Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Pac-Man recreated with a GAN trained on 50k game episodes

blogs.nvidia.com

Post reply on HN
Print view View original

121–130 of 154 posts

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#121

Post by chx »

lumost wrote: post #81
lostmsu wrote: post #76

Earlier quoted context omitted.

A) is just a sign, that further engineering, both model and framework, is required. Also, 50-60 words per second does not sound too bad. It amounts to 4M words per day. I imagine most people only type a thousand words per day. So with a single GPU you can handle 4000 users (assuming batch-style processing). B) This specific example actually shows ML toolchains are orders of magnitude better, than existing ones. Imagi…

On A: Using a p3.2xl for that task leaves you with a minimum ec2 cost of goods sold of 50 cents per user per month for the target task. To make a reasonable SaaS business out of this you would need to be charging a minimum of $5.00 per user month for the service that this GPU is providing, assuming all other tasks have a negligible impact on Cost of Goods Sold, and that you are able to spread the load optimally throu…

> On A: Using a p3.2xl for that task leaves you with a minimum ec2 cost of goods sold of 50 cents per user per month for the target task

https://twitter.com/br_/status/979442438254166016

> "selling AWS at a loss" is crisp shorthand for a lot of startups' business models!

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#123

Post by dougabug »

ookblah wrote: post #122

hmm, train a bunch of UI elements and interactions and feed it underlying codebases and see what comes back... lol.

That’s actually been done at least a couple years back.

https://towardsdatascience.com/code2pix-deep-learning-compil...

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#124

Post by YeGoblynQueenne »

esoterica wrote: post #105
YeGoblynQueenne wrote: post #93

Earlier quoted context omitted.

That is not true for academic papers in computer science and particularly in machine learning were contributions are expected to have some sort of practical application that is at least possible to foresee. Indeed, a common criticism against papers in machine learning is "what is a real-world application of this approach?". If you can't answer that convincingly then it's unlikely your paper gets published, particular…

> That is not true for academic papers in computer science and particularly in machine learning were contributions are expected to have some sort of practical application that is at least possible to foresee. This is 100% not true for computer science and I don't know why you think it's true. There's are entire branches of computer science (e.g. complexity theory) that are highly theoretical and are certainly not gea…

Well of course, it's not 100% true- who said anything about 100%? There are theoretical branches of computer science; but most of computer science is not purely theoretical. Certainly machine learning (and more broadly speaking, much of AI) is like I say.

I'm not sure who brought up a schematic? The bit I quoted is taken from the paper and it's a clear attempt to motivate the paper's approach as having a practical application. To clarify, do you disagree with that?

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#125

Post by kevincox »

fpgaminer wrote: post #96
lumost wrote: post #70

I've observed an increasing trend with modern AI research that A) It's no longer practical to incorporate into real products given the hardware costs for training and inference. In NLP, large transformer based models can easily occupy a 5 thousand dollar GPU at a paltry 50-60 words per second processing pace. B) The Research often demonstrates things that existing non-ML toolchains were already reasonable at doing wi…

> It's no longer practical to incorporate into real products given the hardware costs for training and inference I don't see that trend. Bleeding edge research has always been beyond what's practical, by its very nature. The focus is on pushing the boundaries, not to build something efficient/commercial. It's the followup work that makes it practical. See for example the evolution of WaveNet. It was breakthrough rese…

Exactly. It is sort of like saying the first transistor was too expensive to be practical.

Development always starts with some novel technology that has major flaws, but some good nuggets. Then it is further developed to be practical.

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#126

Post by 101404 »

plutonorm wrote: post #94
101404 wrote: post #92

Earlier quoted context omitted.

Not at all. It just means some people poked a pile of a bazillion if-else statements long enough, until the desired result came out. I believe everyone here knows the xkcd I am referring to. Still, I think it's an interesting result. But, as someone on Fridman's podcast said the other day: recent advances in AI have all been thanks to advances in computing power, not any new ideas. The future if AI seems to depend on…

What makes you think the pile of if-else statements isn't conscious?

Because it's a pile of if-else statements shuffled until it was able to perform a single specialized thing. Any housefly would by a thousand times as "conscious".

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#127

Post by anigbrowl »

lumost wrote: post #70

I've observed an increasing trend with modern AI research that A) It's no longer practical to incorporate into real products given the hardware costs for training and inference. In NLP, large transformer based models can easily occupy a 5 thousand dollar GPU at a paltry 50-60 words per second processing pace. B) The Research often demonstrates things that existing non-ML toolchains were already reasonable at doing wi…

Your comment reminds me of snide dismissals of personal computing in the 1980s - looks worse than [existing thing], very expensive, requires an army of technicians, probably explodes unpredictably, overall just a terrible investment.

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#128

Post by bobly_today »

101404 wrote: post #126
plutonorm wrote: post #94

Earlier quoted context omitted.

What makes you think the pile of if-else statements isn't conscious?

Because it's a pile of if-else statements shuffled until it was able to perform a single specialized thing. Any housefly would by a thousand times as "conscious".

What makes you so certain a house fly doesn’t have a form of rudimentary consciousness? It might be a spectrum.

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#129

Post by Ajedi32 »

Teknoman117 wrote: post #74
ngneer wrote: post #69

Please help me understand the point. Is it to show that nVidia chips can be used to create game AI using GANs? Or is it to aid in game design? Wouldn't a human be in a better position to do both?

If I'm understanding correctly, the machine learning model was able to create a game executable from information about a game.

From the article, it sounds to me more like the AI itself is rendering individual frames of the game. It takes the last frame and the current state of the gamepad as input, and outputs the next frame. They've effectively replaced the entire game code with a neural network.

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#130

Post by contravariant »

lonelappde wrote: post #35
coldpie wrote: post #10

Earlier quoted context omitted.

Yes, absolutely. You can catch a baseball because your brain predicts where it will go before your eyes see its actual movement. This may be a good place to start learning more: https://en.wikipedia.org/wiki/Direct_and_indirect_realism

How is that "not a physics engine"?

It's ridiculously bad at modelling quickly rotating objects.
Post reply on HN