Human computer collaboration

Human computer collaboration:

The article I link here (titled “Prepare for the Next Round of Man vs. Machine” – in hebrew, and an original piece, sorry :S) begins by presenting Eureqa – a program that can build mathematical models to describe various trends in the data provided to it – and then discusses various views on how the integration of computers has and will affect scientfic research. The undertone of the article was “man vs. machine” as the title suggested, but frankly, besides vague arguments saying humans will allways be neccesairy in science, I found it to be more on the practicality of machine aided research rather than true competion.

And there here’s the reason for that – this so-called contest between man and machine is both unhelpfull and impractical. I say it is unhelpful, because computers are tools, and not using them to their advantage because of some ego contest means wasting your own time on something computers are better at by their very nature. I say it is impractical, because we have yet to define the contest well enough, or the parameters by which to measure success.

For example, the article discusses the buzz created around Eureqa’s presentation to the public. The immediate response of non-scientific media to news about it’s “discovery” of the underlying mathematical model of Newton’s second law (f=ma) meant that it could be a better scientist than any man, having discovered this theory on its own. Scientists then critisized the program, saying it still needed humans to define the variables and the form of the function. The developers went on to say they never claimed the program worked out of nothing – the program uses random attmpts to change a basic function involving the variables from the data given to it, nothing less or more. Obviously I’m paraphrasing a little, but the general idea still holds – “inventing” is not the same as “discovering” or “defining” – which means that the arguments discussed incomparable aspects of the scientific process. All this – instead of embracing a new tool that can help verify scientific research, support it, help scientists work faster!

The point I’m trying to make is this – why even compare humanity to machines. Any apocalyptic sci-fi will tell you that making machines more similar to human’s is bad. While obviously this is an oversimplification, I still think that looking to mimic or compete with something that’s allready there is useless. Instead, why not look for ways to complement our thinking. We are allready forced to think about what makes us the way we are to improve machine learning and thinking. I personaly think that we should use this knowledge to recognize processes where computers are better, and find how to integrate them with human reasoning.

If we return to Eureqa’s example – the program can sort out a big batch of data and verify many different equations that represent it, but it can only assess the quality of it’s resultant functions and guess randomly what changes to make to the function in order to improve it. Why not look for ways to take the parts of the program scientists are good at – recognizing and proving function are good – and integrate that process into the program. Why not take it further and allow any human do it, like in FoldIT.

I hope we’ll eventually get over our egos and paranoia and focus on the usability of our inventions, leading to advancements in all fields of research.

via tumblr at January 27, 2012 at 09:57PM. Originally posted on


Tags: , , , , ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: