Group items tagged

Google recently launched the Data Liberation Front, an initiative within the company to ensure every one of its products has a clear, easy option for users to export their data in bulk and take their business elsewhere.

love this one ... it seems to take physicist to explain to the AI crowd what they are actually doing ... Deep learning is a broad set of techniques that uses multiple layers of representation to automatically learn relevant features directly from structured data. Recently, such techniques have yielded record-breaking results on a diverse set of difficult machine learning tasks in computer vision, speech recognition, and natural language processing. Despite the enormous success of deep learning, relatively little is understood theoretically about why these techniques are so successful at feature learning and compression. Here, we show that deep learning is intimately related to one of the most important and successful techniques in theoretical physics, the renormalization group (RG). RG is an iterative coarse-graining scheme that allows for the extraction of relevant features (i.e. operators) as a physical system is examined at different length scales. We construct an exact mapping from the variational renormalization group, first introduced by Kadanoff, and deep learning architectures based on Restricted Boltzmann Machines (RBMs). We illustrate these ideas using the nearest-neighbor Ising Model in one and two-dimensions. Our results suggests that deep learning algorithms may be employing a generalized RG-like scheme to learn relevant features from data.

When Australian mobile provider Telstra offered its mobile customers unlimited data for two separate days this year as compensation for network outages, some customers took it as a challenge to download as much as they possibly could in one day. On Sunday, 27-year-old Sydney resident John Szaszvari outdid himself and everyone else by ploughing through almost a whole terabyte of data.

What happens if you strap NSGA-II to optimize the usage of your abstract data structures?

appears like solid data confirming the link between the two; using also EO data from space

Data-Visualization Firm's New Software Autonomously Finds Abstract Connections --> Annalisa?

I had a nice introduction about Ordinal Regression via Manifold Learning by Francisco last week. It is doubtless a very actual research branch!

I doubt :) The original paper from Liu is from 2011 and has .... wait for it .... 1 quotation (and a self-one)!!! http://scholar.google.it/scholar?hl=it&q=Ordinal+Regression+via+Manifold+Learning&btnG=&lr=

Asimov's Foundation meets ACT's Tipping Point Prediction?

still, some interesting thoughts in there ... "Then you have the 50-year cycles of violence. Turchin describes these as the building up and then the release of pressure. Each time, social inequality creeps up over the decades, then reaches a breaking point. Reforms are made, but over time, those reforms are reversed, leading back to a state of increasing social inequality. The graph above shows how regular these spikes are - though there's one missing in the early 19th century, which Turchin attributes to the relative prosperity that characterized the time. He also notes that the severity of the spikes can vary depending on how governments respond to the problem. Turchin says that the United States was in a pre-revolutionary state in the 1910s, but there was a steep drop-off in violence after the 1920s because of the progressive era. The governing class made decisions to reign in corporations and allowed workers to air grievances. These policies reduced the pressure, he says, and prevented revolution. The United Kingdom was also able to avoid revolution through reforms in the 19th century, according to Turchin. But the most common way for these things to resolve themselves is through violence. Turchin takes pains to emphasize that the cycles are not the result of iron-clad rules of history, but of feedback loops - just like in ecology. "In a predator-prey cycle, such as mice and weasels or hares and lynx, the reason why populations go through periodic booms and busts has nothing to do with any external clocks," he writes. "As mice become abundant, weasels breed like crazy and multiply. Then they eat down most of the mice and starve to death themselves, at which point the few surviving mice begin breeding like crazy and the cycle repeats." There are competing theories as well. A group of researchers at the New England Complex Systems Institute - who practice a discipline called econophysics - have built their own model of political violence and

It's not the scientific activity described in the article that is uninteresting, on the contrary! But the way it is described is just a bad joke. Once again the results itself are seemingly not sexy enough and thus something is sold as the big revolution, though it's just the application of the oldest scientific principles in a slightly different way than used before.

One more reason why we should be much more open about all these EO data we have .... The two citizen scientists, Kian Jek and Robert Gagliano, are listed as authors on the scientific paper recently published. I love this: the digital nature of these data make it far, far easier to analyze the science than it was in the past, and also easier to get the data out to people. Because of this, we have an explosive growth in these kinds of projects. Planet Hunters is great, but then so is Galaxy Zoo, Moon Mappers, Ice Hunters, and so many others. You can find several of these collected at the CosmoQuest website.

Wiktor is also collecting these science games in the ACT-wiki. You should have a look at: http://sophia.estec.esa.int/actwiki/index.php/On-line_Games_4_Science

A sonification experiment taking data from ATLAS and translating it into music. The outcome was played at Montreux jazz fest, listen to the results in soundcloud https://soundcloud.com/sonification-quantizer. The way it works is "A tiny subset of collision data from the ATLAS Detector (in CERN, Switzerland) is being generated and streamed in real-time into a sonification engine built atop Python, Pure Data, Ableton, and IceCast." Code's in github https://github.com/cherston/Quantizer_public

I guess this is for data what oil tankers vs. pipelines are for hydrocarbons. Isn't ESA doing something similar for satellite data? I seem to remember that Leopold mentioned that at some point.

Did you know that papers in space science are among the most quoted? Check how your country is doing .... you will be surprised :)

this might also be an indication / point to an issue with their data concerning space science publications ... quite surprising indeed that all Europeans are doing so well in this field

Something should be wrong, for Spain I can read: Economics & Business 4.54 -28 Only minus 28!

http://arxiv.org/abs/1206.6389v2 for Guido; an interesting example of "takeover" research

European research network GÉANT turns Voyager 1 & 2 data into a musical duet.

Scientists in Germany are using artificial nerve cells to classify different types of data. These silicon 'neurons' could recognize handwritten numbers, or distinguish plant species based on their flowers.

Scientists have fabricated a portable data storage system based on nanostructured glass that seems to be much more stable than current disks. I wonder if it can survive that long in space?

Robert Alexander spends parts of his day listening to a soft white noise, similar to water falling on the outside of a house during a rainstorm. Every once in a while, he hears an anomalous sound and marks the corresponding time in the audio file.

I want this guy's job!

A good start for our brainstorm?

Intuitive, well documented, and very powerful. A library to keep an eye on. Check the example Belgian elections, June 13, 2010 - Twitter opinion mining