During IBM’s virtual AI Summit this week, the company announced updates across its Watson family of products in the areas of language, explainability, and workplace automation. A new feature called Reading Comprehension surfaces answers from databases of enterprise documents in response to natural language questions, assigning a confidence score to each response. A novel module in Watson Assistant called FAQ Extraction automatically generates question-and-answer documents. And AI Factsheets automatically captures key facts on a machine learning model’s performance and generates reports to “foster transparency and ensure compliance.”
According to IBM, Reading Comprehension, which was built atop a top-performing question-answering system from IBM Research, is intended to help identify more precise answers in response to queries referring to business documents. Reading Comprehension provides scores that indicate how confident the system is in each answer and is currently in beta in IBM’s AI-powered search service Watson Discovery.
As for FAQ Extraction, which is in beta in Watson Assistant’s search skill, IBM says it uses a novel natural language processing technique to automate the extraction of Q&A pairs from FAQ documents. The idea is to help businesses keep virtual assistants up to date with the latest answers while reducing the need for manual updates, IBM says.
Alongside the new services is a classification model available in Watson Assistant that is designed to improve interactions with Watson-powered virtual assistants by enabling faster training from less data, as well as more accurate results. In a related development, Watson Discovery now features support for 10 new languages, including Bosnian, Croatian, Danish, Finnish, Hebrew, Hindi, Serbian, and Swedish.
Today IBM also revealed plans to commercialize the IBM Research-developed AI FactSheets in Watson Studio in Cloud Pak for Data throughout the next year. The “factsheets,” which were first proposed in a paper published by IBM researchers in 2018, will answer questions ranging from system operation and training data to underlying algorithms, test setups and results, performance benchmarks, fairness and robustness checks, intended uses, maintenance, and retraining. FactSheets will offer:
“Like nutrition labels for foods or information sheets for appliances, factsheets for AI services would provide information about the product’s important characteristics,” Aleksandra Mojsilovic, head of AI foundations at IBM Research and an architect of AI FactSheets, told VentureBeat in an interview. “The issue of trust in AI is top of mind for IBM and many other technology developers and providers. AI-powered systems hold enormous potential to transform the way we live and work but also exhibit some vulnerabilities, such as exposure to bias, lack of explainability, and susceptibility to adversarial attacks. These issues must be addressed in order for AI services to be trusted.”
IBM also debuted Services for AI at Scale, a framework, methodology, and set of technologies intended to help customers increase the speed of AI deployment and minimize risk. Finally, the company spotlighted new capabilities in Cloud Pak for Data, including simplified administration, an improved user experience, enhanced governance and security, and the availability of OpenPages with Watson, IBM’s governance risk and compliance product.