Takeaways: GA’s ‘Making life easy with algorithms’ class

25 Feb

Class: General Assembly session on “Make Life Easy With Algorithms“, led by Adam Levine of the Metropolitan Museum of Art.

Date: February 22 2012, 20:00-21:00

This was my second class at GA (the first covered mobile strategy). This session, only an hour long, would best be described as “thinking about thinking about algorithms” – it was a high-level overview of the history of algorithms, what they are, their modern applications and a short “how-to” of writing them.

First, some quotes from Levine (an art historian and mathematician) on the nature of algorithms:

  • Algorithms aren’t about complicated math; they’re about logic
  • Algorithms are any systematic, mathematical method for calculation
  • If the [output of the algorithm]’looks’ wrong, it probably is
  • There should always be as many feedback loops as possible built in

Levine spoke briefly about the difference between “deterministic” and “probabilistic” algorithms, and about the utility of visualizing problems in terms of a network (as he put it, networks transform problems into questions of ‘distance‘.)

As for modern applications of algorithms, he cited LTCM and Renaissance Technologies from the finance world (the latter a hedge fund known for high-frequency and algorithmic trading and founded by a mathematician named James Simons), and Google and Match.com (of course) from a pure(er) technology perspective.

In terms of applying algorithms to business cases (the “make life easy with…” part of the session), Levine stressed the importance of “systematizing the process”, going for accuracy over optimization, and asking the following questions:

  1. What is the problem the business is trying to solve?
  2. How do people currently resolve the problem?
  3. Do people believe this is the right approach?
  4. If so, why is there an opportunity?
  5. If not, why is it the current approach wrong?

And finally, if you’re thinking about writing algorithms, here’s a (view from 10,000 feet) how-to:

  • Write the overview of the specification in prose (i.e. what do you want the algorithm to do?) (‘High level’)
  • Write each step in prose (what are the different parts needed to achieve the goal?) (‘Implementation’)
  • Code each step (‘Formal’)

This class definitely left me wanting more – but perhaps that was the point. As Levine said at the beginning, algorithms are about logic. At the very least I left thinking about the world in a more structured way.

See also:
Predicting a Red-Light Runner? M.I.T. Has an Algorithm for That – NY Times Bits Blog
The Dubious Science of Online Dating – NY Times

Advertisements

One Response to “Takeaways: GA’s ‘Making life easy with algorithms’ class”

Trackbacks/Pingbacks

  1. Takeaways: Alex Taub’s ‘Intro II Business Development & Partnership for Startups’ Skillshare class « Ent! - March 5, 2012

    […] on Ent! Takeaways: GA’s ‘Making life easy with algorithms’ class Share this:MoreLike this:LikeBe the first to like this […]

Comments are closed.

%d bloggers like this: