in game feedback, which is the easiest to land in combination with wearable devices. To say that automatic driving and other safety related issues, we have to consider time.