Singularity Pulse  

Bank of America Launches AI Driven Personal Assistance

Bank of America is launching AI powered virtual assistant for its financial services. It is named “Erica”. The personal assistance will be available to its 25 million mobile customers. Erica providing new and interactive ways to clients to do their banking using voice, text or gesture. It is also used in providing personalized solutions to customers on improving credit score or managing their budget. Past transactions, navigation on the app, locating nearest ATM, schedule meetings with bankers, view bills and schedule payments etcetera can be handled by the algorithm driven assistant.


An AI Technology to See Human Bodies Through Walls

Computer Science and Artificial Intelligence Laboratory at MIT has created a system that can see human bodies through walls. It can recreate poses of walking, sitting, or simply standing still. It uses RF waves to sense where person is and then recreates the body as a stick figure called “RF-Pose”. The technology uses neural network to analyze radio signals that bounce off people’s bodies. This system is primarily intended for passive monitoring of a subject inside a room without cameras or other intrusions. It could also help elderly people live more independently, while providing the added security of monitoring for falls, injuries and changes in activity patterns.

Information Technology  

AI Powered Index to Track ESG Activities

Solactive launched its “Solactive Sustainability Index Europe”, which is designed to track performance of environmentally and socially responsible European companies. It is based on data driven by “Arabesque S-Ray” platform. The platform uses ML and algorithm-based technology to rate companies based on principles of United Nations Global Compact: Human Rights, Labor Rights, the Environment and Anti-Corruption. The technology combines 500+ ESG metrics with new signals from over 50000+ sources. Screening like liquidity and dividend-yield filters are applied later on to get final index composition. It also weighs the components based on volatility optimization. The index is priced in Euro.


Brain-Based Circuitry AI to Be Much Faster Than Conventional AI

Researchers at IBM have built a physical neural network which resembles neurons present in human brain. When this new AI was tested on new type of circuitry, it performed better at par with conventional neural nets already being used. But the new integrated neural net system completed the task with 100 times less energy than its conventional AI algorithmic counterpart. The new research modeled two types of neurons: one for quick computations and the other for storing long-term memory.

Information Technology  

Fixed Income ESG Can Improve Risk Management, Returns

Recent research suggests that incorporating ESG factors into fixed income investing can strengthen risk management. Additionally, it should provide more stable financial returns. The report was published by World Bank and Japan’s Government Pension Investment Fund. As per the report, approximately 90% of studies showed a nonnegative correlation between ESG and corporate financial performance. Additionally, a Barclays study was included which found that higher ESG ratings translates into small but steady performance advantage. The report was focused on private sector and made recommendation of broadening ESG investing across fixed-income asset portfolios.


AI Technology to Understand Concept Of 3D Space

Computer vision systems don’t have the understanding of space. But Deep Mind’s latest research published in journal Science describes a neural network, knowing practically nothing, which can look at one or two static 2D images of a scene and reconstruct a reasonably accurate 3D representation of it. Unlike most computer vision algorithms which work on supervised learning, it works entirely independently of any ideas of how to see objects. One half of the system is called “representation” part which observes scenes and encode it into vectors. The other half, called “generative” part, predicts what different part of scene would look like.


A Software to See Several Minutes into The Future

Scientists from University of Bonn in Germany have developed a computer process that can accurately anticipate human behavior minutes ahead. The software was trained to predict what a chef would do next while preparing a meal. It was earlier trained by showing it a number of videos of people making breakfast or a salad. Then, it was shown completely new video of another person preparing a similar meal. Computer predictions were compared against the actual activities of the video. The accuracy was over 40% for short forecasting periods.


ML to Aid Inspections in Auto Auction Apps

TradeRev-H, an app for one-hour online auctions, is utilizing machine learning technology to change the way how vehicle inspection is done. Traditionally, third-party inspectors are required for creating a visual condition report. TradeRev-H aims to change that. It takes video from vehicle walkaround and then automatically creates still photos showing parts of the vehicle. This gets uploaded in standard condition-report order and format. Additionally, if any scratch or dent is large enough to matter, the video will zoom in for a closer look. Cost of damage is also calculated from database of damage costs.

  Dairy Farming

AI Backed Commercial Solutions for Dairy Farmers for Milk Testing

SomaDetect Inc. of Fredericton is preparing to deliver commercial systems which will test milk and use AI to provide insights to maximize farmers’ profitability and management. They have developed what is termed as “Internet of Cows” with sensors installed at each milking stall. It identifies each cow, test their milk and provide farmers with metrics such as protein and fat counts. It also provides with indicators of disease, hormones that manage reproduction and antibiotic residuals.


Facebook AI System Aims To 'Open' Closed Eyes in Photo

Facebook researchers are working on an AI system that can make closed eyes look convincingly open. It presented a Generative Adversarial Network (GAN), a ML tool that tries to fool itself into thinking that its creations are real. After getting trained on data of people with their eyes open, GAN learns what eyes should go on the person on the basis of shape, colour and other features of the eyes of a particular person. When tested, people were unable to distinguish between fake eyes-opened photos and the real ones.

  Civil Engineering

A Smart Power Outlet to Prevent Accidental Household Fires

MIT’s Department of Mechanical Engineering has developed a “smart power outlet”, which analyzes power drawn by appliances and can distinguish between harmless electrical arcs and dangerous ones. By reading a device’s unique electrical pattern, it can identify which device is accidentally left plugged on. Researchers fed the neural network power signature data from four different devices: a fan, an iMac, a stovetop burner, and an air purifier. After getting trained, it was able to identify dangerous electrical arcs with 99.95% accuracy and differentiate between the four devices with 95.61% accuracy.

Contact us:

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