IBM reports that Watson is ready to take on the most challenging aspects of the English language. New natural language processing (NLP) capabilities will be added to the artificial intelligence platform later this year to allow businesses to identify and analyze some of the most challenging aspects of human language.
IBM has been developing these NLP skills with IBM Research’s Project Debater, an AI system that can debate humans on complex topics.
SEE: Cheat sheet: Artificial intelligence (free PDF) (TechRepublic)
The new “advanced sentiment analysis” capability helps Watson to understand expressions such as “better late than never” and “hot under the collar.” This means the platform will be able to read and analyze text-heavy documents such as contracts and project proposals. Now companies will be able to analyze such language with Watson APIs for a more thorough understanding of their business. These advances also will allow companies to incorporate business documents, such as PDFs and contracts into AI models.
“Language is a tool for expressing thought and opinion, as much as it is a tool for information,” said Rob Thomas, general manager, IBM Data and AI. “This is why we believe that advancing our ability to capture, analyze, and understand more from language with NLP will help transform how businesses utilize their intellectual capital.”
IBM showed off one of these new capabilities earlier this year at the Grammys. IBM used the platform’s Summarization technology to display information cards about recording artists in real time as they walked the red carpet. Summarization pulls text from multiple data sources to create a synopsis of a person or event on the fly. At the Grammys, the data was added to on-demand videos and photos on grammy.com. The capability will be added to IBM Watson Natural Language Understanding later this year.
Throughout the year, IBM will integrate these other Project Debater technologies into several Watson products to improve clients’ ability to analyze natural language:
Analysis–Advanced Sentiment Analysis: This allows Watson to understand “sentiment shifters,” which are combinations of words that take on new meaning, such as, “hardly helpful.”
Clustering–Advanced Topic Clustering: The technique also allows subject matter experts to customize and fine tune topics by adding related information to reflect the language of specific businesses or industries, such as insurance, healthcare, and manufacturing.
Documents–Customizable Classification of Elements in Business Documents: This technology enables AI models to more easily classify clauses that occur in such business documents as procurement contracts. The platform can learn from as few as several hundred samples to do new classifications quickly.
Project Debater in action
In November 2019, Project Debater took on the statement, “AI will do more harm than good,” at Cambridge Union in England. Project Debater supplied the opening speech for both the pro and con arguments and then the two human teams took over the rebuttal and summary. At the end of the debate, the audience voted, and 48% agreed that AI will do more harm than good but 51% disagreed with the premise. “At least for now, the world is in favor of AI,” the host of the event said.
Project Debater is the first AI system that can debate humans on complex topics. The system digests massive texts, constructs a speech on a given topic, delivers it, and rebuts its opponent.
IBM’s long-term goal is to help people reason by “providing compelling, evidence-based arguments and limiting the influence of emotion, bias, or ambiguity.”
Earlier this year, researchers from the MIT-IBM Watson AI Lab, Tulane University, and this week the University of Illinois unveiled research that allows a computer to more closely replicate human-based reading comprehension and inference.
Project Debater has been in development since 2012 and represents a significant milestone in IBM’s AI work. IBM has applied its NLP technology to enterprise search, conversational AI, and advanced text analytics.
ESPN Fantasy Football uses Watson Discovery and Watson Knowledge Studio to analyze millions of football data sources to offer millions of fantasy football players real-time insights. With NLP, Watson identifies the tone and sentiment of news articles, blogs, forums, rankings, projections, podcasts, and tweets that cover everything from interviews to injuries.