What are you working on?

jsalvatier

Whpearson recently mentioned that people in some other online communities frequently ask "what are you working on?". I personally love asking and answering this question. I made sure to ask it at the Seattle meetup. However, I don't often see it asked here in the comments, so I will ask it:

What are you working on?

Here are some guidelines

Focus on projects that you have recently made progress on, not projects that you're thinking about doing but haven't started, those are for a different thread.
Why this project and not others? Mention reasons why you're doing the project and/or why others should contribute to your project (if applicable).
Talk about your goals for the project.
Any kind of project is fair game: personal improvement, research project, art project, whatever.
Link to your work if it's linkable

What are you working on?

Here are some guidelines

Focus on projects that you have recently made progress on, not projects that you're thinking about doing but haven't started, those are for a different thread.
Why this project and not others? Mention reasons why you're doing the project and/or why others should contribute to your project (if applicable).
Talk about your goals for the project.
Any kind of project is fair game: personal improvement, research project, art project, whatever.
Link to your work if it's linkable

I'd be interested in your work on recommendation systems. How well does it deal with semi-intelligent spammers? That is spammers that copy other normal people's ratings for the most part but then alter there behaviour to prefer something more than it is worth.

Personally I think a good recommendation system is something that can have a vast impact on society. Mainly through knock-on effects; if you save charity X time and money on deciding which programmer to employ (due to a recommendation system) they can then spend that money on actually helping people.

I'm interested in what you think is more important!

How well does it deal with semi-intelligent spammers?

The most obvious thing distinguishing my work from previous attempts is that it attains all its guarantees even if 99% of all users are arbitrarily sophisticated adversaries. The amount of time depends on the logarithm of the fraction of honest users. So if only one user in 10^10 is honest, it will take about 30 times longer to converge to good recommendations.

The goal is to find the best fixed moderation strategy (ie, block all messages from people X, Y, Z...) for the honest users. Here the definiti... (read more)