This website uses cookies and similar technologies to understand visitors' experiences. By continuing to use this website, you accept our use of cookies and similar technologies,Terms of Use, and Privacy Policy.

Jan 17 2018 - 09:41 PM
Google Cloud AutoML: AI may not be that far as it seems to education

Today, Dr. Fei-Fei Li, Chief Scientist of Google Cloud AI, blogged about the soon-to-be-released Google Cloud AutoML. Cloud AutoML is the self-serving version of the Google Cloud Machine Learning Engine released in 2017. The difference between the Cloud Machine Learning Engine and Cloud AutoML is the latter has a much lower entry barrier to startups and small businesses who are not able to hire machine learning experts to build ML models and finetune the model parameters. Cloud AutoML is capable of improving its own parameters based on the data it receives such as user-generated inputs and tagging. It uses AI to improve AI.

So what that means for education? It means many innovation organizations and individuals in the education sector such as EdLab can use Cloud AutoML in the near future to improve their products and services with AI without spending hundreds of thousands of dollar each year to hire machine learning experts. For example, Vialogues can use Cloud AutoML Vision, the first service of Cloud AutoML to be released, to achieve image recognition of the videos. That makes it possible to judge whether the video discussions are relevant to the video. In addition, Google will also release Cloud AutoML NLP (Natural Language Processing). That can potentially let us implement advanced recommendations in Vialogues to suggest discussions that are "inherently" (as opposed to "superficially") related to each other. Also, if we let users vote for the "productivity" of discussions, Cloud AutoML NLP can automatically summarize the inherent patterns of productive and unproductive discussions. This is where the interesting things begin to happen...

Posted in: Technology|By: Zhou Zhou|1278 Reads