It is increasingly common to read or hear about machine learning or machine learning, a branch of artificial intelligence that basically make machines “learn” alone.
Advances in this discipline have been coupled with the use of huge amounts of data to train these engines and achieve surprising results in many cases. Are there limitations to what automatic learning can achieve? Of course, but all these examples show that there are areas in which this discipline could change our world forever.
The fight against suicide
A team from Cincinnati Children’s Hospital is working on a system that after interviewing a number of people trying to figure out if they were people with some risk of suicide.
The analysis not only took into account the answers, but also other sections such as the intonation or the harmony of the answers, something that according to the experts could contribute much information to the system. According to the study, people were identified with suicide risk with 93% accuracy.
How to Speak English (Non-native)
The speakers of the same language tend to be without problems, but there are situations in which those who have learned to natively and those who have done as a secondary language have problems communicating.
Precisely in this area, it is where the MIT created a large base of data in non-native English phrases. The goal was to improve the processing of language machines. Technology language processing (NLP) can work something curious: a machine does not know how to process the particularities of non-native English.
Infallible medical diagnoses (or not)
One area in which it was expected that these systems achieve significantly impact is medicine, and specifically the medical diagnosis in which machine learning theory seemed very capable of giving valid responses.
The self-diagnostic tests conducted by Harvard University showed that automatic systems work with objective data (radiographs, CT scans, clinical analysis) but not with symptoms offered directly by patients.
The problem is that doctors are able to detect and extract those symptoms from patients information (“everyone lies,” said Dr. House) to effectively diagnose what happens to them. A machine will still have time to solve this problem, it seems.
Image Search: Google Image
Cessation and censored contents in social networks
The overwhelming success of Internet video on YouTube and Facebook Video and other services has enabled both live broadcasts and recorded become a content type appreciated by users, but not all content follows policies use of these services on ethical or moral aspects.
That is why they have developed a system that allows on Facebook detect violent content or nudity during broadcasts live and live their service offering Facebook Live, and automatically blocks them.
In search of habitable exoplanets
A team from the University of Toronto has been inspired by the recommendation algorithms Netflix, which also make use of machine learning, to try to accelerate the search for dynamically stable planetary systems.
The idea is to achieve accelerate the search for exoplanets “up to a thousand times faster than with traditional methods,” say the investigators. In fact it is considering using the algorithm in the mission Transiting Exoplanet Survey Satellite NASA that starts next year.
Instant Object Recognition
One of the areas where automatic learning has demonstrated its validity is that of the recognition systems images. Several are the services that use these systems to organize our photo libraries, but the idea is also applicable to the recognition of objects and people in real time.
It is what makes RadarCat, a kit is based on the Project Soli of Google and allows recognize, classify and distinguish different materials and objects in real time and with high precision.
The secret are the electromagnetic waves and of course, your machine learning algorithm, which is powered by a database in which each object leaves a unique “fingerprint.”
Reading the mind seems feasible
The relationship of machine learning with areas such as neurology and psychology seems unlikely, but a group of US researchers has developed a system approaching very own idea of fiction films: the power read your mind.
Those responsible for this development presented hundreds of faces to 23 people who underwent a functional magnetic resonance, and then repeated the process but offering the algorithm only the resonance data for the algorithm to try to link that resonance with the face they were looking at Participants of the experiment. The result was unique, but not definitive. Perhaps one that directs us effectively to an infallible mind reader.
From the stand-alone car to the stand-alone wheelchair
Autonomous cars are certainly one of the fields in which artificial intelligence seems to have a more significant short – term paper, but the foundations of that machine learning applied in these vehicles also apply to wheelchairs used For example in hospitals.
In Singapore, they are trying to create precisely those chairs autonomous wheel to free up staff to move patients from the hospital so that they can occupy other various tasks. Nissan, incidentally, has a similar project.
This discipline is also starting to become part of our smartphones, those devices that will become more and more ambitious in that objective to simplify the life to users ahead of their needs. Google Assistant, a tool that debuted in Google Pixel, makes use of such algorithms.
Google is not alone in taking advantage of auto-learning. Huawei recently introduced its Mate 9, a smartphone whose software learns the user to improve the user experience. It recycles and classifies applications to manage mobile resources (memory, storage, processing capacity) in the most efficient way possible, and this is surely only the beginning of the application of this type of discipline in our smartphones.
You may also like to read another article on Web2GB: Facebook, Amazon, Google, IBM and Microsoft join forces to monitor the development of artificial intelligence
Beyond facial recognition and voice recognition
The applications of facial recognition technology are increasingly curious. If in principle had been raised as a useful and convenient way to authenticate (Windows 10 and some devices and allow it through Windows Hello) machine learning has enabled algorithms that identify faces are keys to sort photos online and make such as precise searches Google Photos (with controversy included) or Facebook Moments enable their users.
