Ask HN: What's a useful application of AI and/or ML?
4 comments
I think this is a good question to ask. But it might be better put as just "how far have applications of AI really gone".
That said, self-operating vacuum cleaners, lawn-mowers and cars would be obvious answers. Automating X is a fairly easy thing for people to understand. Whether that's where the important applications of AI are going to come from is a different question.
Just as much, do things like walking robots require deep learning or do better with deep learning rather than just regular engineering is a further question. A claim I read here is that Boston Dynamics, of Big Dog fame, was sold by Google because weren't able to convert their walking robots to working primarily via neural networks and I don't know whether a Rhoomba uses "serious" ML or not - they have been around for a while.
That said, self-operating vacuum cleaners, lawn-mowers and cars would be obvious answers. Automating X is a fairly easy thing for people to understand. Whether that's where the important applications of AI are going to come from is a different question.
Just as much, do things like walking robots require deep learning or do better with deep learning rather than just regular engineering is a further question. A claim I read here is that Boston Dynamics, of Big Dog fame, was sold by Google because weren't able to convert their walking robots to working primarily via neural networks and I don't know whether a Rhoomba uses "serious" ML or not - they have been around for a while.
Generating new scientific hypotheses and helping guide scientific discovery. In this regard AI / ML can help us:
develop new materials (stronger lighter materials for airplanes, so the need less fuel)
develop better superconductors (make t3h electric cheaper)
create more efficient solar cells (make t3h electric cheaper)
develop better techniques for desalinizing seawater (make freshwater available in areas where supplies are limited)
discover new drugs (cure t3h disease)
etc., etc., etc.
develop new materials (stronger lighter materials for airplanes, so the need less fuel)
develop better superconductors (make t3h electric cheaper)
create more efficient solar cells (make t3h electric cheaper)
develop better techniques for desalinizing seawater (make freshwater available in areas where supplies are limited)
discover new drugs (cure t3h disease)
etc., etc., etc.
None of the things you mentioned exist yet, though.
AI is absolutely already being applied to scientific discovery. IBM even have an entire product "Waton for Drug Discovery". See also;
https://www.ibm.com/blogs/research/2016/06/artificial-intell...
https://medium.com/@ShaliniAnanda1/ai-and-the-future-of-comp...
http://www.nature.com/news/can-artificial-intelligence-creat...
I can't cite specific instances where it has yielded specific results from the top of my head, but there's no question that AI is already being used in one or more of these areas.
https://www.ibm.com/blogs/research/2016/06/artificial-intell...
https://medium.com/@ShaliniAnanda1/ai-and-the-future-of-comp...
http://www.nature.com/news/can-artificial-intelligence-creat...
I can't cite specific instances where it has yielded specific results from the top of my head, but there's no question that AI is already being used in one or more of these areas.
You encounter a random person on the street and they ask what it's good for. They say you can say anything as long as it is not entertainment related and benefits them directly in some way.
The purpose of the above situation is to show that application cannot be specific to a given socioeconomic status, career, job, gender, etc. It must be something that's useful to more or less anyone in the planet, even if they don't have internet.