Peer Review

How We Teach Computers to Understand Pictures

Material Detail page

Bookmark

Overall Rating:

Content Quality:

Effectiveness:

Ease of Use:

Overview:: How We Teach Computers to Understand Pictures
Type of Material:: Reference Material
Recommended Uses:: This video can be used in-class, individual watching or team discussions. It can be used as part of a lecture or self-paced learning.
Technical Requirements:: Browser with the right plug in to show YouTube video
Identify Major Learning Goals:: - It introduces the basic idea on a computer content understanding on pictures - The youtube video recorded a TED presentation with the aid of illustrations in the background. - As described in the Youtube page, a computer vision expert described the state of the art — including the database of 15 million photos her team built to teach a computer to understand pictures.
Target Student Population:: - The presentation is open public lecturer that was designed for general public. - It covers areas of computer machine learning and artificial intelligence. - It also helps in understanding of computer graphics. - Undergraduate computer science course, or general education course targeting students who are interested in computer science
Prerequisite Knowledge or Skills:: - For the general understanding of the content, no particular background knowledge is needed. - Some background concept of neural networks can further help users in better understanding. - Students or the audience needs to have a basic understanding of computer science terms, algorithm, image modelling, big data etc.

Rating:
Strengths:: TED talks tend to be interesting and engaging. This video started by presenting pictures from real life, drawing analogy to how children learn to see pictures, and ended with how computers can be trained to see pictures. It also gave a detailed description of imagenet.org, "an image database organized according to the WordNet hierarchy (currently only the nouns), in which each node of the hierarchy is depicted by hundreds and thousands of images." The lecture uses many illustrative examples to explain how convolution neural network works on the content understanding. The theory of machine learning used in information extraction from a picture is briefly but clearly presented. Very sufficient set of data for supporting the idea being presented.
Concerns:: The video is quite long and mostly introduces facts and efforts the researchers undertook to train computers to see images. It does not show the technical details, the algorithm or the modelling.

Rating:
Strengths:: - While the presentation is a public lecture requiring no prerequisite knowledge on computing for the general understanding, it would be more interesting for user with some background of the machine learning. - It also updates viewers on the achievements and the limitations of the current computer algorithms and technology in understanding pictures. - It is natural that it does not have any assessment to be given the users, but the lecture inspires viewers to reflect on the impact of neural network in computer picture understanding. - Video is one of the best multimedia one can use in teaching. This video can be flexibly used in various learning settings, individual learning or in-class discussions. It could generate good topics for discussion. The story-telling is quite effective and this video can also be used to teach entrepreneurship or in entrepreneurship education.
Concerns:: This video is informational and non-instructional.

Rating:
Strengths:: - The presentation is very interesting and the viewers are not overwhelmed by the complicated theories or algorithms behind. - The purpose of lecture is conveyed effectively. - Also, the content is easy to understanding and visually appealing. - The video provides very good switching between the captions of the presenter and aiding the slides or pictures. - Quality of this Video is excellent, consistent with TED video qualities, including pictures, sound, and the switch between scenes. - The information is presented at the right pace and in an engaging way.
Concerns:: The video is lengthy. It focuses more on conveying information than teaching new knowledge. The beginning part seems like a publicity for imagenet.