Sep 11 2020
Digital Reasoning, the global leader in artificial intelligence solutions that understand human communications and behaviors, announced today breakthrough, patent-pending research with advanced, intelligent technologies that allow customers to train models in English and immediately transfer the learnings to many other languages.
While speaking to a large audience at a Neural Information Processing System (NIPS) conference (2016), Andrew Ng, the world-renowned AI expert, once remarked that “transfer learning will be -- after supervised learning -- the next driver of Machine Learning commercial success.”
“I am happy to announce that Digital Reasoning’s Applied Machine Learning Research Group, after two years of iterative research, has made a significant breakthrough in the area of transfer learning and will be unveiling this technology soon to the market,” said Dr. Uday Kamath, Chief Analytics Officer at Digital Reasoning. “Working with Dr. Wael Emara, we’ve combined bleeding-edge research in AI and Deep Learning to solve one of the hardest industry problems.”
Digital Reasoning works with many of the largest global banks to analyze all of their electronic communications so they can identify any misconduct that would put their institutions at regulatory or reputational risk. Historically, banks, who have to manage policies across many different regions, have found that building Natural Language Processing (NLP) Models across low-resource languages is costly, error-prone, and time-consuming.
“As we deploy these latest AI breakthroughs, we will dramatically accelerate our customer’s ability to efficiently and effectively apply machine learning across languages and further extend the value of our Machine Education Platform,” said Kamath. “This is groundbreaking work that has the ability to change both the Conduct Surveillance industry and many others. We feel privileged to work alongside many of the largest banks in the world in moving the industry forward.”
Digital Reasoning has been working alongside industry experts to validate and refine their applied research. Industry leader and former Global Head of Monitoring, Surveillance & Controls at UBS, Steve Livermore, has reviewed the work of the Applied Machine Learning Research Group and remarked, “I cannot overstate how significant this progress is. High-quality alerts are key to effective surveillance and to be able to rapidly build new models in various languages, not only extends coverage in a consistent manner, but it can deliver real cost savings at the same time.”
While showcasing the advancements in multilingual models, one executive at a major global financial institution remarked, “I believe once this is out there, banks will have to explain to regulators why they are not using it rather than why they are.”
Digital Reasoning and its partner, PwC, will host a webinar to cover multilingual models and other significant advancements in electronic communications surveillance solutions on Sept. 17 at 2 p.m. Singapore Standard Time.