AXA Investment Managers' quantitative equities arm Rosenberg is integrating advanced modelling techniques known as "neural networks" into its investment approach.
The use of neural networks, which are flexible artificial intelligence models, loosely based on how the human brain processes data, will enhance Rosenberg's data analysis and complement its existing investment process, the firm said.
Rosenberg's Sustainable Equity fund, which is part of the Advanced Factors range, will be the firm's first strategy to incorporate the technique.
The new model aims to improve the strategy's ability to identify stocks that are at risk of extreme price events, thereby mitigating tail risks, while booting returns.
Gideon Smith, Europe CIO of AXA IM Rosenberg Equities, said the move was "a natural progression of the advanced quantitative techniques" the firm had adopted in the last 30 years.
He said: "We believe this is a considered step and consistent with our investment approach of modelling and managing investment risk for clients.
"It's an area where we have a wealth of knowledge and expertise and we are excited to leverage these techniques with the aim of improving client results in all areas of the investment process."
Rosenberg will also be employing "unstructured data sets" to provide deeper insights into stock fundamentals and sentiment.
For example, the firm is developing a "natural language processing model", which will analyse company filings and other text sources to derive great insight in investor sentiment associated with a company.
CEO of AXA IM Rosenberg Equities Heidi Ridley said both developments are a great example of how AI and unstructured data can improve existing modelling techniques.
She added: "[The] innovative use of data and technology leads to better insights and ultimately better outcomes for our clients."
Approaching $1.5bn capacity limit
Leader of People's Party
Listing two global equities strategies
Aims to encourage competition in industry
Markets went into freefall