Singularity Pulse  
     
  Banking

ML to Help Reduce Criminal Risks for The Banking Sector

Mindtree in collaboration with Tookitaki's ML powered platform is helping banks improve their crime detection activity. They have come up with two services namely, “Smart Alert Management” and “Smart Reconciliation Management”. The former is an automated, adaptive model based on AI and ML technology to detect suspicious cases with more accuracy. It is significantly better in reducing false alerts and increases true positives. The latter is an end-to-end approach to reconciliation management across business activities. Match rates, exceptions, audit trails etc are generated through the algorithm. It is capable of auto-correcting systems while managing to stay compliant.

 
  Image  
 

AI Algorithms Learn from Patient Data to Make Treatment Less Toxic

MIT Media Lab researchers have come up with a model that can make dosage regimens for treatment of Glioblastoma, a tumor that appears in the brain or spinal cord, less toxic and still equally effective for the patients. In study of 50 patients, ML model developed treatment cycles that reduced the potency to quarter or half of almost all the doses while maintaining the same tumor-shrinking potential. With an iterative process, it finds an optimal treatment plan, with lowest possible potency and frequency of doses which can still reduce tumor sizes to an extent comparable to that of traditional regimens. At times, it completely skips a dose, scheduling appointments only twice a year instead of monthly.

Health Care  
  Image  
  Banking

Miami Bank to Adopt AI System to Combat Money Laundering

Miami branch of Chile-based Banco de Credito e Inversiones (BCI) bank has entered into a three-year agreement with start-up “QuantaVerse” to monitor for money laundering activities. It comprises of 3 systems wherein, the first one will automate and complete 70% of an anti-money laundering investigation detected by the transaction monitoring system. The second program will detect abnormalities by analyzing years' worth of financial data. The third program will aim on reducing false positives. The conventional detection methods miss, on average, 50-70% of potential criminality. Bank is hopeful to increase the success rate of the program substantially.

 
  Image  
 

New Genres of AI Algorithms Post New Challenges Towards Hacking

IBM Corp. has deployed AI technique to build hacking programs that could easily infiltrate through top-tier defensive measures. Conventional defenses rely on examining what the attack software is doing, rather than analyzing software code for danger signs. The new genre of AI-driven programs, instead, are capable of being trained to stay dormant until they reach a very specific target, making them exceptionally hard to stop. One of the most recognized examples is Stuxnet, which was deployed by U.S. and Israeli intelligence agencies against a uranium enrichment facility in Iran. IBM’s “DeepLocker” is capable of reaching equal precision with fewer resources.

Information Technology  
  Image  
  Information Technology

ML to Remove Biases from Speech Recognition

There are particularly three types of biases in speech recognition algorithms; underrepresented demographics, gender biases and accent biases. Mozilla is currently working on a program they are calling “Common Voice” that aims to eliminate these biases that can happen when there aren't a range of voices used to develop the technology. They are planning to crowd source voice samples for the purpose. It collects plenty of pre-prepared sentences from public in multiple languages. These samples are then used to train their ML model called “Deep Speech”. The samples are freely available as datasets on internet for download.

 
  Image  
 

Algorithm to Identify Mental Illness, Response to Medication

Lawson Health Research Institute and the Brainnetome Center have developed an algorithm that can distinguish between two different types of mental illnesses and further predict patient reactions to those medications. Researchers examined brain scans for 60+ patients who had been previously diagnosed with either major depressive disorder or bipolar disorder and compared them to scans from 33 other participants who had no known history of mental illness. Following this, researches developed an AI algorithm based on this data. Using ML, their algorithm was able to examine scans and determine whether a patient had MDD or bipolar disorder with an accuracy of 92.4%. Algorithm was further applied to 12 additional participants with complex mood disorders for whom no clear diagnosis was available. Algorithm also predicted the diagnosis which was later proved to be beneficial for 11 out of 12 participants.

Health Care  
  Image  
  Logistics

Samsung To Develop A ML Powered Autonomous Ship Platform

Samsung Heavy Industries is developing an autonomous shipping platform and has announced Amazon Web Services (AWS) as its preferred cloud vendor. The platform will be designed to enable self-piloting of container ships, LNG carriers and floating production systems. The company further plans to partner with AWS for developing its ML algorithms, augmented reality, analytics and databases. It will also develop technology which will be used inside the vessel for improving loading mechanism and optimize its own machinery.

 
  Image  
 

Grocery and Dish Recommendations from An AI System

Halla, an LA based start-up, has developed a platform called “Halla/IO” that uses AI to provide its users with recommendations for grocery, restaurant, food delivery apps as well as websites. One can consider this as Netflix for food ordering. The algorithm uses psychographics and data to predict preferences. Predictive analytics is applied to taste and flavor attributes to further deepen the understanding. The database comprises of more than 10,000 grocery items, 20,000 ingredients, 175,000 recipes, and 20 million restaurant dishes.

Miscellaneous  
  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

Graphene Nanoribbons to Exhibit Quantum Properties

Graphene nano-ribbons consists of atoms of single element, but exhibits different properties depending on how the atomic arrangement is. Its electronic character - conductor, semiconductor or insulator depends on shape and width of the element. Via varying the shape, specific local quantum states can be generated. Empa researchers showed that if ribbons are built of alternatively different widths, a chain of interlinked quantum states with its own electronic structure is created by the numerous transitions. The switching distance between the “1” state and the “0” state of the nano-transistor is quite large, but with the new technology it can be set as desired.

 
  Image  
 

Physicists Develop a Circulator 1000 Times Smaller in Size for Quantum Computers

University of Sydney, Stanford University and Microsoft have made use of the phase of matter, “topological insulator” in developing an electrical component called “circulator” which is 1000 times smaller that its current size. This will allow squeezing in more qubits into a small space. Qubits are piece of electronics that uses probabilities of an unmeasured bit of matter to perform calculations. The movement of electrons becomes critical when electrical components are shrunk by such high factors. Until now, the smallest circulator is almost the size of the palm of one’s hand.

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