Monday, 11 February, 2013
Erratic weather, ever changing sources of supply, fickle consumer tastes… what’s an orange juice manufacturer to do so as to keep ahead of the game? Simple, devise an algorithm – that takes into account a mere one quintillion possible variables – in the production of their juice:
The Black Book model includes detailed data about the myriad flavors – more than 600 in all – that make up an orange, and consumer preferences. Those data are matched to a profile detailing acidity, sweetness, and other attributes of each batch of raw juice. The algorithm then tells Coke how to blend batches to replicate a certain taste and consistency, right down to pulp content. Another part of Black Book incorporates external factors such as weather patterns, expected crop yields, and cost pressures. This helps Coke plan so that supplies will be on hand as far ahead as 15 months. “If we have a hurricane or a freeze,” Bippert says, “we can quickly replan the business in 5 or 10 minutes just because we’ve mathematically modeled it.”
algorithms, food, technology
Tuesday, 5 February, 2013
Guessing what will feature in the Hottest 100, and whereabouts, has been open to keen speculation for as long as Australian radio station triple j’s annual countdown of their listeners’ favourite music has been around.
It’s something that a couple of listeners, Nick Drewe and Tom Knox, clearly have an above average interest in however… they trawled through Hottest 100 related Twitter and Facebook posts in the lead up to last Australia Day’s countdown, and were able to, quite accurately, predict the composition of the chart:
Nick Drewe and his mate Tom Knox analysed about 35,000 Hottest 100 votes submitted to Triple J via social media sites in the run-up to Australia Day and published their own chart, the Warmest 100. The site accurately forecast 92 of the 100 songs on the chart. Messrs Drewe and Knox forecast all 10 songs in the Hottest 100 top 10, including five in the correct position.
algorithms, music, social media
Friday, 28 September, 2012
A sign of things to come? Some companies have taken to hiring staff by way of algorithms that tap into a job applicant’s skills, personality, and even their use of social networks, rather than the more traditional methods of evaluating past experience and conducting interviews.
Some employers have noticed a distinct increase in staff retention when recruiting staff in this fashion, particularly those looking for call centre workers, who have found attrition rates have dropped by about 20 percent.
When looking for workers to staff its call centers, Xerox Corp. used to pay lots of attention to applicants who had done the job before. Then, a computer program told the printer and outsourcing company that experience doesn’t matter. The software said that what does matter in a good call-center worker – one who won’t quit before the company recoups its $5,000 investment in training – is personality. Data show that creative types tend to stick around for the necessary six months. Inquisitive people often don’t.
algorithms, recruitment, technology, work
Friday, 17 August, 2012
If the geolocation data provided by our mobile phones isn’t quite enough to help others discern our future movements – assuming our activities adhere to some sort of recurring pattern – then tapping into such data on our friend’s phones may help pin us down in the event we unexpectedly change schedules:
Studies have shown that most people follow fairly consistent patterns over time, but traditional prediction algorithms have no way of accounting for breaks in the routine. The researchers solved that problem by combining tracking data from individual participants’ phones with tracking data from their friends – i.e., other people in their mobile phonebooks. By looking at how an individual’s movements correlate with those of people they know, the team’s algorithm is able to guess when she might be headed, say, downtown for a show on a Sunday afternoon rather than staying uptown for lunch as usual.
Slightly concerning, don’t you think?
algorithms, mobile-phones, privacy, technology
Tuesday, 1 May, 2012
Despite being a touch rigid, some of the stories written by algorithmic reporters, or computers programmed to write news reports, are anything but robotic or mechanical.
OK, it’s not Roger Angell. But the grandparents of a Little Leaguer would find this game summary – available on the web even before the two teams finished shaking hands – as welcome as anything on the sports pages. Narrative Science’s algorithms built the article using pitch-by-pitch game data that parents entered into an iPhone app called GameChanger. Last year the software produced nearly 400,000 accounts of Little League games. This year that number is expected to top 1.5 million.
Also of interest is the way these electronic journalists craft their news accounts, as they harvest data submitted by people, via smartphone apps, who are say spectators at a sporting event, and then use algorithms to weave divergent strands of information into a report.
algorithms, journalism, smartphones, technology
Monday, 12 March, 2012
Mathematicians – whose ranks most certainly do not include me – who play Sudoku, will be pleased to know that the popular number placement puzzle can feature sixteen starter clues, maybe even fewer, and still be solvable. This after it was widely believed at least seventeen clues were required to produce a solvable game.
Rather than searching through every sixteen-clue subset of a given Sudoku square, desperately looking for one that is actually a proper puzzle, we need only consider sets of sixteen starting clues containing at least one cell from each known unavoidable set. Finding those particular sets of starting clues is a specific instance of a more general problem, known to mathematicians as the “hitting set problem.” The really clever part of McGuire’s work is the development of algorithms for solving the hitting set problem in a reasonable amount of time. Solving the minimum-clue problem for Sudoku was just an application of this new algorithm.
algorithms, games, mathematics, puzzles, Sudoku
Monday, 19 December, 2011
Either elevators arrive in packs or they’re no where to be seen when you’re looking for one… is it possible to develop a more efficient algorithm though for controlling idle lifts?
All elevator algorithms solve the same type of optimization problem: given that a building has n floors and m elevators, how could we most efficiently move people up/down the floors? I’m sure you already know of the simple algorithm that every elevator implements, but one can definitely improve on this.
algorithms, elevators, lifts
Monday, 12 December, 2011
Facial recognition software is becoming increasingly sophisticated, and ever more capable of positively spotting family members by comparing different photos. Who knows, if advances in the technology continue at their current pace, it could even end up supplanting other forms of parental testing.
The most predictive features were the darkness and colour of the eyes, the darkness and colour of the skin, and the distances between the nose and mouth and the eyes and nose. The team tested the abilities of human volunteers to spot family ties, and found that humans correctly identified parent-child matches and mismatches 67 per cent of the time, slightly worse than their algorithm’s 71 per cent.
algorithms, facial recognition, family, photography
Monday, 5 December, 2011
Massachusetts Institute of Technology (MIT) researchers have devised an algorithm that can work out which cars are likely to run red traffic lights, a development that could boost safety at intersections.
In order to reduce the number of accidents at intersections, researchers at MIT have devised an algorithm that predicts when an oncoming car is likely to run a red light. Based on parameters such as the vehicle’s deceleration and its distance from a light, the group was able to determine which cars were potential “violators” – those likely to cross into an intersection after a light has turned red – and which were “compliant.”
algorithms, cars, driving, road safety, traffic lights
Tuesday, 30 August, 2011
A Cornell University team is developing algorithms that can help detect fake reviews that are posted online for goods and services such as books and hotels.
So the team developed an algorithm to distinguish fake from real, which worked about 90 percent of the time. The fakes tended to be a narrative talking about their experience at the hotel using a lot of superlatives, but they were not very good on description. Naturally: They had never been there. Instead, they talked about why they were in Chicago. They also used words like “I” and “me” more frequently, as if to underline their own credibility.
algorithms, feedback, review