Singularity Pulse  
     
  Information Technology

Facebook to Automatically Translate Posts Into Different Languages

Researchers at Facebook have developed a quicker and more accurate way of translating low-resources languages like Urdu and Burmese using Artificial Intelligence. Facebook has found a way to translate material on its social network from base language to a different language with enhanced speed and accuracy. The development is based on AI method called, “Convolutional Neural Network”. Currently, the technology is already being deployed for image processing and other ML tasks. It will also be used for chatbots. With time, the social media giant is planning to use multilingual posts to improve its machine translation capabilities. Currently, Facebook translates posts in more than 45 languages.

 
  Image  
 

Bloomberg To Use Data Science To Move Markets

Gideon Mann, former scientist at Google is developing data science tools and techniques at Bloomberg which will help company’s journalist automate analysis of news, social media feeds, company financial documents and press release. The stress is more on unstructured data now along with fundamental economic information such as market prices and stock exchange data. The new process involves computers coming up with first draft of the story based on major shifts or deviations. A human journalist will then edit and elaborate the same. The system uses optical character recognition (OCR) and deep learning techniques to read numerical data points from images of charts and transform them into tables.

Finance  
  Image  
  Research

AI Enabled Causal Models To Mitigate Risk For Oil And Gas Industry

DNV GL, an international accredited registrar and classification society, has released a position paper, “Entitled AI + Safety”, on use of AI for safety of oil and gas industry. Currently, system’s safety is ensured through “control theory” which considers predefined set of rules and current states of the system as decision making factors. In contrast, AI tries to learn reasonable rules based on previous experience. The research paper challenges the conventional approach and calls for a supplement to these data-driven models by generating physics-based casual data. it will be significantly closer to the safe implementation of AI in safety-critical systems.

 
  Image  
 

Clickbait Pattern Detection To Be Done Via NLP Technique

“Clickbait” is one of the most prevalent technique used by websites to brag users’ attention who are searching for information on particular topics online. Clickbait headlines are designed to capture reader’s interest in what is promised. The developed model is aimed to extract features which constitute as clickbait. Features can be both syntactic and semantic in nature. The technology uses multiple classification algorithms and classifiers to distinguish between actual news and a clickbait.

Information Technology  
  Image  
  Civil Engineering

Deep Learning Algorithm to Predict Aftershocks of an Earthquake

Scientists at Harvard and Google have developed an algorithm which can predict size, timing and location of aftershocks after an earthquake has hit an area. The neural network was trained on database of more than 131,000 “mainshock-aftershock” events to look for patterns. Following that, it was deployed for making predictions on a database of 30,000 similar pairs. The new model outperformed existing “Coulomb failure stress change” in making predictions. It scored an accuracy of approximately 85% as compared to 58% for existing model.

 
  Image  
 

Predictive Policing To Put Hold On Criminal Activities

Belgian police department are creating tools and building databanks and algorithms to predict both, time and place prone to a criminal activity. The new policing computer system will be called iPolice and is expected to be launched by 2020. The data needed to make these predictions will be supplied by the police department along with third parties like Met Office. This data will then be introduced to allow creation of links connecting possibility of a crime activity taking place with its predicted time of occurrence.

Civil Services  
  Image  
  Digital Imaging

AI Powered Background Removal, Social Filters For Images

Researchers from MIT, Adobe, and Microsoft have developed a technology that is capable of removing backgrounds from images in one click. The process is called “semantic soft segmentation” and uses ML neural network to analyze images using texture and color along with factors like object recognition to automatically segregate images into multiple layers. It creates a transparent background, place the subject on an entirely different background, and then apply changes either to the subject, or the background. Currently the process works for still images but the team in hopeful to use it for videos too.

 
  Image  
 
Image
“Quantum Computing, a revolution in technology, has shown great promise towards solving complex computing problems currently outside the capabilities of current computers. Though in its infancy, we at Decimal Point Analytics strongly believe that it is going to grow exponentially in the near future. It holds the potential not only to boost the AI revolution but also transform the way data is synthesized. The articles on Quantum computing will give an insight into the recent developments in this space.”
 
  Image  
  Quantum Computing

“Bose-Einstein Condensation” To Make Quantum Computing Scalable

There are three types quantum computing; gate quantum computing, adiabatic quantum computing and quantum walks. Gate quantum computers are the most common. Quantum walks are found in nature. It refers to the way in which electron-transfer step occurs in photosynthesis works. These walks can be implemented with light. Researchers have found a condensation called, “Bose-Einstein” condensation (BEC) where roles of light and matter can be reversed. In quantum walk, problem is encoded as series of paths. Quantum state takes all possible paths simultaneously but the path that encodes solution has higher probability compared to others. BEC version is most flexible one where light kick move the BEC along in free space, while the microwave pulses act like couplers between different paths. The light and microwave pulse can be changed, hence making the route programmable.

 
  Image  
 

Multifunctional Solid-State Quantum Memory Achieved

Researchers from CAS Key Lab of Quantum Information have developed a multi-degree-of-freedom (DOF) multiplexed solid-state quantum memory. This allowed reliable storage and coherent manipulation of quantum states with matter-systems. To attain useful communication rates, multi-mode quantum memories will be used to construct a multiplexed quantum repeater. The new developed technology can also serve as quantum mode converter with high fidelity. The device is capable of performing arbitrary pulse operations within time and frequency DOF.

Quantum Computing  
  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