Ai Dreams Forum

Artificial Intelligence => AI News => Topic started by: infurl on May 30, 2020, 07:08:23 am

Title: Measuring the hype.
Post by: infurl on May 30, 2020, 07:08:23 am

https://tech.slashdot.org/story/20/05/29/2048201/eye-catching-advances-in-some-ai-fields-are-not-real (https://tech.slashdot.org/story/20/05/29/2048201/eye-catching-advances-in-some-ai-fields-are-not-real)

It should be obvious to most of us that there is a lot of baseless hype surrounding artificial intelligence and the past decade has been no exception. Lately it has been starting to look like the most expensive snake oil in history and I fear that we will be seeing another AI winter soon. All the scavengers who have been flocking to machine learning in order to get rich quick will have to move on and good riddance to them.

Title: Re: Measuring the hype.
Post by: LOCKSUIT on May 30, 2020, 02:04:35 pm

I've seen such papers myself. Most say they get superior results.

But without actually testing the algorithm's usefulness or prediction accuracy, you'll never know (if they haven't built up trust).

Also, papers need to explain the idea fully and clearer. If you can't understand the idea, it is "able" to be a lie.