Sunday, January 31, 2016

Machines algorithms

The Internet of Things will generate 44 terabytes in 2020 but does it make sense to store all the information? During how much time? How you can harness all that information generated?
More than 600 experts in technology and analytical data from around the world representing companies like Linkedin, Facebook, Google, Stratio, MongoDB or Zalando, among others, will meet for the fourth edition of Big Data Spain (www.bigdataspain.org ) (# BDS15) held in Kinepolis Madrid in the Ciudad de la Imagen in Pozuelo de Alarcon on 15 and 16 October, on how Big Data is impacting people's lives and productivity of companies.
Machine learning
Mobile phones, banking, or social networks have become not only the new forms of communication, but a huge source of information in which we reflect our daily life. Technologies, and offer good services to end users, also provide powerful sources of information companies. The study presented Alejandro Llorente, co-founder of Piperlab analyzes on Twitter behavioral characteristics of more than 145 million messages pinpointed distributed among more than 340 different economic regions in Spain, to infer the level of unemployment in a particular region or to predict whether users are in a better position to attract better opportunities and position information.


Similarly, analysis of the emails that circulate daily by a company can help improve employee happiness, as shown Miguel Romero, head of HP Hadoop architecture, and Alberto Santos HP scientific data. Emails circulating organizations are an interesting asset, not only for its content but by other functional details that respond to the following questions: Which communities in my company? Where they do occur? What are they talking about? Do you do activities together? Do they treat customers well? Knowing the answer to these questions can provide activities or actions that enhance the well-being at work.

By 2020 more than 44 zettabytes of information will be generated, according to IDC estimates. The rise of so-called Internet of Things will present data analysts are unable to analyze all the information that is received in real time infinity of connected devices. As explained Marco Laucelli, CEO of Novelti, the need to assess whether it makes sense to store all the information, how long is needed to store and, most importantly, how all that information generated pays off. And it challenges as the information received come from different sensors, which do not always send the same data type, which may occur disconnections in the transmission of data, structured data and unstructured combined add up .. .
Machines algorithms

In this context, there is growing interest in developing machines and self-learning algorithms using artificial intelligence based systems. Always associated with super robots, the truth is that today few of us interact with bipedal robots or give a ride in a car without a driver, and yet, in our daily lives are increasingly present artificial intelligence systems that recognize speech, images or analyze patterns in real time to detect credit card fraud or customize ads showing a web. Precisely Amparo Alonso, president of the Spanish Association for Artificial Intelligence (AEPIA), working with machines and self-learning algorithms and their application in various fields such as predictive maintenance engineering or prediction of gene expression in bioinformatics.

And in the world of health, Big Data, statistical analysis, self-learning models and visualization can bring benefits that translate into improved quality of life for patients, cost rationalization and better use of resources in the public health system and expose Dr. Javier Gómez Pavón, scientific data of the Central Hospital of the Red Cross.

In short, Big Data Spain (# BDS15) show how instant intelligence can calm the impatience of an increasingly doomed to the need to make decisions in real time economy. Other prominent speakers of Big Data Spain (# BDS15) are: Paco Nathan, consultant O'Reilly, with his vision of Big Data in 2016; Jason Sundram (Facebook), Kartik Paramasivam (LinkedIn) and William Vampenepe (Google) featuring in the first person as their companies face the challenge of managing so many concurrent real-time data; Jim Webber Neo Technology to teach interpret graphical database in fraud detection; or Roland Zalando Vollgraf of employing Bayesian statistics to save logistics costs.

Thursday, January 28, 2016

Passport national identity

Passport, national identity card (DNI), fingerprints and iris are several conventional biometrics used to identify airports and other border crossings. However, in the future a new technology 'made in Spain' could allow identifying an individual by his body odor, thanks to the investigations of the Group of Biometrics and Security Biosignals (GB2S) in Integrated Automation Center of the Polytechnic University Madrid (UPM).

After analyzing samples -8 thirteen men and 5 women, scientists found that there are consistent patterns in an individual odor. Although the error rate is high, 15%, compared to more reliable as taking a fingerprint or iris, Gonzalo del Pozo Dancer analysis coordinator assures the world that "a refinement in the sensors and algorithms could serve as a new method to identify individuals ".

