Singularity Pulse  

Machine Learning, And Predictive Science to Provide Real Time Borrowing and Lending

Hyderabad based start-up, AnyTimeLoan is an on-demand peer-to-peer lending platform. It uses advanced data science tools and technologies from Machine Learning to ensure real-time, 24/7 dispersion of loans. ATL offers personal or consumer loans, educational loans and MSME business loans. When an eligible borrower applies for a loan, LOBOT (chat bot) automatically maps the inquiry with lenders. In turn, lenders rely on the evaluation outcome and accordingly initiate loan disbursement within five minutes to the respective borrower’s bank account. It ensures better asset classes for its lenders as well. Currently, the start-up has raised $4 million in Series A funding from Spice Digital. It has dispersed over 43,000 loans, aggregating over Rs 80 crore with gross non-performing asset ratio of less than 0.63 percent.



AI Powered Chatbot Speeds Up Finance Admin

University of Sydney has developed a chatbot which assists its finance department in completing administrative tasks, hence improving the efficiency of the department. The university partnered with Ignia to develop the chatbox using Microsoft cognitive services. It assesses university's financial system to query data and communicate appropriate purchase orders, invoices, and supplier-related queries, if any. For better user experience, adequate attention is given to details like chatbot’s tone, personality, and user interface. After successfully completing its pilot project, university is now targeting to deploy it to other areas as well, where it can provide similar help to reduce administrative back logs.



Algorithm to Count Number of People Present in Demonstrations

Computer Vision Centre presented a new technology, based on artificial intelligence, that could accurately count the number of people that appear in still or video images in big gathering or demonstration with 80-90% accuracy. It was presented at the Mobile World Congress by the company. Project leader Joost Van De Weijer explained that problem with correctly estimating number of people in a demonstration is that they will not all be in a single image. Consequently, algorithm is trained on a learning method that trains vision algorithms to compare images. It checks that number of people in the small or cropped image is always less than that in the large image.



AI to Make GDPR Compliance Better

IBM has launched its governance portfolio titled as “IBM Information Server”. It makes use of machine learning technology to perform data analytics and compliance. Madhu Kochar, VP analytics product development at IBM, explains that instead of employees cleaning the data machine learning models will classify it quickly and efficiently. Regulations like General Data Protection Regulation (GDPR) will come into effect from May 2018. New regulations can put a drag on speedy processes of data-gathering. The new technology will minimize cases of GDPR slips. The algorithm will automatically tag data as personal, sensitive or whether data is needed for compliance.

Information & Technology  

  Chemical Engineering

Machine Learning to Assist Doctors in Diagnosing Heart Attacks

Traditional detection of heart attach includes electrocardiograms (ECGs). But they only include a small portion of the entire electrical signal of the heart. Each person has a unique resting ECG signal. Multiple variations in measurement make it very difficult for automated systems to predict heart conditions. To overcome this, researchers fed machine learning algorithm with multiple sets of categorized data with different features. It then detects changes in the cardiac signal indicative of a heart attack. It isolates the electrical signals from the heart and examines changes before, during, and after simulated heart attacks. The two categories the algorithm isolates are “having a heart attack” and “not having a heart attack.” It was 32% more accurate at detecting the early signs of a heart attack.



Google’s AI Is Being Used by US Military Drone Program

US military is using Google’s artificial intelligence technologies for its drone projects. TensorFlow AI systems are being used for Project Maven. Large number of images shot by multiple drones are being analyzed using AI algorithms. It then detects objects of interest and flag them for review. The algorithm provides open source TensorFlow APIs that can assist in object recognition on unclassified data. It is estimated that this will increase an analyst’s capacity by thrice the original limit. Project Maven is part of US Department of Defense’s (DoD) data processing program; which has seen $7.4B already been spent on it. Reportedly, Project Maven has already been put to work against Islamic State.

Information & Technology  


AI to Provide Recruitment Solutions to HR Firms

Midas IT, a South Korean company, has launched an enterprise solution called “inAIR” that allows clients to use AI to conduct recruitment interviews. It automates the application process and uses an AI bot “MIGHTY” to conduct interviews of applicants. It asks questions along with conducting games for applicants. Through its facial recognition technique, which analyses 68 muscles in the face it determines veracity of the responses. Voice tone and verbal patterns are also analyzed for better diligence. Later, it can also make suggestions on which department is best suited for the candidate. The company claims that it showed accuracy of 82 percent to find high-performing talent amongst all those it recruited.



Astronauts to Be Guided by A Floating AI Brain

IBM and Airbus has developed CIMON (Crew Interactive Mobile Companion), a 3D-printed AI assistant or DLR Space Administration. Its entire structure is made of 3D printing. It uses IBM’s Watson AI technology as its processing unit. The AI powered computer brain can alert astronauts with technical glitches. Astronaut Alexander Gerst has been testing it out via various experiments. CIMON has given customized voice and an onscreen face based on Gerst’s input. Multiple experiments are due to be conducted in microgravity environment for further enhancements. In its first mission, CIMON will be deployed on a probation basis with only limited capabilities switched on.



AI Predicted Oscar Winners with Over 90% Accuracy

Unanimous AI has continued with its form of making accurate predictions. This time it aimed at the 90th Academy Awards and published its predictions a week prior to announcements. It is based on “Swarm AI technology” which combines real-time human input with its algorithms to optimize a group’s combined knowledge and intuition into a unanimous output. It does so by building artificial “hive minds” that amplify the intelligence of human populations to create an artificial super-expert that can outperform traditional experts. For the Academy Awards, researchers created a swarm of 40 movie enthusiasts using AI algorithm in real time. In 15 of the 16 categories, the AI was able to perfectly predict the winner of the Oscars.



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