Open Mind Common Sense

New Blog Site!

2010-05-12T15:10:00.000-07:00

If you've wondered why we haven't seemed active lately, it's because we moved the blog over to csc.media.mit.edu and forgot to tell everybody! :) So if you're still reading this blog, come check out the new one instead!

(We may one day figure out how to syndicate posts back here, but for now, this blog is archived.)

New SQLite database

2009-09-28T22:26:00.000-07:00

We finally have an updated SQLite database for ConceptNet. It's designed for use with ConceptNet 4.0b8 (just released). This will fix the long-standing "best_raw_id" bug.

This database now includes data that was imported from the online game Verbosity. It also includes the initial import of the Chinese ConceptNet. This comes to us thanks to our collaborators in Taiwan, Jane Hsu, Yen-Ling Kuo, Edward Shen, and the many people playing the online games they developed.

We've also cleaned up our documentation, and written tutorials for some key things you may want to be able to do with ConceptNet and Divisi, at http://csc.media.mit.edu/docs/ .

Verbosity, and one meeeelion sentences

2009-08-20T10:26:00.001-07:00

How did we just get nearly 200,000 new statements in Open Mind Common Sense?

We've just imported a whole lot of data from Verbosity, one of Luis von Ahn's Games with a Purpose. Verbosity collects common sense knowledge through a game: one person is given a word, and needs to get the other person to guess that word by listing common-sense facts about it.

The data is rather noisy in places, but after some filtering, we've got a list of new statements about as reliable as the other score-1 statements in OMCS. These include a number of useful "is not" statements, describing things that are different, which we've never prompted for on OMCS before, as well as many examples of a new relation, "SimilarSize", expressing the statement "X is about the same size as Y".

A side effect of this is that it's pushed our total sentence count for English over one million! Of those, we can parse about 542,600 so far (we've still got a lot left to try to parse from the original Open Mind), and those translate to about 504,700 unique assertions in ConceptNet.

Thank you to all our contributors (especially those who are patient enough to try to deal with our current web site), and to all the players of Verbosity!

Welcome back, Catherine Havasi!

2009-08-10T10:33:00.001-07:00

Catherine Havasi co-created the Open Mind Common Sense project, as an undergraduate researcher working with Push Singh way back in 1999. For the last five years, she's been working on a doctorate in computational linguistics at Brandeis University. She's been doing a lot of cross-campus research with this group.

Last month, she finally earned her Ph.D (congratulations!). Now, she's returned to the Media Lab as a post-doc, where she'll once again be able to work on Open Mind and its applications full time. It's great to have her as an official part of the group again!

How to make Fink work when it has the wrong URL

2009-06-24T17:21:00.001-07:00

I hit a stumbling block today, and this is one of those things that really should be Googleable.

If you're installing libraries on a Mac, you might be doing it through Fink. And Fink has the unfortunate property that a lot of its download URLs give 404 errors, leaving you stuck. This was the case for the "cloog" library. I don't know what it does, but it's required by hdf5, which is needed by pytables, which we need to store Divisi results on the disk instead of in every instance of the Web server. All the URLs that Fink looks for when it tries to download "cloog" are broken.

The workaround is to Google for the file yourself, and download it into the /sw/src directory.

I hope this helps someone else who runs into the same problem.

Divisi for Windows

2009-06-24T17:19:00.001-07:00

The theme of this week is "make it so that our underlying code can actually be run by other people". One recent accomplishment: I finally figured out how to make a Windows installer of Divisi, our machine learning library. (The hard work to make Divisi compile on Windows at all was done by contributor Akshay Bhat. Thanks, Akshay.)

