algorithms

An Algorithm and 7 Million Unique Nutella Jars

readingbyeugene Leave a comment

Earlier this year, Milan-based design agency Ogilvy & Mather partnered with Nutella manufacturer Ferrero to unveil its “Nutella Unica” jars. The agency created an algorithm that generated 7 million unique variants of Nutella jars, from an assemblage of various patterns and colors. Ferrero sold these jars in Italy throughout the month of February; each of the 7 million unique jars sold out in a month.

Here’s a brief video of the manufacturing process showing the unique designs:

A 30-second spot in Italy highlighting these unique jars:

Due to the success of the campaign in Italy, Ogilvy & Mather and Ferrero have decided to sell these jars elsewhere in continental Europe, beginning with France.

It will be cool to see the unique jars make it to the United States. Also, from a coding/machine learning perspective, it would be really neat to see the source code/implementation of this algorithm.

 

What Happened in the Markets on September 18, 2013 at 2PM?

readingbyeugene Leave a comment

This is an intriguing analysis at Nanex of what happened in the financial markets (equities and futures) on September 18, 2013 milliseconds before the FED announcement of “no taper” at precisely 2:00PM.

One of Einstein’s great contributions to mankind was the theory of relativity, which is based on the fact that there is a real limit on the speed of light. Information doesn’t travel instantly, it is limited by the speed of light, which in a perfect setting is 186 miles (300km) per millisecond. This has been proven in countless scientific experiments over nearly a century of time. Light, or anything else, has never been found to go faster than 186 miles per millisecond. It is simply impossible to transmit information faster.

Too bad that the bad guys on Wall Street who pulled off The Great Fed Robbery didn’t pay attention in science class. Because hard evidence, along with the speed of light, proves that someone got the Fed announcement news before everyone else. There is simply no way for Wall Street to squirm its way out of this one.

Before 2pm, the Fed news was given to a group of reporters under embargo – which means in a secured lock-up room. This is done so reporters have time to write their stories and publish when the Fed releases its statement at 2pm. The lock-up room is in Washington DC. Stocks are traded in New York (New Jersey really), and many financial futures are traded in Chicago. The distances between these 3 cities and the speed of light is key to proving the theft of public information (early, tradeable access to Fed news).

We’ve learned that the speed of light (information), takes 1 millisecond to travel 186 miles (300km). Therefore, the amount of time it takes to transmit information between two points is limited by distance and how fast computers can encode and decode the information on both sides.

Our experience analyzing the impact of hundreds of news events at the millisecond level tells us that it takes at least 5 milliseconds for information to travel between Chicago and New York. Even though Chicago is closer to Washington DC than New York, the path between the two cities is not straight or optimized: so it takes information a bit longer, about 7 milliseconds, to travel between Chicago and Washington. It takes little under 2 milliseconds between Washington and New York.

Therefore, when the information was officially released in Washington, New York should see it 2 milliseconds later, and Chicago should see it 7 milliseconds later. Which means we should see a reaction in stocks (which trade in New York) about 5 milliseconds before a reaction in financial futures (which trade in Chicago). And this is in fact what we normally see when news is released from Washington.

However, upon close analysis of millisecond time-stamps of trades in stocks and futures (and options, and futures options, and anything else publicly traded), we find that activity in stocks and futures exploded in the same millisecond. This is a physical impossibility. Also, the reaction was within 1 millisecond, meaning it couldn’t have reached Chicago (or New York): another physical possibility. Then there is the case that the information on the Fed Website was not readily understandable for a machine – less than a thousandth of a second is not enough time for someone to commit well over a billion dollars that effectively bought all stocks, futures and options.

The conclusions the authors draw? The announcement was leaked:

The Fed news was leaked to, or known by, a large Wall Street Firm who made the decision to pre-program their trading machines in both New York and Chicago and wait until precisely 2pm when they would buy everything available. It is somewhat fascinating that they tried to be “honest” by waiting until 2pm, but not a thousandth of a second longer. What makes this a more likely explanation is this: we’ve found that news organizations providing timed release services aren’t so good about synchronizing their master clock – and often release plus or minus 15 milliseconds from actual time. Their news machines in New York and Chicago still release the data at the exact same millisecond, but with the same drift in time as the master clock. That is, we’ll see an immediate market reaction at say, 15 milliseconds before the official scheduled time, but in the same millisecond of time in both New York and Chicago. Historically, these news services have shown a time drift of about 30 milliseconds (+/- 15ms), which places the odds that this event was from a timed news service at about 10%. 

Something does sound fishy based on the charts provided by Nanex. I’ve read some of their analyses before and they have been overwhelmingly convincing. We’ll see how this one unfolds pretty soon, I think.

The Lucrative Business of Hollywood Script Doctors

readingbyeugene Leave a comment

After the success of House of Cards on Netflix, reading this piece is both interesting and disturbing:

A chain-smoking former statistics professor named Vinny Bruzzese — “the reigning mad scientist of Hollywood,” in the words of one studio customer — has started to aggressively pitch a service he calls script evaluation. For as much as $20,000 per script, Mr. Bruzzese and a team of analysts compare the story structure and genre of a draft script with those of released movies, looking for clues to box-office success. His company, Worldwide Motion Picture Group, also digs into an extensive database of focus group results for similar films and surveys 1,500 potential moviegoers. What do you like? What should be changed?

“Demons in horror movies can target people or be summoned,” Mr. Bruzzese said in a gravelly voice, by way of example. “If it’s a targeting demon, you are likely to have much higher opening-weekend sales than if it’s summoned. So get rid of that Ouija Board scene.”

