Tuesday, December 03, 2013

The rise of phones that read your mind

Google Now, pictured, uses what's called 'predictive analytics' to guess what its users want.
Google Now, pictured, uses what's called 'predictive analytics' to guess what its users want. It combines location information with calendar entries, emails and more to predict what a person will do next, before offering tailored help and support - and all without being asked

Actively searching for directions, weather reports or even what to have for dinner is so last year. 
The future lies in apps that can read your mind and offer up this information automatically, with Google leading the way with its Now app.
Technically Google Now doesn’t read a user’s mind, but it is capable of predicting what they will do next, before offering relevant help and information at each step without ever being 

Predictive analysis isn’t a new concept; financial services have used it for years to work out a person’s credit score and how likely they are pay back a loan, for example.
 It does this by using what’s called ‘predictive analysis’. 
Yet Google Now, along with other intelligent personal assistant apps including Osito and Grokr, uses this technology to predict how everyday smartphone owners are going to behave.

    The software is based on an algorithm that combines various snippets of information to form a prediction. 
    In the case of these personal-assistant apps, the software scans calendar entries to work out where a user is, or should be. 
    It combines this with location data, taken from the phone or tablet’s GPS unit, as well as posts on social networks, email information and more.
    It will then present the user with Cards designed to offer support or help it thinks they need. 
    For example, if a calendar on an Android phone contains a diary invite, Google Now will create a reminder. 
    It will then check to see whether the user is en route to the meeting and add directions to this reminder, or work out how long it will take them to get there based on current traffic. 

    The prediction software, used in apps such as Osito pictured, scans calendar entries to work out where a user is, or should be.
    Other prediction apps including Osito, pictured, can show bus timetables as as users pass a bus stop, for example. It can show the latest deals as they enter a supermarket, or open a digital boarding pass as they approach an airport
    A Card could also be created to show what the weather is like where they are going and even advise taking an umbrella. 
    Elsewhere, as users pass bus stops, prediction apps can show bus timetables, or show the latest deals as they enter a supermarket. The latest updates to Google Now are an extension of changes the tech giant recently made to its search algorithm. 
    Now when you search Google it no longer just searches for individual keywords; it can now respond to questions and statements in the same way a human might, by looking at the whole query - even if it doesn't have an exact answer. 
    For example, asking Google 'Tell me about impressionist artists' will show a list of artists but also now let you click to learn about the impressionist movement, browse individual artwork, or switch to abstract artists using filters.
    Called Hummingbird, the company claims it’s 'the most significant algorithm change' in years and will affect 90 per cent of searches. 
    Users can now ask Google to compile playlists, compare items or find out facts about landmarks and the search giant will create context for each one. 
    For example, if searching for information about Big Ben, the searcher only needs to say 'Big Ben' once; Google will know that if someone asks for trivia about the landmark, before asking for directions, that both questions relate to the same thing.
    When used with Google Now, a card will show these directions automatically as soon as they search for the landmark, as long as they are within relatively close distance. 
    Back in 2012, researchers from Birmingham University created software that could successfully predict a person’s future locations with an error margin of just 60ft.

    A screenshot taken from the Far Out mapping software. A pair of researchers from Microsoft are able to predict where a person will be years from now using this system.
    A screenshot from the Far Out mapping software. Researchers from Microsoft and Google can predict where a person will be years from now using this system. The program tracks a person using GPS and learns their routine. It then uses this information to accurately plot their future locations
    While mobile phone networks can already track where a handset is in 'real time', the scientists developed the algorithm to forecast our future movements.
    They compared data from one individual and their closest social network to predict a person’s future location based on places and areas visited in the past and the frequency of contact between those studied. 
    Earlier this year, researchers from Microsoft and Google similarly developed software called Far Out that can learn a person’s routine by tracking them using GPS.  
    It then uses this information to accurately guess their future locations and will adapts its predictions even if someone changes their job, relationship or moves house.

    No comments:

    Post a Comment

    Related Posts Plugin for WordPress, Blogger...