This capacity could be applied, however, to much more sensitive subjects such as the identification of suspects of a crime, but not by facial recognition, but by voice recognition. There is much ground covered in this sector, and researchers of various technology companies already have systems and APIs that allow reaching identify a person by his voice even when this is distorted.
Goodbye, Manolo Lama, we do not need you
The sports commentators could also be affected their work by the arrival of the artificial intelligence. Machine learning algorithms are able to recognize what is happening in various sports and add comments to video being broadcast with little delay.
As in any system of this type, the secret lies in the training of these solutions, which make use of hours and hours of small plays of several sports and then offer those aids. In fact that’s the key word, because the commentators should not be threatened: these systems are geared to help them in their work such as analyzing the plays from other angles quickly.
Caution, Conductor Friend
Autonomous cars are never more appropriately around the corner, but maybe before seeing these vehicles on the road we meet based on machine learning systems are intended to prevent our next miscue behind the wheel.
In this type of application, a camera analyzes our body language and our gestures to try to understand what reactions we have before different maneuvers. According to tests conducted by researchers from Cornell and Stanford, it was possible to predict such changes lanes with 90% accuracy, and seem very interesting applications in the field of freight transport, for example.
Machines that speak like us
Another clear segment in which machine learning can bring clear advantages is its ability to help the communication between humans and machines.
Train an artificial intelligence system with all kinds dialogue results in an algorithm able to maintain a relatively coherent conversation, which showed in Google from talks classic movies. This type of application will be more and more striking, especially now that all the big ones seem to want to put their particular version of the Amazon Echo in our salon.
Now in the bill of light (and in many others, hopefully)
The good thing about this machine learning that is conquering all kinds of scenarios is that it has enormous potential to optimize a bunch of processes and decisions. Some, by the way, as useful as getting us to save money on our bill.
Google has managed to do in their data centers, and who has advised improvements in the way people consume more energy efficient in these centers has been an algorithm developed by Jim Gao, a developer of Google. THE system learned the behavior of each of the servers to adjust the PUE (Power Usage Effectiveness) in each case and make eating out just right.
Musical recommendations that will ring you to glory
Maybe Google is being involuntary protagonist of this article, but as we said one of his most responsible in this area that “no department in Google that does not benefit from machine learning”. Practical applications to users are many and varied, but the most recent practices and is affecting the Google Play Music, its music streaming service.
This new system uses machine-learning systems to discover the kind of music we like and mix it with data as our location or activity and even the state of meteorology to create a custom playlist. These recommendations go beyond what other systems are trying to do based on what we like or listen to frequently, and the analysis of those additional factors (you cannot activate this option if you prefer to protect your privacy) can be a plus for many users.
Social networks with less spam and less false news
The problem with most popular social networks is that their own philosophy causes a lot of false content that many users can mistake for legitimate content. The spam and the recent phenomenon of false news have shown that here is a way to go, but that machine-learning algorithms also provide solutions in this regard.
It is what they think at least on Twitter, where their BotMaker system has been a couple of years trying to fight against spam in this microblogging network. Their leaders then emphasized that the system had been able to reduce by 40% the spam posted on Twitter, and this recent efforts by Facebook are joined by dealing with the scandal of fake news. Mark Zuckerberg company will use machine learning systems to detect this type of news, although there will be more resources to prevent problems such as those that have recently given.
Tremble, Beethoven: come the machines that make up classical music
One might believe that the arts are safe artificial intelligence yet. The intuition and the kind of factors that come into play in the creation of any piece of art are something that for now did not seem within the reach of automatic learning systems, but some are challenging that idea.
DeepBach is a system that after learning the way they were composed cantatas of Bach is now able to create pieces that mimic these compositions and even take existing songs to harmonize them under that same style. It is not the only project in this regard, and recently we interviewed Francisco Vico to discuss Iamus a computer cluster capable of composing contemporary music, which already has been performed by the London Philharmonic.
More and better medicines
Among those practical applications that speak of a better future is the research in the pharmacological field: the discovery of new drugs could be much more efficient and faster thanks to the use of artificial intelligence systems and, in particular, machine learning.
A recent study allowed for example minimize and even avoid the side effects of a drug before conducting clinical trials in humans, which significantly reduces the cost of this type of process. As expected, Google is also investigating in this area in collaboration with Pande Lab, a laboratory at Stanford University, and there is a good set of similar initiatives, which will hopefully help address before and better all kinds of diseases.