3 Followers
8 Following
31 Posts
I am a director of a group working on AI/ML/robotics projects at a small company. We're especially focused on collaborative decision-making for autonomous vehicles.
Never before has the 'understanding' piece for #AI been more important “It is that “decoding” process that is the ‘U’ in NLU — that is, understanding the thought behind the linguistic utterance is exactly what happens in the decoding process” https://buff.ly/3s4jc29 #voicefirst
Machine Learning Won't Solve Natural Language Understanding

An argument that it is time to re-think our approach to natural language understanding, since the ‘big data’ approach to NLU is implausible and both technically and theoretically flawed..

The Gradient

@Ryder So I think talking about that end of it, it's basically like Dennett's "intentional stance" - For humans, at some point a system becomes complex enough that it's easier (and effective) to explain it using theory of mind principles.

But that's separate from the machine itself having enough meta-level awareness of its own functionality to 'transfer' that for use as the model of how a human is going to react, which is what I feel would be "theory of mind" in the machine.

@Ryder Ohh, are you talking about the machines triggering the humans' theory of mind?
@Ryder Does #TheoryOfMind in its typical definition really include predictive behavior models and the implicit models that you're talking about? I was always under the impression that #TheoryOfMind was more about entities capable of reasoning that another entity was "like me, but with different beliefs/desires/intents". If these models don't have the capacity to reason about their own nature, do they really have a sufficient theory of self to use for inverting others' mental states? #discussion
@Ryder Previous classes of WMDs leave signs that can be detected by inspectors... So it's possible to enforce agreements. But can AI-driven weapons be detected? Or could the work being done to build them basically hide in plain sight in any civilian context? And if nations agree globally to a ban but terrorist groups start doing their own work along those lines, will the nations stay in line?

@Ryder It wouldn't hurt, but isn't that Pandora's Box already open at this point? I think at least as helpful would be solutions that mitigate the geopolitical game-theoretic configurations that drive this arms race. You can tell countries to stop what they're doing, but 1. It's a Prisoner's Dilemma, so everyone wants to talk peace and then cheat; and 2. It's not just nation-states - this stuff is accessible enough that random groups can do it. Any ideas? What do people think?

#AI #discussion

@solalnathan Ah lol, I actually wondered if it was the French spelling, thanks.
@solalnathan Just curious, what does "IA" stand for in this context? It's coming up in sporadic hashtags...
@andrey @thegradient Hey #SigmoidSocial team, maybe consider adding this in the server announcements, or the site-level blurb? 'In the web client you can click the control panel icon in the upper right over a hashtag column and "pin" it. If you click that icon again over a pinned column, you can choose "local only".' This seems like it could be a good way for people to get engaged in #discussion and #question answering relevant to this instance, if you could advise them as such. #mastodon
@andrey I don't know if you figured this out already, but in the web client you can click the control panel icon in the upper right over a hashtag column and "pin" it. If you click that icon again over a pinned column, you can choose "local only". This seems like it could be a good way to get discussions and questions going on just this server, if people would start doing this. #mastodon #question #discussion #SigmoidSocial