[ml] Project idea: generate a list of items given examples

Wladyslaw Zbikowski embeddedlinuxguy at gmail.com
Wed Aug 31 13:01:05 PDT 2011


As you may know, Google has decided to shut down its Labs program,
including the very interesting Google Sets - a "mechanism for quickly
and efficiently generating lists of items given one or more example".
In other words if you type "brian eno" and "mike oldfield" it suggests
"kraftwerk", "philip glass", etc. Some terms work better than others!

I'm not aware of any other generalized large-dataset suggestion
software out there, so I thought it would be fun to start a Google
Sets replacement as an open source project. I'm pretty comfortable
with Linux system software, SQL, and web application programming, but
I could definitely use help in:

- scalable database & computation backend (Hadoop, LevelDB, etc)
- the actual pattern-matching inference magic (hence my post to the ML list)

Any alternative ideas or interest in collaborating on something like this?

http://labs.google.com/sets
http://modernhacker.blogspot.com/2011/08/google-sets-experiment-cities-1337-l4m3.html


More information about the ml mailing list