Google says it has made improvements to Google Assistant that bolster the company’s artificial intelligence assistant’s natural language capabilities, including correctly understanding and pronouncing unique names.
Starting over the next few days, users can teach Google Assistant to enunciate and recognize names of contacts the way they pronounce them. Without keeping a record of the user’s voice, Assistant will listen to the pronunciation and remember it.
“This means Assistant will be able to better understand you when you say those names, and also be able to pronounce them correctly,” the company said in a blog post. “The feature will be available in English and we hope to expand to more languages soon.”
The company also said it fully rebuilt Assistant’s NLU models so it can more accurately understand context while also improving its reference resolution to know what users are trying to do with a command.
According to Google, the upgrade uses machine learning technology powered by state-of-the-art Bert, a technology Google invented in 2018 and first used in the company’s search engine to process words in relation to all the other words in a sentence.
“Because of these improvements, Assistant can now respond nearly 100 percent accurately to alarms and timer tasks,” Google wrote in the blog. “And over time, we’ll bring this capability to other use cases, so Assistant can learn to better understand you.”
The company also applied BERT to improve the quality of conversations with Assistant, including using previous interactions to understand what’s currently being displayed on a smartphone or smart display to respond to follow-up questions.
For example, if a user is having a conversation with Assistant about Miami and asks about the best beaches, Assistant will know the user meant specifically in Miami.
“Assistant can also understand questions that are referring to what you’re looking at on your smartphone or tablet screen, like [who built the first one] or queries that look incomplete like [when] or [from its construction],” the company said.
“There’s a lot of work to be done, and we look forward to continue advancing our conversational AI capabilities as we move toward more natural, fluid voice interactions that truly make everyday a little easier.”