As a bridge between humans and computers, the use of natural language processing in compliance demonstrates how this branch of AI is adding value across the financial services industry.
- Natural language processing and text analytics when used together can meet a range of challenges in financial services.
- In compliance, data is structured and categorized to help with due diligence screening for AML.
- Human error is removed and the strain on resources diminished, freeing staff to concentrate on areas of higher value.
For more data-driven insights in your Inbox, subscribe to the Refinitiv Perspectives weekly newsletter.
Natural language processing (NLP) falls under the wider umbrella of artificial intelligence (AI) and essentially uses algorithms to help computers understand the everyday language of humans — both spoken and written.
As such, NLP is a fundamental bridge enabling interaction between computers and humans, and allowing machines to understand commands and input from humans in a seamless and streamlined manner.
Once a machine can understand a human’s everyday means of communication, the potential to add value becomes almost limitless.
Some real world applications
The Center for Cognitive Computing is constantly researching ways to perfect and advance different areas of AI, including machine perception, reasoning, knowledge management, and human-computer interfaces.
The focus is always on designing algorithms that can offer practical solutions to real-world challenges. Many of these solutions make use of NLP.
One such example of NLP in action is News Analytics, an NLP system that cuts through the noise for financial market participants dealing with volumes of media content by analyzing Reuters news.
Through harnessing machine power, the system can track breaking news and developing stories affecting thousands of companies, all in real-time.
For traders, this is an invaluable resource that supports quantitative strategies and frees time for asset managers to concentrate on higher value add tasks, such as asset allocation decisions.
None of this would be possible without NLP.
Other examples of NLP applications in everyday use include the converting of spoken words into text and text into speech, and automatic translation from one language to another.
Combining NLP and text analytics
In the age of big data, the sheer volume of content that must be analyzed and managed in any industry can be overwhelming.
Machines can help to structure this content and cut through the noise, pinpointing the information that is relevant to an individual user at that time.
Enter text analytics, where content is grouped and structured and large volumes of content are analyzed to detect themes or patterns.
Text analytics and NLP are often used together to solve a range of challenges and the combination can be applied across myriad industries.
Due diligence screening
For compliance professionals, conducting due diligence screening in line with AML regulations frequently involves analyzing and managing significant volumes of content.
By harnessing the combined power of NLP and text analytics, data can be structured and categorized.
For example, documents or individual sentences can be tagged so they become searchable; events can be clustered; or duplicate content can be highlighted.
Download Expert Talk – How is AI rising to the challenge in the fight against financial crime?
All of this saves significant amounts of time because staff no longer need to manually trawl through volumes of content.
Additionally, the human error factor is removed and, perhaps most importantly, the strain on often scarce resources is diminished, with staff freed to concentrate on areas of higher value-add.
Challenges and limitations
It is important, as always, to maintain realistic expectations.
In our recent report, 2018 AI Predictions, an example of current NLP limitations was illustrated by looking at the capabilities of Siri, Apple’s well-known ‘intelligent personal assistant’.
In the example, Siri was able to interpret and act on the simple command ‘play music I like’.
However, the more complex command ‘play music my wife likes’ resulted in the system defaulting to a web search, despite the fact that the user and his wife shared the same favorite music list on a family music plan.
Both commands use natural language, but the more complex command resulted in failure.
Read Alex Paladino’s blog – Tech must apply roboethics to keep AI from learning our mistakes
This simple example highlights that there are still many challenges and limitations to AI in general — and NLP in particular — that users need to take account of.
However, ongoing research and development and an ever-growing appetite for all things AI, mean that new capabilities are developing in exponential time.
