Make your idea brighter by leveraging on crowd-wisdom

Crowdsourcing is now being leveraged in several fields, as I had the opportunity of posting from time to time. There is also another trend developing, the mixing of biological intelligence with the computer based intelligence (AI…).

Add these two trends together and you understand what researchers at Rochester University are doing with their Chorus system.

Crowdsourcing has started as a way to amass the wisdom of many people using mechanisms to average out the responses and to pick those that might be most valuable. What Chorus does is to intercept chat discussion to extract the most meaningful answer to a query.

Chorus is a crowd powered chat system where the computer (Chorus) is actually the one chatting with you as you voice (type) your queries. The perception is to be talking to a real person, since the answers are a concentrate of answers provided by real persons on that topic.

There are now a number of companies working to leverage crowdsourcing power, like MobileWorks. We are moving one step up the ladder. The first one was to have the FAQ (frequently asked questions) prepared by an expert, the second was to open up discussion board having people asking and answering and providing a search mechanism to look at the various answers. Now, given the growing number of available answers coming from different people researchers are learning to apply Artificial Intelligence techniques to extract meaningful content from those crowdsourced data banks.

I guess a further step will come as more of these answer creating systems will become available so that meta answering systems will start to emerge, and one further step will be the AI answering systems evaluating the usefulness of the answers and challenging one another to improve them.

As things are becoming part of the Web, and Gaia goes on line, my bet is that we will see more and more objects participating in the crowdsourcing data creation…

Moving towards crowd-movies…

A new app is ready for you on the Apple Store. It is Vyclone, an app that takes advantage of the location and time information  attached to shootings you take with your iPhone.

Suppose you are watching a sport game and shooting. As you, there may b tens, hundreds of other spectators shooting the same event, although each one from a different position. Here comes Vyclone.

You can ask Vyclone to mash ups, in a completely automated way, your shooting with those of three other people that were filming at your same time. It can be done instantly, provided you and the others have an Internet connection, like a WiFi, and are on Vyclone as well, or it can be done the next day or the next year. Vyclone checks the location and time stamp on your video and looks for identical data in other video.

Then it mashes up the clips together providing a movies resulting from a multi camera shoot. It keeps one of the sound tracks so that you don’t get a confusing audio.

The effect is really nice, as you can see in the video below. Of course, once created, you can post the video to your social network, YouTube and the likes.

The present limitation of mashing up a maximum of four movies is going to be relaxed in the future, according to Joe Summer, one of the founder of Vyclone, who is also the Chief Creative Officer (CCO, I like the new acronym), and the system can scale up to accommodate several more cameras.

What I like is this idea of social shooting and the possibility to create something through crowd sourcing. It makes the world so much more connected and let eat of us see it with some others’ eyes. Also, I can imagine in the future that the mash ups will actually be producing as many layers as there are movies so that when you watch it you can decide to take the images from a different camera. Of course that would make the resulting film much bigger in terms of Bytes (but who cares tray about a file size?) and also much more demanding in terms of bandwidth if you are planning to stream it (but who’s gonna care about bandwidth in a few years time?).

The Cloud was born out of the possibility of sharing computer resources thus sharing investment and operation cost. Amazon has been one of the first to offer this service and it has been followed by many others.

The Cloud is therefore becoming  commodity but it is now moving to higher sky by providing additional features that, at least for me, were quite unexpected, such as the crowd sourcing cloud.

Enter mechanical turk, a tool provided by Amazon that let people collaborate to a project independently of one another, without knowing each other. Each one works independently and an artificial intelligence based tool, mechanical turk, integrate the individual results in something meaningful.

It is very easy, both to become part of the “crowd” and to exploit the “crowd”.

When you join the crowd you offer your intelligence and when needed you will be asked to perform a certain task requiring basic human intelligence (no rocket science). And you get paid for that, some 12 cents a minute, about 7$ per hour.

There are still many activities that are better done by humans than by machine. For sure one of this activity is to evaluate how a hum a respond to an interface. So if your company is about to launch a product and you want to test your interface, its seamlessness and understandability, you may want to test it on mechanical turk!

It is quite easy to make use of this tool, you just need to register on line, provide the money that will be used to pay the “crowd” (and Amazon) and load your task with instruction. Then lean back and let the “crowd” work on it.

When I checked this morning there were 236,331 HITs available (a HIT is a work item you can take up). And the system let you, as a potential worker, specify what you are good (or interested) at and will thereafter present you with those HITs that match your interest.

Every HIT indicates how much money you can make by participating in that activity. It is just a bit, actually, but most people participate for the pleasure of working on a certain thing. The beauty of the system is that it makes possible to harvest the willingness of individuals and direct it towards a goal.

Researchers at the Carnegie Mellon and Rutgers University have been using this tool to screen various applications on cell phones from the point of view of privacy to provide an intuitive feedback to people using those apps on their possible effect on private data.

Our dog can be a good sentry to our home and its barking can discourage intruders and alert us. But what if we are not home or if the intruders put the dog to sleep? Or if we have to go on a long trip and the dog cannot be left “on duty”?

My hight tech pet on patrol

Technology is coming to assistance in the form of a robot watching our premises. TAke a look at the paper:

http://www.mercurynews.com/business/ci_16534286?nclick_check=1

What hit me is the fact that today one can use o toy robot, spending less than 200$, to create a real sentry roaming the home and sending back images of what is going on. Clearly, at that price, the reliability is, to say the least, suspicious but, at that price, one could afford to buy more than one of them making up with volume what is missing in reliability.

A nice thing is that these “toys” come with customization capability and one can interact with them in many different ways. One of the researchers mentioned in the paper has decided to steer the robots through a headset he is wearing so that b turning his head the robot head turns and the camera picks up the images from that direction giving the feeling that one is actually there. To get more insight on what you can do with these robot toys you should spend some time on robocommunity.com.

I found interesting spending some of my time in that community: it made me realize how sophisticated toys have become and how they can actually become a platform for third parties to develop amazing applications.

It is yet another area where the concepts of crowd sourcing and ecosystems are changing the rules of the game.