RiskLab has completed successful research into creating NLP (natural language processing)algorithms for extracting ESG-related twitter data. These ESG (environmental, social, and governance) algorithms are in pursuit of creating an “ESG score” for a corporations. The largest companies that evaluate corporations on ESG are predominately performed in a manual ad-hoc manner such as checklists. Natural language processing and other machine learning algorithms are particularly well-poised to overcome the current methodologies’ weaknesses. We’d like to continue this endeavour with several generalizations of our first approach.
Truly successful applications of machine learning in finance are far and few between. That being said, with enough financial specificity, significant improvements in classical financial models are possible with machine learning. We’ve found significant success thus far in using ML to predict future covariances via “financial neural network embedding”, construct optimal portfolios, and determine market sentiment. We’d therefore like to continue this research and find further applications of AI in finance.
An economic focused investigation is currently underway for how fast one can replicate an ML model via querying its API and training a copy of said model via the queries. This being a fairly new area of research being pioneered at RiskLab, we’re very interesting in continuing this research.