Feeds

Mathematica man brews 'AI' Google Killer™

A New Kind of Pseudo-Science

Build a business case: developing custom apps

Stephen Wolfram - the lovable George Costanza of the mathematics community who developed the invaluable Mathematica suite and wrote the much talked about but quickly forgotten "A New Kind of Science” - is trying his hand at artificial intelligence.

His new project, Wolfram Alpha, set to go live in May, combines natural language processing with machine understanding. You'll be able to get succinct answers to questions like "When was Google's stock at $300 per share?" or "How much did it snow in New England last year?” Allegedly.

It's a noble goal, to aggregate human knowledge in a large machine brain that's able to answer questions. The problem is: We already have Wikipedia and Google that - together - get the job done well enough. So, Wolfram Alpha seems more like a science fair project than a serious stab at machine intelligence. In a company blog post announcing it, Stephen Wolfram himself says that his new project started as a wouldn't-it-be-neat-if brainstorm.

"Fifty years ago, when computers were young, people assumed that they’d quickly be able to handle [systematic knowledge]. And that one would be able to ask a computer any factual question, and have it compute the answer. But it didn’t work out that way. Computers have been able to do many remarkable and unexpected things. But not that.”

Yes, true, computers have not yet been able to process knowledge on the level promised by science fiction novels of the nineteen fifties, but why not? It's not for lack of trying. AI researchers decades ago spent a lot of time and money and programmed a lot of Lisp, only to come up with a therapist named Eliza with whom you can hold a hollow conversation without the $150 per hour fee. It's not as if this year computers finally became powerful enough for proper machine understanding.

No, computers haven't solved this problem because there are no people who actually need it solved.

Stephen Wolfram has a history of being the answer to a question that nobody asked. His 2002 self-published manifesto, “A New Kind of Science” (abbreviated NKS for those of you who prefer the Church of Scientology method of using acronyms to make yourself sound more serious), ruffled some feathers in the scientific community. Wolfram argued that studying simple cellular automata, similar to Conway's Game of Life, will lead to greater discoveries in science. NKS was criticized for being an answer without a question, and it's possible that Wolfram is using this new project as a justification for his book:

"I’d always thought, though, that eventually [machine understanding] should be possible. And a few years ago, I realized that I was finally in a position to try to do it. I had two crucial ingredients: Mathematica and NKS. With Mathematica, I had a symbolic language to represent anything—as well as the algorithmic power to do any kind of computation. And with NKS, I had a paradigm for understanding how all sorts of complexity could arise from simple rules."

If you focus your attention and listen closely, you can hear his ego approaching critical mass, preparing to implode on itself.

Business school lecture aside, Wolfram does deserve some credit where it is due. The Mathematica software package has been - and will continue to be - a critical resource for many people who work in quantitative science. Wolfram Alpha is built on top of Mathematica, which shows how wide a range of problems this software can solve. So, is machine understanding a new and upcoming feature in the next version of the world's most expensive scientific calculator? Unlikely. From the announcement of Wolfram Alpha:

"Some people have thought the way forward must be to somehow automatically understand the natural language that exists on the web. Perhaps getting the web semantically tagged to make that easier. But armed with Mathematica and NKS I realized there’s another way: explicitly implement methods and models, as algorithms, and explicitly curate all data so that it is immediately computable."

If you've ever dealt with real world machine learning, your snake oil detector should be deafening you right now. Explicitly curate all the data? Surely, Stephen, you have come up with an elegant mathematical solution to do this? After all, anybody who has done real world machine learning will tell you that the vast majority of your time is spent cleaning the data. It's a cruel twist that academics who teach machine learning gloss over the most important part and simply focus on the clean mathematical models. So, if this were a real breakthrough in machine intelligence, the input would be essentially arbitrary, but how does Wolfram Alpha do it?

"Every different kind of method and model—and data—has its own special features and character. With a mixture of Mathematica and NKS automation, and a lot of human experts, I’m happy to say that we’ve gotten a very long way."

That sounds an awful lot like the marriage of some Python scripts with a few hundred bucks spent hiring third world workers through Amazon Mechanical Turk.

Given that, Wolfram Alpha doesn't seem terribly innovative. Correct me if I am wrong, and I know you will, but there was a programming language in 1972 called Prolog that could take carefully curated declarative statements and allow you to run logical queries over them. Something like "If the standard rate of chucking is 10 logs per minute, and all woodchucks can chuck, how much wood could a woodchuck chuck if a woodchuck could chuck wood?”

Fortunately, the full on boot to the face of media hype hasn't started yet. This is a bit of a curiosity, because Wolfram Alpha makes for a good "underfunded, smart guy taking on Google.” We heard that story with Powerset and Cuil, both of which amounted to nothing more than a comedy act. Who knows, maybe Stephen Wolfram really is cooking up something. Maybe he's just feeding his ego. In May, we'll get to see how useful the system really is. Whichever the case may be, answering the question "When was Google's stock at $300?" is a parlor trick. Answering the question "When will Google's stock be worth $300" - that might be worth something. ®

Ted Dziuba is a co-founder at Milo.com You can read his regular Reg column, Fail and You, every other Monday.

Gartner critical capabilities for enterprise endpoint backup

More from The Register

next story
'Stop dissing Google or quit': OK, I quit, says Code Club co-founder
And now a message from our sponsors: 'STFU or else'
Why has the web gone to hell? Market chaos and HUMAN NATURE
Tim Berners-Lee isn't happy, but we should be
Microsoft boots 1,500 dodgy apps from the Windows Store
DEVELOPERS! DEVELOPERS! DEVELOPERS! Naughty, misleading developers!
Mozilla's 'Tiles' ads debut in new Firefox nightlies
You can try turning them off and on again
Apple promises to lift Curse of the Drained iPhone 5 Battery
Have you tried turning it off and...? Never mind, here's a replacement
Uber, Lyft and cutting corners: The true face of the Sharing Economy
Casual labour and tired ideas = not really web-tastic
Linux turns 23 and Linus Torvalds celebrates as only he can
No, not with swearing, but by controlling the release cycle
prev story

Whitepapers

Top 10 endpoint backup mistakes
Avoid the ten endpoint backup mistakes to ensure that your critical corporate data is protected and end user productivity is improved.
Implementing global e-invoicing with guaranteed legal certainty
Explaining the role local tax compliance plays in successful supply chain management and e-business and how leading global brands are addressing this.
Backing up distributed data
Eliminating the redundant use of bandwidth and storage capacity and application consolidation in the modern data center.
The essential guide to IT transformation
ServiceNow discusses three IT transformations that can help CIOs automate IT services to transform IT and the enterprise
Next gen security for virtualised datacentres
Legacy security solutions are inefficient due to the architectural differences between physical and virtual environments.