Bugfixes and improvements

2009-06-01T16:09:00.000-07:00

I know we're a bit quiet on the PR front, but as usual a lot is happening under the hood. You can see what we're up to by watching Launchpad, e.g., Divisi trunk and ConceptNet trunk. I'll highlight a few recent examples:

Divisi works on Windows. According to Akshay. We haven't tried it. Though properly supporting Windows means a double-click .exe installer -- which setuptools can apparently make, but we haven't figured out yet.
We finally renamed u_distances_to to u_dotproducts_with in SVD results; the name has been wrong ever since I wrote that code maybe a year and a half ago.
I wrote csc.divisi.util.PickleDir, which I've found really helpful for hanging onto temporary data that's a little longer-lived than an ipython session.
I improved how Divisi summarizes SVD results (in two commits).
I refactored how Analogyspace is built, making it easier to try out different combinations of things. Object-oriented code definitely improves things, but I still don't think I hit the sweet spot; certain customizations are still too hard. Any input from software architects? It's a set of mostly composable operations, though certain things only make sense in certain cases...
top_items had been effectively ignoring its new key parameter. Fixed.
We had been returning squared magnitudes for tensors. Oops. Fixed. Fortunately, I don't think this was used.
Finally got around to implementing the (pretty trivial) decomposition of a vector into the parallel and perpendicular components to another vector. That required filling out some other routines, fixing tests, etc.; I think they call that 'yak shaving'; my real goal was to figure out why AnalogySpace was coming out differently than a few weeks ago.
All this stuff is begging a new release. In time... for now, you can use the bzr head; we try not to break the trunk too often.

We also have a bunch of awesome documentation that hasn't gotten linked to in the main pages (e.g., conceptnet.media.mit.edu). More yak shaving: I went to edit the page and realized that we hadn't committed our local website changes into our web svn, so I committed some things, running into a svn bug and writing about it. But before I linked to the docs, I really wanted to move them to csc.media/docs, but that required mucking with the Apache configuration. I remembered that we had wanted to try out nginx, so I got that set up on a backup port on the server and got a dynamic site (csc.media) configured for it. I actually should have just stuck to getting the static config working and forward to Apache for the dynamic config, because that's what each is good at! Anyway, that required futzing with a fastcgi socket permissions issue (I tried using a Unix domain socket -- cool things, but as documented, they have permissions issues.) etc., etc.

I had planned to work on my thesis, but... speaking of thesis, Jayant just graduated. Thankfully he's staying a little while longer to wrap things up, so maybe he'll post something about his thesis.

Speed issues

2009-05-05T11:48:00.000-07:00

The Open Mind Common Sense website is currently really, really slow, and I'm sorry about that.

As we acquire more users and try to do more complicated reasoning behind the scenes, clearly what we need to do is ~~spend the piles of money that we have just sitting around on a huge fancy server~~

Sorry, I meant to say: clearly what we need to do is keep finding ways to cache lots of stuff and using whatever computing power we can find. Anyway, I'm working on it.

New site.

2009-04-22T15:23:00.001-07:00

We've got a new version of the Open Mind Common Sense site: openmind.media.mit.edu

The big changes:

It's based on the Pinax web framework. This should make it easier to add features to the site.
It's running on ConceptNet 3.5 instead of 3.0. (So was the old site, kinda, but it was a hack that wasn't sustainable.)
It distinguishes between "assertions", the normalized connections between concepts that Open Mind learns from, and "statements", the roughly parsed text that people have typed in. You can vote on both of them. This is a key step toward putting back the free text box.

New Mailing List

2009-02-26T14:54:00.000-08:00

Also in the realm of new and exciting stuff is our announcement mailing list. It's we'll use it for any big news we have and to announce workshops, symposiums, and software releases.

Subscribe yourself here!

Launchpad and Bazaar

2009-02-26T07:48:00.000-08:00

We're on Launchpad now. We can host our version control there, track bugs, and answer questions from users.

For people who work on Open Mind within the Media Lab (and possibly even others), here's a guide to hacking on the code using Bazaar.

IUI

2009-02-19T12:25:00.001-08:00

Hi everyone!

We've been quite busy lately and I'm going to make a point to update this more often. I really intend to, and since all I have to do this semester is graduate and transition to being a post-doc on top of Open Mind I should have lots of free time. What have we been up to? Well...

Most recently, Henry, Erik Mueller , and I ran a workshop on story understanding at IUI last week. Rodger Schank gave an interesting keynote on designing user interfaces using stories. He focused on how people interact and convey information naturally using stories rather than using the constructs which are common in user interfaces today. He talked about how "cavemen" communicated (ie, what are the modes of interaction we've been using all along), just-in-time information and making interfaces more goal-directed. We had a lot of good discussions, saw papers presented, had a demo session and many of us are still communicating by email. I'll try to post a more comprehensive summary later.

