Solving the Natural Language Understanding (AI) problem in Ed-tech
Details
If you are in an ed-tech startup in Sydney, please consider coming to this meetup, which will sell out quickly. It is going to be "a cracker".
On 5 October the Sydney Machine Learning Meetup Group will host a talk by Beth Carey (https://www.facebook.com/Beth-Carey-359793174167713/?ref=gs&fref=gs&hc_location=group) and John Ball of PAT Inc., a Sydney, Bay Area and RoP based startup of Australian background.
PAT Inc. has made substantial progress in solving the "natural understanding problem" in AI. They have recently published a substantial technical paper suggesting their API can pass serious independent tests of language understanding, such as the Facebook bAbi tests.
Practically speaking, their technology may lead to voice interfaces for consumer electronics actually being able to handle context and thus problems such as ambiguity etc.
Two of the early financial backers of PAT Inc. have been Chinese and Australian education companies, who can see the potential of this technology in education, particularly ESL education in the first instance. Personally, IMO, a large proportion of ed-tech companies need to look at tools such as the PAT Inc. API.
As you probably know, machine learning tools have leapt ahead in recent times in the area of "visual intelligence", as represented in iPhoneX and Google Photos.
Less progress has been made in language understanding. Basically, the error rates for actions based upon "free form" verbal or written instructions, are still unacceptably high (even if deep neural network based voice to text recognition is good). There is considerable controversy as to whether deep neural networks will ever get there.
The meeting will be at Haymarket HQ (https://www.facebook.com/haymarkethq/?ref=gs&fref=gs&hc_location=group), and is supported by the Shenzhen Economic and Trade Office. It will sell out quickly, given that normally around 200 folks turn up to this meetup group.
Please sign up at the below Eventbrite group:
Solving the Natural Language Understanding (AI) problem in Ed-tech