This is a guest post by Peter Sarlin, CEO at Silo AI.
With rapid developments in digitization and digitalization, the world is being filled with digital data, with ever more created every day. Much of that data is unstructured – it lacks pre-defined standards and structure that would allow computers to process it efficiently and accurately.
A common layman’s example of unstructured data is a text-heavy contract written in Word and saved as a PDF. Yes, it probably has titles and paragraphs, even a table of contents, but it still misses several qualities that would enable automated processing.
So, how can we process high amounts of unstructured documents efficiently to enable finding information quickly and generally getting more value of that data?
NLP bridges humans and computers
Dating back to the 1950s, natural language processing (NLP) involves the computers’ ability to process written language created by humans. Today, NLP is a mixture of linguistics, computer science, information engineering and artificial intelligence. In the past years, NLP has developed into a practical, powerful tool that helps companies make better use of vast amounts of unstructured data. One common focus area of NLP is to tap into the potential to create rich data sets from a wide variety of different textual data sources.
For example, a NLP-powered system can analyze a PDF contract document, understand certain elements such as contract parties or contractual details, and automatically pick up that data into a system in which it can be further processed. And together with machine learning, NLP systems can also improve over time as they process more data.
Value-add and potential of NLP
Developing NLP solutions requires deep knowledge and capabilities in both data collection and data processing as well as NLP and machine learning. While certain parts of document processing can be standardized into a product, such as Zefort, we at Silo AI focus on developing bespoke NLP solutions to tackle custom needs and purposes.
With top-notch capabilities in NLP, we have experience from developing NLP-driven AI solutions for a wide variety of purposes. Silo AI employs approximately 100 leading professionals, including a team of NLP experts with NLP professor Filip Ginter at the head, as well as a full development and deployment infrastructure to rapidly build and launch NLP solutions.
We see that NLP-powered solutions have the potential to create significant value in many fields. So far, the most prominent application areas have been for use cases in the finance and legal sectors. To understand the potential, we describe here two use cases from Silo AI’s customer projects in finance.
Case: Generating investment leads with NLP
In this collaboration with Infranode, a Nordic long-term infrastructure investor, we co-created a NLP-powered solution that analyzes a vast amount of data sources to identify potential investment leads for further validation.
The discovered opportunities are then curated and catered as a simple email to Infranode’s investment team. This allows the team to improve their efficiency in identifying the investment opportunities that are most relevant for the company’s strategy. And feedback to be collected to further improve the solution.
Case: Collecting data on extreme scenarios related to climate change risks
Together with Riskthinking.AI, a Canadian science-based scenario generation company, we created AI-powered scientific approaches to understanding climate change risks. Scenarios allow organizations, such as financial institutions, to understand and prepare for the future.
In the first risk scenario, we created an intelligent system that analyzes thousands of textual articles with the help of NLP. As a result, the solution is able to extract published predictions regarding the rise of sea levels and use it as an input to a multifactor scenario generation model, in order to understand underlying risks of extreme scenarios.
CEO at Silo AI
Silo AI is the largest AI solution and service provider in the Nordics that solves the most strenuous challenges in machine learning, computer vision and natural language processing. Silo AI aims at being a trusted partner that brings AI into product development and delivers AI-driven solutions and products. The company serves clients across several industries on four continents.
This is a guest blog by Clemens Brunner, CEO of…Read More
Our Contract Challenges series has previously talked about search metadata,…Read More
Following the ERP boom in the early 2000s, IT people…Read More