Also at IUI, Jayant and I gave a main conference talk on a paper by pretty much all of us using mixture models to play a game of twenty questions with the user. It asks questions which are selected to help AnalogySpace infer information about a new concept.

There's lots of stuff in the pipeline:

There are thoughts of an AAAI symposium in a year for the entire common sense community.
I'm working on a new technique for infusing normal data and reasoning techniques with common sense. It's working really well.
We're planning on putting up a lot of documentation soon and possibly some videos. Keep watching.

Assertions, sentences, and ratings

2008-08-11T12:02:00.000-07:00

Ken mentioned that we're in the middle of reorganizing the database. I'll fill in some more details about what we're doing.

Currently, our users give their ratings to assertions, the things that make up the links of ConceptNet. Many sentences can yield the same assertion: for example, "dogs are mammals" and "a dog is a kind of mammal" both turn into an assertion that can be expressed as IsA(dog, mammal). The ratings on these assertions are useful to representations such as AnalogySpace.

The problem is that the OMCS web site doesn't want to show you IsA(dog, mammal), it wants to show you something in natural language. And some of the natural language we've collected is of

What should matter to OMCS isn't just how good the abstracted assertions are, it's how good the sentences are.

So we're reorganizing the database. After this, your ratings will apply to the sentences you see, and the scores on assertions will come from aggregating those ratings. We'll display each assertion using its highest-rated sentence. This puts our users in charge of which sentences show up, instead of arbitrary decisions by the computer.

The hard part of the reorganization is that we have to take all of the existing ratings and find out where they came from. If they came from a user on the new site, for example, we need to know what sentence they were looking at when they gave the rating. The database generally has this information, but not necessarily recorded in a smart way.

So what we've really been doing is cleaning up messes in the database while we track down where the ratings should go. And most of which were created by me a couple of days before giving a demo. Sorry about that.

Facebook

2008-08-11T09:14:00.000-07:00

Hi everyone, I'm Catherine.

I will post some more substantive soon, but first-off I'd like to say we have a facebook group now. I'll hopefully update it with pictures and such as things become available.

In other news Rob, Henry, and I went to AAAI in Chicago to present AnalogySpace which went rather well. I had a great conference and really enjoyed the city of Chicago.

A Brief Note on Status

2008-08-08T21:07:00.000-07:00

First, thanks to everybody who has been contributing! We're running all sorts of cool analysis stuff with our data, and most everything you put it makes the analysis a bit better. And that's just the beginning...

We've gotten some feedback recently that basically suggests that our interaction with our user community has been lacking. That has definitely ratcheted up in our priorities, and we have a few things in the works. But for the moment, I thought I'd hit on a few items that people have asked about recently.

Usability: Obviously the site has usability flaws -- some larger than others. Specific feedback is most helpful. We've been doing a lot of grimy back-end work, but with your help we won't neglect the front end.
Speed: will improve a lot when we move to a server that doesn't have a game port in the back (really). There are certainly some optimizations to do also, but we're prioritizing increased functionality. Stay tuned...
Acquiring common sense from elsewhere. Though we haven't shared any of the work yet, we've actually done a lot of work in combining our dataset with other data to get interesting new results. Soon we'll be leveraging those tools to pull in large amounts of (hopefully useful) information from several user-contributed large databases of knowledge. One is, yes you guessed it, Wikipedia.
The "fix" flag is just a temporary flag until we implement the UI to edit stuff. It's not quite as simple as it seems because of the interaction with ratings, etc. We're reorganizing the database right now so that such things will become possible.
Stats are actually available on http://commons.media.mit.edu/en/stats/ which we haven't promoted yet because it's not complete or well-explained. We do have the raw data to do graphs and other spiffy stuff, but just haven't gotten around to it. Suggestions for neat graphing libraries, or just straight-up code contributions, are welcome.
Community involvement is very important to us, though we haven't been showing it yet. We're trying to run this as an open-source project. We haven't officially released the main website code, but we can send a tarball on request, and we're considering ways of doing better. If you can code, you'll be welcome to help out, or help recruit others. The site is written in Python, using Django and the ConceptNet and Divisi libraries that are already available (see the links on the home page).