Bowling scenes tend to pop up in films that fizzle, Mr. Bruzzese, 39, continued. Therefore it is statistically unwise to include one in your script. “A cursed superhero never sells as well as a guardian superhero,” one like Superman who acts as a protector, he added.

His recommendations, delivered in a 20- to 30-page report, might range from minor tightening to substantial rewrites: more people would relate to this character if she had a sympathetic sidekick, for instance.

How soon before this goes mainstream and all scripts are run through an algorithm?

For what it’s worth, I agree with the writer quoted in the piece:

It’s the enemy of creativity, nothing more than an attempt to mimic that which has worked before. It can only result in an increasingly bland homogenization, a pell-mell rush for the middle of the road.

On Algorithmic Gatekeepers

readingbyeugene Leave a comment

Writing for the opinion section of The New York Times, Evgeny Morozov characterizes “one dimensional” algorithms which censor the Web:

The proliferation of the Autocomplete function on popular Web sites is a case in point. Nominally, all it does is complete your search query — on YouTube, on Google, on Amazon — before you’ve finished typing, using an algorithm to predict what you’re most likely typing. A nifty feature — but it, too, reinforces primness.

How so? Consider George Carlin’s classic comedy routine “Seven Words You Can Never Say on Television.” See how many of those words would autocomplete on your favorite Web site. In my case, YouTube would autocomplete none. Amazon almost none (it also hates “penis” and “vagina”). Of Carlin’s seven words, Google would autocomplete only “piss.”

Until recently, even the word “bisexual” wouldn’t autocomplete at Google; it’s only this past August that Google, after many complaints, began to autocomplete some, but not all, queries for that term. In 2010, the hacker magazine 2600 published a long blacklist of similar words. While I didn’t verify all 400 of them on Google, a few that I did try — like “swastika” and “Lolita” — failed to autocomplete. Is Nabokov not trending in Mountain View? Alas, these algorithms are not particularly bright: unable to distinguish between Nabokov’s novel and child pornography, they assume you want the latter.

Why won’t tech companies let us freely use terms that already enjoy wide circulation and legitimacy? Do they fashion themselves as our new guardians? Are they too greedy to correct their algorithms’ mistakes?

Thanks to Silicon Valley, our public life is undergoing a transformation. Accompanying this digital metamorphosis is the emergence of new, algorithmic gatekeepers, who, unlike the gatekeepers of the previous era — journalists, publishers, editors — don’t flaunt their cultural authority. They may even be unaware of it themselves, eager to deploy algorithms for fun and profit.

“I, for one, welcome our new algorithmic overlords.” –Said almost no one.

Algorithms Invading Our Lives

readingbyeugene Leave a comment

From this Wall Street Journal piece, we learn about the proliferation of algorithms. I am not convinced about algorithms picking out creative works (music hits and potential blockbuster movies), but I found this bit interesting:

Algorithms also have invaded areas of our lives that might seem too personal for mere automation. We are all familiar with the words “this call may be recorded for quality or training purposes.” Though that message may sometimes mean just what it says, it often means that an algorithm has been invited in for a listen.

Using only the words you say in a three-minute conversation, more than five million eavesdropping algorithms, created by a company called Mattersight, determine your personality type, what you want and how you might be most easily and quickly satisfied by the customer-service agent. The electronic psychological analysis divides people into six sorts of personalities. Steve Jobs, for instance, was a “reactions-based” person, someone who responds strongly to things: “I hate that!”

The next time you call, the algorithms, recognizing your phone number, will route you to an agent with a personality similar to your own, which results in calls that are half as long and reach happy resolutions 92% of the time, compared with 47% otherwise, according to an assessment of 1,500 customer service calls at Vodafone, the European telecom company.

What have algorithms done for you lately?

Link of the Day (01/25/10)

readingbyeugene 1 Comment

There is one article I want to highlight for today. It is so interesting that it deserves to stand on its own as the link of the day.

(1) “The Chess Master and the Computer” [New York Review of Books] – an incredibly well-written and thought-provoking piece by Garry Kasparov, perhaps the greatest chess player of all time. In the article, Garry Kasparov discusses his play against computers, from the 1980s to the showdown with Deep Blue in 1997 to playing against modern computer chess programs.

Most intriguing to me are Mr. Kasparov’s thoughts on the possibility of solving chess. Imagine this scenario: you make a move in chess, and the computer would be able to calculate the best move under the circumstances and predict the likelihood of achieving mate (and in how many moves it will occur). The concept of solving chess is something I have been thinking about for over ten years, so it’s refreshing to read a Grandmaster’s opinion:

Another group postulated that the game would be solved, i.e., a mathematically conclusive way for a computer to win from the start would be found. (Or perhaps it would prove that a game of chess played in the best possible way always ends in a draw.) Perhaps a real version of HAL 9000 would simply announce move 1.e4, with checkmate in, say, 38,484 moves. These gloomy predictions have not come true, nor will they ever come to pass. Chess is far too complex to be definitively solved with any technology we can conceive of today.

So Mr. Kasparov is not excluding the possibility of chess being solved one day; he simply argues that it is inconceivable to solve the game of chess with the hardware we have today. Mr. Kasparov goes on to explain:

The number of legal chess positions is 1040, the number of different possible games, 10120. Authors have attempted various ways to convey this immensity, usually based on one of the few fields to regularly employ such exponents, astronomy. In his book Chess Metaphors, Diego Rasskin-Gutman points out that a player looking eight moves ahead is already presented with as many possible games as there are stars in the galaxy. Another staple, a variation of which is also used by Rasskin-Gutman, is to say there are more possible chess games than the number of atoms in the universe. All of these comparisons impress upon the casual observer why brute-force computer calculation can’t solve this ancient board game.

If you are at all interested in chess, computer science, or algorithms, I highly encourage you to read the entire article.