The algorithm that won an Oscar

Hollywood likes a good explosion. Now with the help of an open source algorithm called Wavelet Turbulence, filmmakers can digitally create smoke and fire effects that used to be time-consuming and difficult to control.

UCSB’s Theodore Kim (along with three collaborators) picked up the Academy Award in Technical Achievement for Wavelet Turbulence. So far, it has been used in a number of major Hollywood films including Avatar, Iron Man 3, Man of Steel, and Super 8 (pictured above).

Kim’s work is a true melding of art with physics and computer science. Even though he uses a lot of math and numerical data in his work, there are no quantitative answers. In the end, it all comes down to the way the effect looks. 

If something doesn’t look quite right, it’s not always clear what the problem. You need to use your intuition and some visual imagination to figure out how to make the effect look believable.

Curious how his software works? Check out the video below:


Pleasant Places

Tech artwork by Quayola explores the subject of natural landscapes and, with algorithms, display abstractions caused by natural forces:

Titled like the first series of landscapes prints produced in Holland in the Seventeenth century, Pleasant Places consists of a series of digital paintings exploring the boundary between representation and abstraction.
Inspired by the work of Vincent Van Gogh, Quayola has returned to the same countryside of Provence 125 years later. The landscapes serve as a point of departure - a pretext to shape an inner motion and vision.
Through the misuse of image-analysis and manipulation algorithms, Pleasant Places challenges the photographic image and proposes alternative modes of vision and synthesis. Familiar landscapes - filmed in Ultra-High-Definition - is shown with meticulous attention to details and to the anthropomorphic shapes of the trees. Then, through the use of custom-software, the detailed texture of the foliage is reduced to two-dimensional masses of volume veering towards abstraction. As the outlines of trees and shrubs get blurred, nature becomes dense and almost impenetrable. The resulting compositions remain, suggestively, suspended between representation and abstraction, between the depth of the natural scenery and the surface of the screen.
In contrast to this vision, raw data-visualisations of colour and motion information follow in sequence the contemplative digital paintings to remind us what really lies beneath the surface. Pleasant Places pays homage to the modern tradition of Western art that takes landscape as a point of departure towards abstraction, reducing the complexity of the world into new alternative synthesis.

More Here


Uber has shaken up what it takes to get from point A to point B in cities across the country with a simple premise: If you need a ride, a driver nearby could pick you up within minutes.

Behind that idea is an algorithm, which promises to keep supply and demand in constant balance, encouraging drivers toward busy areas and tempering customer requests by increasing the price of each ride. It’s calledsurge pricing.

Those who have used Uber know that surge pricing is a temperamental beast. It changes quickly, varies seemingly unpredictably and has gotten heat fromconsumers, regulators and even drivers themselves. Uber says without surge pricing, the whole premise of a ride in minutes falls apart when there’s a crush of demand.

But how exactly does Uber’s algorithm work? The company doesn’t say. A team of researchers at Northeastern University decided to find out by doing what they call “algorythmic auditing.”

They found that for customers, it pays to be patient — or to walk a few blocks to a less crowded area.

Uber Surge Price? Research Says Walk A Few Blocks, Wait A Few Minutes

Images: Courtesy of Christo Wilson

Euclid and the First Non-Trivial Algorithm

Euclid is this guy:


He is best known for the derivation of all geometric truths using an axiomatic system (referred to as Euclidian Geometry).

He is also credited with writing the first non-trivial algorithm. His algorithm solves the Greatest Common Divisor (GCD) problem*. There had been algorithms developed before Euclid’s algorithm for GCD, such as for adding two numbers, or multiplying two numbers, but Euclid’s stands out. Before going to much into the “non-trivial” part, lets look at the algorithm itself.

This function takes two numbers, a and b, and recursively** finds the greatest common divisor. At first glance, this function doesn’t exactly look like it would work, because it doesn’t look like it is doing much. But this algorithm harnesses the power of the modulo!

“Mod” is an arithmetic operator that takes two operands. You can think of it like division, except instead of keeping the number of times that a goes into b, you keep the remainder. So 8%3=2 and 8%2=0. If a%b returns 0, then b is a divisor of a. So when b == 0, we have found the largest number that divides both a and b.

This is why Euclid’s GCD algorithm is considered the first non-trivial algorithm. It uses recursion, seems to require a formal proof in order to be proven, and can indeed be proven to be correct for any two positive integers.

