The Gradual Public Awareness of the Might of Algorithms

Soylent Mauve writes "The trend toward data- and algorithm-driven tuning of business operations has gotten a lot of attention recently — check out the recent articles in the New York Times and the Economist. It looks like computer scientists, especially those with machine learning training, are getting their day in the sun. From the NYT piece: 'It was the Internet that stripped the word of its innocence. Algorithms, as closely guarded as state secrets, buy and sell stocks and mortgage-backed securities, sometimes with a dispassionate zeal that crashes markets. Algorithms promise to find the news that fits you, and even your perfect mate. You can't visit Amazon without being confronted with a list of books and other products that the Great Algoritmi recommends. Its intuitions, of course, are just calculations -- given enough time they could be carried out with stones. But when so much data is processed so rapidly, the effect is oracular and almost opaque.'"

Comments

Oracular, opaque...

[Anonymous Coward]

and often hilarious or silly. People already trust computers too much.

Re:Oracular, opaque...

[mahmud]

The day the computers can read what we're looking at and know us well enough to offer an even remotely successful guess at what comes next will be the day the computer decides it doesn't need us anymore. And I think we all know what happens when the computers decide they don't need us anymore. No.

Don't apply your intuition concerning human beings to other intelligent systems. A true AI may or may not decide it doesn't need us, depending on how it's programmed.

You ignore the fact that stand-alone sentience has little to do with our evolution-dictated habits (e.g. getting rid of competing group/species/whatever). You assume that all the evolution-dictated behaviour and thinking patterns embedded in human brain will somehow automagically manifest themselves in a true strong-AI machine, a view with which I disagree.

Boy They're Slow

[PingPongBoy]

Whereas algorithms are instantly aware of their own prowess.

Is management starting to wonder (again) whether a computer can really do a better job making the important decisions? But can it yet? There is so much data that needs to be acquired in order to return a meaningful answer.

Some of the most powerful organizations are probably making deals to combine as many databases as possible. Interesting to see (if they would let us see) if that will give them the answers they're looking for. As data acquisition becomes more accurate and less expensive, there will be less privacy but more creative computer output, a trade-off in the value of personal information leading to the possible marginalization of humanity.

Heuristics are not the same as algorithms

[pedantic bore]

Sheesh! Someone needs to spend some time with a dictionary.

If only we could have a gradual (or sudden) awareness of the power of heuristics and modeling ...

The Importance of a CS Degree

[Enonu]

This is one of the reason why getting a CS degree is important, despite what the ignorant masses say in the IT industry. Sure writing lame CRUD applications will satisfy your average customer's needs, but sophisticated algorithms are what provide value beyond a simple shopping cart.

If you're still entrenched in the thought that a CS degree "isn't needed for what I do," then let me propose a somewhat common problem. Suppose your client wants the built in reporting in your web application to minimize the amount of noise introduced by users who forget their password and create a new account rather than resetting it. It's up to you to write code to detect these duplicate accounts. How do you begin doing this beyond simple string comparisons?

Re:The Importance of a CS Degree

[neonfreon]

Why would anyone ever do this? "Excessive noise"? Oh, you mean more orders? Last time I checked, more entries in the database never hurt anyone (not like every user is going to create duplicate accounts to the point where you're running out of resources, user records are tiny anyhow). Writing some 'intelligent' algorithm to detect duplicate accounts will invariably lead to marking legitimately separate user's accounts as duplicates and eliminating business.

Ahh, but experience matters too..

Re:The Importance of a CS Degree

[CodeBuster]

How do you begin doing this beyond simple string comparisons?

It is also useful to realize that just because one can does not mean that one should, especially when the cost of an error is high. There is a tendency, sometimes, among the computer scientists towards too much cleverness, particularly in algorithms, when something much simpler and more reliable would have been better. I cannot tell you how many times bad assumptions about automated processes and the algorithms which control them have lead to inappropriate behavior and blown user expectations under the worst possible conditions. The real world is not the same as the CS labs in your algorithms course and the simpler solution often has much to recommend itself over the efficient and elegant, but hopelessly complex and slightly unreliable algorithm that one learns in the AI courses during their university CS education.

For example, suppose that your online banking application assumes that you really do want that regular payment upon receipt to go through automatically, because that is how it has happened before, when in fact you, the user, know that a one time payment for an unrelated expense, which has not yet been posted but will be shortly, must be made first. The automated agent makes the deduction for the regular payment automatically while the one time payment, which goes through several days later, is unexpected and overdraws the account. The user curses the system for being too "clever" instead of just carrying out his instructions. Cancel or allow?

Re:Heuristics ARE algorithms

[Chmcginn]

GP: "Heuristics are not the same as algorithms"

P: "Heuristics ARE algorithms"

Both of these statements can be true. (Depending on the exact meaning of the GP.) For instance:

Humans are not the same as animals.

Humans are animals.

A more exact statement than either is that heuristics are a subset of algorithms, as humans are a subset of animals.

Re:This Just In

[GuyMannDude]

Math is a really really powerful tool.

While that may be obvious for slashdot readers, it's news to the general public. I remember an endless number of conversations, even as recent as a few years ago, in which people would ask "Can you do anything with that degree other than teach?" upon learning that I was a mathematician. I think it's great that the public is starting to realize that math makes the world go around. God forbid, the gradual public awareness of the power of math might even lead to kids wanting to pay attention in class. While there are drawbacks to this (e.g., the deluge of college kids taking business-oriented mathematics programs with the expectation of a six-figure salary once they graduate), I'm generally happy to see math and computer science get their days in the sun.

GMD

Re:Software Patent Propaganda.

[ShakaZ]

Without algorithms, there can be no computing but there's nothing really special about any one in particular. Algorithms are just instructions, and there are many ways of achieving the same result. There are indeed many ways of achieving the same results. However some algorithms are much more efficient than others and when a large program is made up of many of those optimised algorithms there will be a huge speed improvement over other implementations.
This is especially important in resource-hungry applications, scientific calculation or on systems with constrained resources such as embedded systems.

Saying there's nothing special about any algorithm is simply dumb.