News for hackers. Stuff that matters.
Re:Zoom lied to users about end-to-end encryption for years,... by Jasper_ on Tuesday November 10, 2020 @10:55AM (#25047351)
Given how most actual AI solutions work under the hood, this might not even be a lie!
Re:Zoom lied to users about end-to-end encryption for years,... by TeMPOraL on Tuesday November 10, 2020 @11:12AM (#25047568)
That many (most?) companies in this space lie through their teeth doesn't excuse the lie.
Re:Zoom lied to users about end-to-end encryption for years,... by wil421 on Tuesday November 10, 2020 @10:58AM (#25047380)
Our system has an AI module.

AI module:

If something

Else if something else

Else if

Else if

...

Else Call Human

Re:Zoom lied to users about end-to-end encryption for years,... by Jasper_ on Tuesday November 10, 2020 @11:14AM (#25047603)
To be honest, before the modern machine learning approach, this was known as a decision tree and was thought to be a valid way to approach "artificial intelligence". Lots of "AI" hype in the 80s was based around "Expert systems" and "Decision trees".
Re:Zoom lied to users about end-to-end encryption for years,... by zelphirkalt on Tuesday November 10, 2020 @11:27AM (#25047738)
And there are even modern tree based approaches, that beat some of the modern artificial neural network approaches! It's not like it has become an absolutely unusable class of algorithms.
Re:Zoom lied to users about end-to-end encryption for years,... by FridgeSeal on Tuesday November 10, 2020 @07:04PM (#25053858)
People seem to think that the presence of neural nets and deep learning means that most of types of models are practically superseded whereas in my experience if some non-deep-learning model even gets you mostly there, then the efficiency and explain-ability wins make it worth it.

At my previous job I had a ML-based service that used a basic random-forest model instead of a neural net because it was faster to train and operate, not to mention easier to maintain and had equivalent accuracy with little to no effort required on my part. It was a solid little service.

Re:Zoom lied to users about end-to-end encryption for years,... by SAI_Peregrinus on Wednesday November 11, 2020 @03:47PM (#25062782)
And you can even do both. Decision tree to get to a smaller problem space, then NN on the output. You end up with a bunch of neural nets, each of which performs better than a single monolithic net due to solving a simpler problem.
Re:Zoom lied to users about end-to-end encryption for years,... by FridgeSeal on Wednesday November 11, 2020 @06:12PM (#25064145)
Yeah! Stacking models and playing on each types strength is a great strategy.
Re:Zoom lied to users about end-to-end encryption for years,... by p1necone on Tuesday November 10, 2020 @06:41PM (#25053634)
Decision trees/behaviour trees are still the most widely used way to build video game AI. Of course video game AI just has to appear smart, really you're solving a different problem a lot of the time.
Re:Zoom lied to users about end-to-end encryption for years,... by TeMPOraL on Tuesday November 10, 2020 @11:14AM (#25047598)
We practice responsible and safe approach to AI, by detecting situations the AI can't handle and deferring to human response.

(It just so happens that the set of things the AI can handle is empty.)

Re:Zoom lied to users about end-to-end encryption for years,... by WalterBright on Tuesday November 10, 2020 @12:50PM (#25048761)
> Given how most actual AI solutions work under the hood, this might not even be a lie!

Even back in the 80's, the computer algorithms that played the other side in computer games was called "the AI".