NLP is a subset of artificial intelligence, it involves teaching computers to read and interpret words and sentences. It’s been around since the 1970s, but recent advances in deep learning have made it a more powerful weapon for analyzing text.
Insight from a fund manager
Why choose natural language processing?
Fund managers are searching for untapped sources of alpha.
Malcolm is the lead fund manager of a global equity fund who wants to reduce the growth of human resource costs.
Volume and diversity of data
Malcolm’s analysts spend a significant amount of time digging through an expanding pool of data. The goal is to discover untapped sources of alpha. This involves examining unstructured data to extract actionable information. Converting raw data into insight remains a difficult challenge for investment professionals. This level of difficulty arises not only from the volume of data but the diversity of data as well. For instance, Malcolm expects analysts to synthesise information from traditional sources. Traditional sources include company filings and earnings calls. But, they also include alternative sources such as consumer spending and lifestyle data.
This presents Malcolm’s team with a billion-piece jigsaw puzzle. Unstructured text data represents a box of puzzle pieces. Malcolm’s analysts must compose these into a clear picture of the investment landscape.
The expanding puzzle
The challenge for Malcolm’s team is constantly changing unstructured text data. Every time Malcolm’s team makes some progress, it feels as though a saboteur sneaks into the puzzle room and dumps a few million more pieces in the box. For good measure, the scoundrel knocks over the table and slinks out. This means Malcolm’s team struggles to maintain a clear picture of the investment landscape.
Volume and diversity of data
Malcolm reached a point where the business could not sustain extra human resources expenses. The problem was two-fold. On one side, it was not feasible to keep adding analysts for the composition and interpretation of an increasingly complex puzzle. On the other side, Malcolm saw that his most talented analysts were spending too much time putting puzzle pieces together and not enough time interpreting the picture that the connected pieces reveal.
It became clear that as data volumes increase, Malcolm’s team would need more sophisticated tools to optimise their portfolios.
Implementing knowledge graphs
Malcolm implemented an NLP solution that allowed his asset manager to automatically capture and process text-based data sources. This solution allows his team to leverage information that is unknown to others.
The information that the NLP model produces renders a knowledge graph. This graph represents the relationships between different data points. By connecting data in this way, Malcolm’s knowledge graphs mimic the human ability to derive contextual meaning and produce surprising insights that were previously impossible. By connecting data in this way, Malcolm’s team can harmonise and quantify their assessments by minimising the uncertainty introduced by human bias.
This transformation has reduced margin pressure and increased alpha for Malcolm’s clients.
|Systems||Supervised Machine Learning||Intelligent Automation|
|Innovation||Unsupervised Machine Learning||Expert Systems|
|Management||People Analytics||Supervised RPA|
|Operations||Process Analytics||Unsupervised RPA|
|Resources||Natural Language Processing||Learning Bots|
What's Your Story?
Address71-75 Shelton Street, Covent Garden
London, WC2H 9JQ