AIDA Technologies: Where Predictive Analytics Meets Human Intelligence

Dr. Andrew Ang, COO, Dr. Tan Geok Leng, CEO and Prof. Shonali Krishnaswamy, CTO

“Human ingenuity will outsmart the rule engine, each and every time.” This is how Dr. Tan GeokLeng, CEO of AIDA Technologies began his presentation at the Singapore MAS FinTech Hackcelerator 2016. What followed was an insightful explanation on how Artificial Intelligence Driven Analytics (AiDA) can be leveraged to address the issues facing Traders’ Risk Management. Dr. Tan’s pitch further covered how AIDA technology can analyze traders’ communication patterns to spot the usage of unusual words that might signal suspicious activity. The innovative idea not only won AIDA the prestigious MAS Global FinTech Hackcelerator Award but also marked the beginning of AIDA’s exciting start-up journey in Nov 2016.

AIDA was founded by a team of PhD-level scientists, to assist businesses around the world that struggle to effectively utilize massive volumes of data from heterogeneous data sources, largely focusing on banks and insurance industries in the region. AIDA offers AI-driven analytics to address such challenges through the automation of knowledge work by augmenting human-expert abilities to make decisions from large amounts of heterogeneous information. The company brings proprietary technology innovations, which help extract-valuable information from both structured and unstructured data and develop both supervised/predictive and unsupervised machine learning solutions. The technology blends a multi-layer machine learning approach to achieve superior levels of predictive accuracy. AIDA also has sophisticated unsupervised machine learning technologies to reduce false positives when detecting outliers, a specialized ability that detects conceptual drift or gradual change in data, a deep learning engine for text analysis, and an expert ability to extract information from documents including tables and graphs.

Beyond precision or accuracy, businesses also need another functionality of using machine learning solutions effectively for interpretability and explainability of the derived results.

At AIDA, we have developed mechanisms to provide interpretability and explainability for the sophisticated machine learning approaches

“At AIDA, we have developed mechanisms to provide interpretability and explainability for the sophisticated machine learning approaches. Focusing on business metrics, we are able to extract reasons for a machine’s scoring or decisions,” asserts Prof Shonali Krishnaswamy, CTO at AIDA.

The AIDA technology first extracts information from text bodies and uses it’s proprietary deep learning engine to conducts emantic text mining. It then combines the information with its outlier and concept drift detection engines for unsupervised machine learning. This also improves predictive precision by intelligently combining structured and unstructured data. For organizations, this machine learning approach leads to higher precision, data-driven decision-making, and improved productivity.

AIDA, with its machine learning innovations facilitates highly precise solutions for the financial services industry. To mitigate a wide range of bottlenecks pertaining to claims processing between insurance companies and health claims decision support teams, AIDA designed, implemented, and validated its AI-based, highly precise machine learning engine—AI-CLAIMS. The engine establishes a splendid correlation between human expertise and machine learning to significantly improve claims processing productivity and efficiency.

To cite an instance, AIDA’s advanced AI-based solution, AIDA-KYT (Know Your Trader), expedites trading floor irregularities detection for compliance and audit teams, necessitated by current regulatory requirements. The KYT system is based on advanced AI and machine learning coupled with deep analysis capabilities to detect unusual communications (beyond traditional dictionary-based text mining) and unsupervised concept drift to detect behavioural change. It promotes an effective KYT perspective by automatically and continuously monitoring trading floor transactions and communications data, as well as highlighting outlier behaviours and unusual communications.

Dr. Tan continues, “With our expertise and experience, we are able to rapidly model problems and correspondingly employ our machine learning tools, techniques, and methods to develop impactful and transformational solutions.” As the key factor constituting to AIDA’s competitive advantage Dr. Tan acknowledges his team comprising of more than 10 PhD level scientists and technologists. With the consistent efforts of its team, AIDA has gained a foothold in Singapore and across, Malaysia, Indonesia, Thailand, and India in a span of less than a year. Having already secured Tier-1 banking and insurance clients in these countries, the roadmap of AIDA is to go global by their third fiscal year.