But how the smell of a person is captured? Gonzalo explains that "there is a sensor in the hot air pipe that extracts the smell of hand. Because it has a mass spectrometer [a machine to analyze the chemical composition of various elements], the analysis is based on identifying the amount of each compound and the relationship between them. "

The odor samples were collected on different days and at different hours-28 meetings in total as "body odor can be affected by many different factors such as nutrition, metabolism of the person, their mood and even some diseases ", explains the researcher.
National
The tube hot air in the sensor extracts the smell of the hand | UPM
Dancer mind that people did not perform any physical exercise plus all previously had to wash their hands to prevent signals from perfumes or creams. "The effect of this type of cosmetics will be discussed in future revisions of the project," he says.

National identification number
A noninvasive technique

The authors of the research, which was born three years ago, say thanks to improved sensors have been able to obtain more reliable data. "This is the last experiment has been carried out after refining the sensor. Other previous campaigns gave problems that occurred in the measurement method. Sometimes instead of measuring body odor picked small samples of air in the room, thereby losing reliability, "acknowledges Gonzalo.

After analyzing the results, the success rate was 85%, while the error of 15%. The scientist found that this figure is promising because if advancements and improvements occur in the sensors, this technique would not be invasive identification. People only have to go through a booth at the airport -instalada or other positions fronterizo- control in which sensors would be integrated.

Several years before joining

However, implementation of this technology still take several years. In fact, for Cabadillo Arturo Alvaro, one of the members of the company SEADM SM has participated in improving the sensor, "it is frankly difficult to implement a new system in a normalized system such as control of airport security. The equipment used usually have to go through a lengthy certification process for the end user (the police control in this case) can access if you want to purchase one of them. "

In addition, the expert notes that the ethical issues, such as invasion of privacy, you may also hinder the implementation of such systems: "In this type of project, I would say that technology is not the limiting factor. If devote the necessary resources, needed to implement a system of olfactory recognition technology could be ready within a few months or perhaps years, but a system of this type takes in the best several years to be used. "

USS Zumwalt

In naval warfare the key factor has always been the distance: the sea is the only theater of war in which it is almost always possible to refuse to fight and take refuge in the immensity, and if there fighting the distance between opponents is essential. The speed and maneuverability of the Greek triremes and the ability to approach the ships from the Spanish Armada Rome sixteenth century defined the battle at sea for millennia.

Later firearms, especially guns, imposed the struggle to ever greater distances (tens of kilometers in the battleships of the two world wars) until the early twentieth century the carrier extended to hundreds of kilometers the space between the warring fleets, which no longer came to see. The missile, cruise or ballistic hypersonic has reached its highest level of development in the early XXI. But the future of war at sea is electric. Right now there is a race for the development of electromagnetic pulse technology that will change from designing warships mode against each other.
USS Zumwalt
The two main technologies involved are the railgun (or 'railgun') and the mass driver. Its design and use are different, but they have two things in common: they need huge amounts of electricity to operate and can revolutionize naval artillery and aircraft of the future.

The brand new and so far only member of its class, the USS Zumwalt, just made its debut with a video of her sea trials published by the US Navy. With its galactic appearance, protracted development and its high cost, the new boat (classified as a destroyer, though size and function would be more of a light cruiser) has attracted much attention. But there is an interesting detail: the new ship artillery batteries comprises two cannons missiles and high-tech AGS (Advanced Gun System) type.

The AGS have a caliber of 155 mm, as much of the artillery American camp, but are very different: you can not use the same ammunition, are completely automatic in operation and are designed so that they can launch a volume of fire equivalent to a full battery of six shells of the same caliber campaign. They are capable of firing long distance (LRLAP) to more than 100 kilometers away, and its mission is to provide artillery support to forces much further from the coast than was possible until now land.

But the AGS was not to be, originally, the main weapon of the Zumwalt class, which was designed to be equipped with guns that were to be electrical unlike conventional powered explosives. These boats are designed to install on them railguns, but the development of this technology has been delayed and costs have increased so much that they were replaced by the AGS, for now.