Singularity Pulse  
     
  Legal

Machine Learning Based Credit Scoring System

Equifax launched its patent-pending machine learning technology “NeuroDecision Technology” (NDT) which models credit scoring using advanced neural network modeling technique. NDT is the first regulatory-compliant machine learning credit scoring system. Currently the technology is deployed in products for customers in commercial and consumer risk models. NDT is also used in Commercial Insight Delinquency Score calculation. Generated scores are used to predict likelihood of a business incurring severe delinquency, charge-off or bankruptcy on financial accounts within the next 12 months.

 
  Image  
 

AI to Catch Liars While Dispersing Loans

Ping An, a Chinese insurance company, has created an app that allows its customers to apply for loans directly. Prospective borrowers are required to answer series of questions regarding their income and repayment schedules. AI powered app then scans approximately 50 tiny facial expressions to determine whether a customer is lying or not. The company is expecting that these current algorithm and ML powered app will help bank reduce its exposure to bad loans as well as its default rate.

Finance  
  Image  
  Space Research

McAfee Uses Machine Learning in Enterprise Security Solutions

McAfee's updated Enterprise Security Manager has added various ML powered solutions like “Behavioral Analytics” and “Advanced Threat Defense”. McAfee Behavioral Analytics uses machine learning to discover new and unusual high-risk security threats without requiring extensive configuration. Out of billions of security threats, the algorithm cuts it down to just few hundreds of anomalies to investigate. Company is hopeful that the solution will streamline workflow and improve detection rates to reduce response time on raised ticket.

 
  Image  
 

Facial Recognition to Impose Fines Via Text Message

Intellifusion, a Chinese AI firm is working with local police and installing AI powered cameras in city of Shenzhen with the aim of identifying jaywalkers and penalizing them with fines. With the use of facial recognition technique, fined individual will be identified and accordingly be informed about it via text message through a widely used system such as WeChat or Sina Weibo. The cameras, linked into a database of people, catch pedestrians crossing the road when traffic lights signal they should not be.

Healthcare  
  Image  
  Technology

Google Planning to Put AI In-Charge of Food Production

Alphabet’s secret X lab is exploring ways in which AI could dramatically improve food production. They are trying to bring together machine learning, drones and robotics to get themselves enabled in using high tech farming practices. One of the critical parameter for the company is decision-making regarding “harvesting time” and the “amount of water required for irrigation”. The technology will also help farmers with crops that are being grown in environments where sudden climate change affects the accuracy of weather forecast.

 
  Image  
 

AI to Enter in Aerospace Manufacturing

According to the latest annual edition of Accenture’s Technology Vision report, more than 80% of leading executives in aerospace and defense companies expect to see AI systems working alongside their human employees in next few years. Accenture is already using AI systems in forecasting decisions. Another leading aircraft Original Equipment Manufacturer (OEM) was using AI to connect distant engineers and factory managers via augmented reality, cutting down on travel time and costs between the sites.

General 
 
  Image  
  General

AI Cardiologist Outperforms Human Counterpart

Rima Arnaout, assistant professor and practicing cardiologist at UC San Francisco has created an algorithm to analyze heart scans. It is done by evaluating an echocardiogram (image produced by bouncing sound waves off the heart). Her team used convolutional neural network to train an AI system that can classify echocardiograms according to the type of view shown. Cardiologist would look at high-resolution video of the echocardiogram, while AI was given still images taken from video clips to learn. AI achieved an accuracy of 92% compared to human accuracy of 79% in sorting images into 15 categories of views.

 
  Image  
 

AI to Invent New Medical Tests

John J. Dillon and Paul A. Friedman of Harvard University have discovered new way of detecting fluctuations of blood potassium levels with the help of AI. This method does not require patients to draw blood. Massive datasets from Mayo Clinic were used to generate an automated computer algorithm for the new testing method. Method combines smartphone technology with algorithmic analysis of ECG recordings. Algorithm detects subtle changes in the morphology of the ECG waveform to determine serum potassium levels in real time.

Neuroscience
 
  Image  
  General

AI to Predict Biological Age Based on Wearables Data

According to findings published in the Journal of Scientific Reports, AI can predict ageing by gathering physical activity data from smartphones and other wearables. Researchers analyzed physical activity records from 2003-2006 US National Health and Nutrition Examination Survey (NHANES) and trained neural network to predict biological age and mortality risk of the participants. Convolution Neural Network was used to understand motion patterns and establish their relation to general health and recorded lifespan.

 
  Image  
 

AI to Automate Digital Effects for Movies and Games

A start-up called “Arraiy” is developing an algorithm on computer-generated imagery which can automatically create digital effects for movies and games. Conventional way of building digital effects is a tedious process. Using neural networks, algorithms can learn tasks by analyzing vast amounts of data. They can edit unclear noise sections or images. After filming in front of the conventional “green screen,” engineers can rotoscope thousands of images to add to the data collection. Once the algorithm is trained, it can rotoscope images without help from a green screen, thus reducing production time.

Neuroscience
 
  Image  
  General

Machine Learning Algorithms Are the New Match-Maker

eHarmony has been using human psychology and relationship research to recommend partners for singles looking for a meaningful relationship. Now it is using data-driven technology backed with big-data, cloud computing and ML algorithm to improve recommendations it makes to its users. Prateek Jain, head of technology, explains how matchmaking service calculate 'affinity', a process of generating behavioral data using machine learning (ML) models. It runs 20 affinity models to improve matches. It is also using ML in increasing match satisfaction across the user base.

 
  Image  
 
 
 
     
  Banner  
 
Contact us:

5A, B-Wing, Trade Star Building, J. B. Nagar, Andheri-Kurla Road, Andheri (East), Mumbai - 400 059, Maharashtra, India
W: www.decimalpointanalytics.com  
E: info@decimalpointanalytics.com   T: +91 22300 15200