36 ms·
>> Our field isn't quite "artificial intelligence" True, but so what? We call it AI and that's that, really. We've been calling it that for 70 years now and it
by YeGoblynQueenne 7y ago
>> Our field isn't quite "artificial intelligence"
True, but so what? We call it AI and that's that, really. We've been calling it that for 70 years now and it's never been a problem.
And let's be absolutely clear that it's not the _name_ that's confusing the public but the way that industry luminaries promise autonomous cars and robotic maids in the next -3 years, or the way that the technology press -the technology press- can't get its shit together to figure out the difference between "machine learning", "deep learning" and "AI" as fields of research and as category labels. Of course the lay public is going to be confused if people who are paid to elucidate complex concepts make a mess of it.
- joe_the_user 7y ago> "True, but so what? We call it AI and that's that, really. We've been calling it that for 70 years now and it's never been a problem." That isn't even ... true. AI became "machine learning" in the late 90s/early 2000s and that change happened because the chorus of criticism of "artificial intelligence" had become extremely loud and a less ambitious term served as a refuge.
- YeGoblynQueenne 7y agoAI was renamed into many things in the '80s and '90s, for example "Intelligent Systems" or "Adaptive Systems" etc, and that indeed was done to dissociate research from the bad rep that had accrued for AI. But "machine learning" has been the name of a sub-field of AI since the 1950's and it's never stood for the whole, at least not in conferences, papers or any kind of activity of the field. For example- two of the (still) major conferences in the field are AAAI and IJCAI: the conference of the "Association for the Advancement of Artificial Intelligence" and the "International Joint Conferences of Artificial Intelligence". Neither of those is in any way, shape or form a conference for machine learning only and neither uses machine learning a byname for AI. By contrast, machine learning has its own journal(s actually) and there are specific conferences dedicated to machine learning and deep learning (NeurIPs and ICLR). Additionally, there are many sub-fields of AI that are not machine learning, in name or function: intelligent agents, classical planning, reasoning, knowledge engineering etc etc. The only confusion between "AI" and "machine learning" exists in the minds of tech journalists and the people who get their AI news exclusively from the tech press. P.S. As a side note, the name for what the tech press is doing, referring to the field of AI as "machine learning", is "synecdoche": naming the whole by the name of the part.
- sgt101 7y agoNo. Some people started saying things like that, more in about 2013, but all the time many people have been working on topics like MAS, answer sets, causal logic and other stuff. At that time the big trend was actually rebranding maimed logical inference as The Semantic Web.
- 6gvONxR4sf7o 7y agoProbably most of the things being marketed as AI have been around and called statistics for a long time, and it's never been a problem.