- Could you give us your backstory?
I’ve been involved with artificial intelligence and machine learning startups since leaving M.I.T. I founded Basis Technology to focus on software globalization and American companies entering Asian markets. The company grew and expanded into text analytics, natural language processing, and digital forensics. We recently launched an incubator program to make our technology available to startups.
- What is Natural Language Processing? How does it work? How has it evolved over the years? What’s around the corner?
Natural language processing (NLP) is the field of computer science focused on human-machine interaction. Basis Technology’s area of focus within NLP is deriving meaning from written text in digital formats.
In most NLP systems, computers first collect data to create training sets tagged by humans to predetermined categories. Based on that training data, statistical models are built to recognize future documents that fall into those categories, and a runtime engine processes previously unseen documents to generate the desired answers.
Recent advancements in NLP, such as our Rosette platform, use a combination of new and pre-existing data to offer discrete functional services including language identification, entity extraction, sentiment analysis, and identification of names of people or things. Higher levels of analytics are moving toward cognitive computing and machine intelligence.
The more proficient machines become in understanding human language, the better they will be able to connect with people.
- Could you explain the relationship between NLP and sentiment analytics?
Most of what I’ve learned about sentiment analytics has come from working with Prattle and other partner companies.
Many sentiment analyzers are naive and incapable of extracting insights from nuanced text. For example, consider how a simple model might interpret this hotel review: “the room was great, but the service was terrible.” If the model is unable to assign value to individual parts of text in an informed manner–an ability which many models lack–the score produced is useless. At Basis Tech, we’re working on evaluating the sentiment of specific concepts within statements.
- What are the current and potential applications of these textual analysis technologies to finance?
The applications are limitless.
For example, every day we hear a mistaken click story–a trader pushes the wrong button or enters the wrong numbers and causes a significant loss. We need an intelligent agent that can “look over the shoulder” of traders and verify that certain information is correct.
Here’s another example. Companies that are legally required to disclose cyber security vulnerabilities will often issue complex legal documents that are difficult to understand. Text analysis is being used to parse these documents and find important, relevant disclosures.
Finally, high-speed trading is an area rife with possibilities for the application of textual analysis. Traders are always trying to recognize market-moving events and capitalize on the resulting changes immediately. Text analytics are enabling high-speed trading algorithms to understand the meaning of financial documents faster than other analysts and make the correct moves to avoid losses.
- Where do you see us fitting into the space?
Prattle is changing the way financial institutions think about central banks. Historically, economists have relied on highly skilled human analysts for “Fed watching.” This new application of a rigorous, algorithmic approach to analyzing written communications provides the advantage of machines’ inherently unbiased and thorough results. Prattle combines human experience and wisdom with these advantages, creating a truly unique signal pertinent to many aspects of financial activity.
- Thanks so much for your time. Before you go, do you have any closing thoughts you’d like to leave with us?
Text analytics is a story that is just getting started. The level of activity from big players in machine learning and language analysis is rapidly accelerating.
It’s a very exciting time.
The Signal Team
About Carl Hoffman
Carl Hoffman is the CEO and co-founder of Basis Technology Corp., the leading provider of components for information retrieval, entity extraction, and entity resolution in many languages. Before founding Basis Technology in 1995, Carl spent 10 years at MIT as both a student and on the research staff of the Laboratory for Computer Science. He also worked as an independent consultant in Boston, New York, and Tokyo to international clients in finance and knowledge management.
About Basis Technology
Analyzing text—the hardest part of big data—is critical to verifying identity, understanding customers, anticipating world events, and uncovering crime. Companies such as Airbnb®, Luminoso®, Recorded Future®, Tamr™, and Yelp®, and agencies across the U.S. intelligence community, use Rosette to solve their toughest human language problems. For over 20 years, Basis Technology has been at the forefront of natural language processing applied to enterprise search, social listening, e-commerce, and e-discovery. Our cyber forensics team pioneers better, faster, and cheaper techniques to extract digital evidence, keeping government and law enforcement ahead of exponential growth of data volumes. For more information, visit www.basistech.com or write to firstname.lastname@example.org.