*GCD problem = the problem of finding the largest number that divides two positive integers

**recursion = when a function calls itself (more here)


Automatic Orchestra

Audio installation put together for Resonance 2015 explores algorithmic and network musical performance using internet connected synthesizer units:

“Automatic Orchestra” is an audio installation orchestrated by networked machines and people. In the months leading up to the festival, a group of students and alumni from University of the Arts Bremen (HfK) and Copenhagen Institute of Interaction Design (CIID) will work with PROTOSEQ synthesizer units, an augmented version of the CHEAP, FAT and OPEN (CFO) audio platform. The group explores algorithmic composition, networked music, and affects of sensor interaction.

For the duration of the festival, an orchestra, comprised of 12 PROTOSEQ units, will be programmed, equipped with sensors, and assembled into a networked organism adapted to the site-specific situation.

More Here and Here

Meshing a heart yields a set of points withing the region. Together with the joining bars they make a graph. This is Euclidean, pretty dense and connected, meaning that every point vertex is (in)directly connected to any other. Inhere exists a curious subgraph which doesn’t disconnect the points and which has the smallest total of edge lengths among all possible connected subgraphs. It’s called the minimum spanning tree.

The Sweet Sound of Algorithms

Do you know your Mozart from your microprocessor? The distance is rapidly narrowing between the abilities of the long-dead master of music and the silicon and circuits of a computer chip.

Back when massive robotic arms lifting heavy assembly line payloads were our only example of automated labor, it was easy to think that machines would never rise to be much more than the brawn to our human brains. But it turns out that even our most prized art forms are nearing the grasp of machines as computer hardware and software becomes more advanced and researchers learn better ways to harness these instruments.  

In one recent example, a Yale computer scientist named Donya Quick has been developing a music-composing program whose output can fool listeners into believing a human made it. One hundred people of varying musical abilities listened to 40 short musical phrases, some produced by humans and others made by computers. The test subjects were then asked to rate each piece on a scale from “absolutely human” to “absolutely computer.” Respondents on average judged Quick’s program, called Kulitta, to fall on the human-sounding side of the scale.  Learn more and hear Kulitta’s music below.

Keep reading


Planning and Navigation for Drone Flight

Research from MIT CSAIL lets drones compute the best way for them to navigate complex obstacles:

Via MIT News:

Getting drones to fly around without hitting things is no small task. Obstacle-detection and motion-planning are two of computer science’s trickiest challenges, because of the complexity involved in creating real-time flight plans that avoid obstacles and handle surprises like wind and weather.

In a pair of projects announced this week, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrated software that allow drones to stop on a dime to make hairpin movements over, under, and around some 26 distinct obstacles in a simulated “forest.”

… “Rather than plan paths based on the number of obstacles in the environment, it’s much more manageable to look at the inverse: the segments of space that are ‘free’ for the drone to travel through,” says recent graduate Benoit Landry ‘14 MNG '15, who was first author on a related paper just accepted to the IEEE International Conference on Robotics and Automation (ICRA). “Using free-space segments is a more ‘glass-half-full’ approach that works far better for drones in small, cluttered spaces.”

More Here

Machine learning plays a part in every stage of your life. If you studied online for the SAT college admission exam, a learning algorithm graded your practice essays. And if you applied to business school and took the GMAT exam recently, one of your essay graders was a learning system. Perhaps when you applied for your job, a learning algorithm picked your résumé from the virtual pile and told your prospective employer: here’s a strong candidate; take a look. Your latest raise may have come courtesy of another learning algorithm. If you’re looking to buy a house, will estimate what each one you’re considering is worth. When you’ve settled on one, you apply for a home loan, and a learning algorithm studies your application and recommends accepting it (or not). Perhaps most important, if you’ve used an online dating service, machine learning may even have helped you find the love of your life.

Algorithms recommend movies, books, dates — even job candidates. In the future, they might cure disease

Autonomous MIT drone avoids obstacles at up to 30 MPH

Everyone is building drones these days, but nobody knows how to get them to stop running into things,” says CSAIL PhD student Andrew Barry, who developed the system as part of his thesis with MIT professor Russ Tedrake. “Sensors like lidar are too heavy to put on small aircraft, and creating maps of the environment in advance isn’t practical. If we want drones that can fly quickly and navigate in the real world, we need better, faster algorithms.”

source @ Roboticstrends