If there are any other things you'd suggest we do to improve interaction with the user community, please share your views in the comments.

Also, feel free to ask anything -- about the site, the project, us, etc.; we'll try to respond.

-Ken

Introduction

2008-08-05T15:22:00.000-07:00

Hi, I'm Rob.

I've been working on stuff related to Open Mind since 2005, but I've had a bit of a low profile recently, because I thought I was going to go work on a different research project after I finished my master's thesis last year. After several months of dabbling in other projects and getting nowhere interesting, I found myself pulled back into OMCS. Moral of the story: if your grad school career ain't broke, don't fix it.

Probably the most visible things I've done are: I put OMCS back on the Web, replacing the 2000-era web site with a Rails site called Open Mind Commons. (When I left, I passed the torch to Ken, and he re-did it in Django, making the code incredibly cleaner in the process.) Also, along with Catherine Havasi, I developed AnalogySpace, the reasoning tool that learns from patterns in ConceptNet and comes up with the "Open Mind wants to know..." questions.

So now I'm back, and my focus is on the multilingual aspect of OMCS. We've got the infrastructure we need to build a ConceptNet in any language. Now our reasoning tools need to catch up.

I've seen the messages on the mailing list speculating that OMCS is stagnating. Well, it's not. It's changing in huge ways on our end, but it takes a while before we can put new features on the Web site. After all, we'd prefer the Web site to stay up.

But I understand that our users want to be in the loop. So I'm planning to write a few blog posts over the next few days about the new multilingual features I'm working on. Stay tuned.

All happy

2008-02-22T20:48:00.000-08:00

No news is good news? It seems everything is working well; a few bugs here and there still, but we'll get to them ;)

Feel free to play with the Explore Concepts demo... think Google Sets, but in a semantic network. I'm not sure why the web frontend is slow; the backend code is plenty fast. We'll figure that out sometime; for now, just be patient.

Any bugs, feature requests, ideas, etc., welcome; feel free to comment here.

-Ken

Server under siege

2008-02-08T23:29:00.000-08:00

Us and web robots are not getting along well. Our server has gotten overloaded several times recently to the point where we have had to bring it down until things settled. Apologies for the downtime; we'll get this stabilized soon.

But there is a new feature on the site... you're welcome to try it if you can find it, but it's still a little slow.

Adding knowledge enabled

2008-02-05T14:18:00.001-08:00

After fixing the last few bugs (hopefully!), we've turned adding new knowledge back on.

Ever wished your computer could get a clue? Well, now you can now to "Add New Knowledge" and give it one. If you put in something that someone else already said, we'll take it that you agree with them.

Better and more fun ways of clue-giving are coming...

-Ken

OpenMind Common Sense / ConceptNet: Welcome!

2008-01-29T13:05:00.000-08:00

For years we've enjoyed tens of thousands of good contributions of from our users, in terms of knowledge and ratings. So we thought that it's about time our users get some feedback from us. This blog is one way that you'll get to hear about what's going on with the site (like, why was it down?) and also see some of the ways that we're using the data you've been giving us.

So first, status: a batch of particularly malicious spam by one or two people hit our site two weeks ago. Since at the time we didn't have some of our spiffy tools quite ready, we decided to pull it down until the code monkeys caught up. To push them along, we went ahead and deployed the new version of the site, now powered by Python and Django.

We're still working hard to get everything working with the new framework. Some cool stuff is already available:

The site is now available in multiple languages! There's two parts to that: the user interface of the website itself, and the actual common sense data in the database. We have a snapshot of the Portuguese database collected by our Brazilian collaborators, and are starting projects in Dutch, French, and possibly Arabic.
Ratings are now much easier. Thumbs-up if it's good as-is. Thumbs-down if it's spam, nonsense, not something that any 10-year-old would know, or just not true. Yellow flag (meaning "Fix") if it would be true after some changes -- fixing grammar or spelling, or adding a qualifier ("sometimes", "occasionally").

Have fun with the new ratings system! We'll turn adding assertions back on real soon (though "Open Mind wants to know..." will take a bit longer).

back to coding,
-Ken
(a 1st-year Masters student... more later)