The main challenge in the knowledge industry is to identify relevant data hidden within layers of noise and to build meaningful inferences – assessments, ratings, opinions – based on this data. In the ESG (Environmental, Social and Governance) research and ratings industry, this challenge is especially salient these days, as companies worldwide want to show to their stakeholders that company ESG performance is concordant with increasing stakeholder expectations.
We will show how a leader in the ESG industry Sustainalytics, a Morningstar company, has built a unique combination of document embedding and fast similarity detection to identify relevant corporate disclosure and to build automated inferential procedures to texts to the reality these texts describe.
Recommender systems are becoming an integral part of our lives, and they help us filter through the information overload. But they are not a silver bullet and might create problems in the long run. Reinforcement learning might be able to help. In this presentation, I will introduce Recommenders and reinforcement learning and then discuss how RL might alleviate some of these challenges.
Like the name suggests Artificial Intelligence of Things (AIoT) is the combination of artificial intelligence (AI) technologies with the Internet of Things (IoT) infrastructure to achieve more efficiency in IoT operations, improve human-machine interactions and enhance data management and analytics.
Current IoT systems can only react to an event while AIoT systems can proactively detect failures and events. The infusion of AI in IoT systems delivers the promise of predictive maintenance which will help organizations save millions of dollars in support and maintenance of equipment. Moreover, the future of industrial automation lies in the convergence of AI and IoT. The importance of AIoT to Bosch can be gauged by the fact that the Artificial Intelligence of Things will impact almost every industry vertical including automotive, aviation, finance, healthcare, manufacturing and supply chain.
A goal of commercial insurers is to accurately classify small businesses according to the risks they face and thereby determine the correct insurance premiums. Setting the proper compensation is essential because charging too much or too little can harm insurers and their customers. 50% of the business applications required the industry classification code correction. This mainly was a manual exercise of looking up business details on the internet or purchasing business classification codes from 3rd party providers. The recent advancements in computational power and machine learning have led to vast improvements. However, the accuracy level was still less than the industry norm until the Cognizant team experimented with business ontology. The Cognizant team coded the ontology model for restaurant business using the NAICS and ISO industry manual. With details around the business name, address, and URL Cognizant team extracted the data about the company from its corresponding website and 3rd party data providers. It was able to classify the restaurant business with >90% accuracy.
This session describes ontologies and their use in computational reasoning to support the precise classification of small businesses for insurance applications.