Alchemy – Open Source AI | Hacker News Reader

Alchemy – Open Source AI

alchemy.cs.washington.edu

143 pointschewxy22 comments

Shameless self-promotion post: if you're interested in AI based on Markov, I maintain an open-source library here https://github.com/Svalorzen/AI-Toolbox. It can be used for general autonomous agents which need to act in response to environmental stimuli. I've used it to learn to play StarCraft and to control cameras to predict people's movements around buildings. Any comments and suggestions are welcome =)

> I've used it to learn to play StarCraft and to control cameras to predict people's movements around buildings.

If you've described / published either of these examples somewhere, you should submit them to HN!

I actually have a video [0] and code [1] of the first (which was a one-month project), while the second is my thesis which is going to be finished soon. Maybe I'll post them as a separate entry one day, but I always think these things are too simple and trivial to be interesting..

[0]: https://www.youtube.com/watch?v=TrKMBIR82Qw

[1]: https://github.com/Svalorzen/DESOLATOR

Loved watching the Dragoons dance around their opponents while taking almost no damage. They seemed to have slipped a few times where they could've doubled up on fire but just one was firing instead. I might have looked at it wrong though.

Thank you for this, I will look into it :). I was actually going to tell these people to give me a github link or go away, haha. I don't need any kind of license page that looks like http://alchemy.cs.washington.edu/license.php?action=2, thank you very much!

By the way, you can talk to my AI guy if you have a recent Chrome/Chromium browser. It is actually an entire OS in a browser, and the AI guy is an app called Bertie. My current implementation of the OS is called "The Native Client Proving Ground", and can be found here: https://nacl-pg.appspot.com/desk. The OS should install in less than 2 seconds after clicking the link. Then you just open the Applications folder and click on Bertie's face. He will then introduce himself, and you can start talking to him.

Ah, I was confused for a second—I'd thought Markov was a library, but you meant the markov assumption—the topics are actually loosely related and orthogonal. Your excellent looking library deals with reinforcement learning agents that model environment/agent interactions as a (PO)Markov Decision Process, where as the Alchemy library combines FOL with network representations of particular (satisfying certain markov properties) probability distributions to perform inference.

More pertinent to your post, Sutton's working on an updated RL book here: http://people.inf.elte.hu/lorincz/Files/RL_2006/SuttonBook.p...

If anyou have the time, Chapter 15 (pdf pg 273) of the above link is a fascinating read. In particular, TD-Gammon had already achieved impressive results using NNs in the early 90s; reaching world class levels in Backgammon with zero specialized knowledge.

Shameless naive question: are these techniques suitable for time series models?

You mean data-fitting? Possibly (every problem can be modeled as an MDP) but it's definitely not the most efficient nor the most intuitive way to go about it, so I wouldn't recommend it.

The NLP group at Washington is one of the best in the world (imo likely top 5), and did most of the seminal work on open information extraction, the technologies backing this service.

I'm surprised they chose this name. The commercial service, AlchemyAPI, that was bought by IBM is likely to be much less accurate. I can't say for sure, though --- their terms of service prohibit evaluation. All I know is, that's not a good sign...

Alchemy was launched much ealier than AlchemyAPI.

Very interesting!! I had no idea.

Washington's Open Information Extraction project - Ollie[1] - is really, really good. For example I've found it much more effective that the OpenIE tool in Stanford CoreNLP.

I think Alchemy is pretty old though? Most people seem to use Tuffy for MLN now.

[1] https://github.com/knowitall/ollie

Tuffy is used in DeepDive[1], which is a very interesting project.

[1] http://deepdive.stanford.edu/doc/advanced/markov_logic_netwo...

Weird. I submitted this yesterday. Must have been the capitalization that got me: https://news.ycombinator.com/item?id=10630551

Anyway, just read "The Master Algorithm" by Pedro Domingoes. Fantastic read. The most interesting part to me was his survey of the five tribes of machine learning: symbolists, connectionists, geneticists, Bayesians and analogists.

After the survey he goes on to talk about some other aspects. And then discusses Alchemy and the possibility of uniting the techniques of all five tribes into one algorithm (hence the title).

I found his writing on a dense subject easy to read and great at conveying the concepts. Well worth checking out.

Edited for typos

I couldn't get past the introduction of "The Master Algorithm", is the rest of the book a bit more grounded?

The introduction is a bit breathless, but most of the rest of the book is more down to earth.

Related? http://www.alchemyapi.com/

no, the one you linked is an analysis API -- now owned by IBM

Yeah slightly confusing naming because Alchemy APIs (owned by IBM now) is promoted alongside the Watson APIs which are in the same general space as this project..

The project looks dead. Last updates in 2013?

This doesn't seem so uncommon for big projects at that department. I'm thinking for example about all the various tools developed by Noah Snavely and colleagues for the Photo Tourism project http://phototour.cs.washington.edu/ , that were later acquired by Microsoft for their Photosynth branded tools.

Beautiful plumage, sir.

Reply on news.ycombinator.com