Showing posts with label AI. Show all posts
Showing posts with label AI. Show all posts

Thursday, 29 November 2018

AI helps in detecting water leackage

Scientists have developed an artificial intelligence (AI) technology that can detect even small leaks in pipes, and could potentially help municipal systems significantly reduce water losses.

The system designed by researchers at the University of Waterloo in Canada combines sophisticated signal processing techniques and AI software to identify telltale signs of leaks carried via sound in water pipes.

The acoustic signatures are recorded by hydrophone sensors that can be easily and inexpensively installed in existing fire hydrants without excavation or taking them out of service.

"This would allow cities to use their resources for maintenance and repairs much more effectively. They could be more proactive as opposed to reactive," said Roya Cody, a PhD candidate at Waterloo.

"Major problems such as burst pipes are revealed by pressure changes, volume fluctuations or water simply bubbling to the surface, but small leaks often go undetected for years," said Cody.

In addition to the economic costs of wasting treated water, chronic leaks can create health hazards, do damage to the foundations of structures and deteriorate over time.

"By catching small leaks early, we can prevent costly, destructive bursts later on," said Cody.

Researchers are now doing field tests with the hydrant sensors after reliably detecting leaks as small as 17 litres a minute in the lab.

They are also working on ways to pinpoint the location of leaks, which would allow municipalities to identify, prioritise and carry out repairs.

"Right now they react to situations by sending workers out when there is flooding or to inspect a particular pipe if it's due to be checked because of its age," Cody said.

The sensor technology works by pre-processing acoustic data using advanced signal processing techniques to highlight components associated with leaks.

That makes it possible for machine learning algorithms to identify leaks by distinguishing their signs from the many other sources of noise in a water distribution system.

Sunday, 11 November 2018

Start Learning AI

Introduction To Artificial Intelligence 

Artificial Intelligence is a way of making a computer, a computer-controlled robot, or a software think intelligently, in the similar manner the intelligent humans think.

Applications  of AI:
AI can be implemented in various fields, such as in a self driving car, Cancer Prediction, Robot as Lecturer  etc.

Two Subsets of AI:

AI subsets
Artificial Intelligence


  1. Machine Learning 
  2. Deep Learning

  1. Machine Learning  :

    Machine Learning is a subset of Artificial Intelligence.
    ML systems learn how to combine input to produce useful predictions on never-before-seen data.
    In Machine Learning , We train a model using a large set of labeled data, later on this model is used to test a set of test data i.e unlabeled data. Then accuracy is calculated that how accurate  the prediction is.

    Suppose that we need an AI to check whether this object is dog or not, then we provide various sets of  sample inputs during the training of our model.
    inputs consist of feature and label
     Dog can be identified by the way they: speak, run, height, eyes, tail etc.
    Here label is DOG and Features are sound, speed, height, and eye.

    When we say labeled data it means that a large data set  that contains Feature and Label.
    So ML moves around mainly three terminologies:
     (i) Label  (ii) Feature (iii) Model

    (i) Label:

    A label is the thing we're predicting—the y variable in simple linear regression. The label could be the future price of wheat, the kind of animal shown in a picture, the meaning of an audio clip, or just about anything.  
     

    (ii) Feature:

    A feature is an input variable—the x variable in simple linear regression. A simple machine learning project might use a single feature, while a more sophisticated machine learning project could use millions of features, specified as:  x1, x2, ...., xn
     Feature could be a property possessed by that thing which is useful to identify that thing uniquely/partially.

    (iii) Model: 

    A model defines the relationship between features and label.

    • Training means creating or learning the model. That is, you show the model labeled examples and enable the model to gradually learn the relationships

      between features and label.
    • Inference means applying the trained model to unlabeled examples. That is, you use the trained model to make useful predictions (y'). 

     Labeled Example:  

    This example is an instance of training data. 
    {x, y} or      {Feature, Label}  

    In order to train our model, we fed a large set of such instances with features and labels. 

    Unlabeled Example:

    This example is an instance of testing data. 
    {x, ?}  or   {Feature, ?}
    Finally we get a trained model that takes the unlabeled test data and make predictions for each test data. 

    "Regression vs. classification"

    A regression model predicts continuous values. For example, regression models make predictions that answer questions like the following:
    • What is the value of a house in California?
    • What is the probability that a user will click on this ad?

    A classification model predicts discrete values. For example, classification models make    predictions that answer questions like the following:
    • Is a given email message spam or not spam?
    • Is this an image of a dog, a cat, or a hamster?   
       
      Model can learn using two approaches: Supervised Learning and Unsupervised Learning.

      " In supervised learning model is being trained under the supervision of input labeled data.But in unsupervised learning no hint is provided during the training of model. Model learns by its own."

       

      "Machine Learning changes the way you think about a problem. The focus shifts from a mathematical science to a natural science, running experiments and using statistics, not logic, to analyse its results." - Peter Norvig - Google Research Director

       

Friday, 9 November 2018

Artificial intelligence technique can boost brain scans to predict Alzheimer early - details inside

The results showed that the algorithm was able to teach itself metabolic patterns that corresponded to Alzheimer's disease.

Artificial intelligence (AI) can help improve the ability of brain imaging techniques to predict Alzheimer's disease early, according to a study. Researchers from the University of California in San Francisco (UCSF) trained a deep learning algorithm on a special imaging technology known as 18-F-fluorodeoxyglucose positron emission tomography (FDG-PET).

They included more than 2,100 FDG-PET brain images from 1,002 patients and on an independent set of 40 imaging exams from 40 patients. The results showed that the algorithm was able to teach itself metabolic patterns that corresponded to Alzheimer's disease.

It also achieved 100 per cent sensitivity at detecting the disease an average of more than six years prior to the final diagnosis. "We were very pleased with the algorithm's performance. It was able to predict every single case that advanced to Alzheimer's disease," said Jae Ho Sohn, from UCSF's Radiology and Biomedical Imaging Department.

"If FDG-PET with AI can predict Alzheimer's disease this early, beta-amyloid plaque and tau protein PET imaging can possibly add another dimension of important predictive power," he added, in the paper detailed in the journal Radiology. While early diagnosis of Alzheimer's is extremely important for the treatment, it has proven to be challenging.

Although the cause behind the progressive brain disorder remains unconfimed yet, various research has linked the disease process to changes in metabolism, as shown by glucose uptake in certain regions of the brain.  These changes can be difficult to recognise.

"If we diagnose Alzheimer's disease when all the symptoms have manifested, the brain volume loss is so significant that it's too late to intervene," Sohn said.  "If we can detect it earlier, that's an opportunity for investigators to potentially find better ways to slow down or even halt the disease process," he noted.

Sohn explained that the algorithm could be a useful tool to complement the work of radiologists - especially in conjunction with other biochemical and imaging tests - in providing an opportunity for early therapeutic intervention.

How to implement Facebook audience network banner ad in android app using Android Studio

Its very simple to implement Facebook audience network banner ad in android app using Android Studio. We will complete it in 4 Steps. 1. ...