1. Q & A robot:
Information retrieval, search
Keywords: TFIDF, SVM, Simple Bayes, RNN, CNN
Knowledge map
Store knowledge and knowledge between knowledge and knowledge in the graphical database, convert the question and answer into query statements, and can realize reasoning
Common implementation ideas
Voice turn text
Intent identification, domain identification, text classification
Slots fill: For example, the robot to buy tickets uses the command to identify the padding from {location} to {location} 2 positions.
Reply management, reply policy
Natural language generation
Text rotation

2, visual and image processing

Image Processing
Image processing is intended to process the original image to apply some kind of transformation. Its goals are typically improving images or inputs them as an input to a particular task, and the target of computer vision is to describe and interpret images. For example, noise reduction, contrast, or rotation operation These typical image processing components can be performed on the pixel level.
Machine visual
It is mainly to refer to visual research in the industrial sector, such as the visual of autonomous robots, for detection and measurement of visual. This shows that in this field through software hardware, image perception and control theory tend to achieve high efficiency robot control or various real-time operations with image processing.