Skip to content Skip to sidebar Skip to footer

Source: cnet as of 10-03-2020

In 2018, IBM debuted technology called IBM Debater that used artificial intelligence technology to read lots of documents and take on a human in a competitive debate about issues like whether we should subsidize preschools. In 2019, Big Blue pitted IBM Debater against a champion debater in a live-streamed competition.

And now you can use IBM Debater technology to find out if people like you on Twitter. That’s because IBM has begun commercializing aspects of the IBM Debater technology, the company announced Wednesday.

It’s an illustration of how today’s AI, in particular technologies like neural networks and deep learning, can be used to help computers better navigate the messy human world. Computers are very good with ordered data, like spreadsheets with consistently formatted numbers in carefully labeled columns. But AI has worked wonders for computing tasks with untidy information — a self-driving car recognizing its surroundings or a phone understanding human voice commands, for example. That’s why companies like IBM, Microsoft, Google, Amazon and countless others are trying to cash in on the technology.

You can use the Debater technology in three ways:

  • For sentiment analysis, in which IBM’s software judges how positively or negatively people feel about a subject, based on natural speech and writing. It’s smart enough to tell that being called “hardly helpful” is not a compliment.
  • For summarizing information. For example, IBM used it to digest information in 18 million blog posts, articles and biographies to then generate information about celebrities on the Grammy competition red carpet.
  • For “clustering,” a process of analyzing raw data to figure out when topics are related to each other. That should be useful for insurance, health care and manufacturing, IBM believes.

The Debater technology grew out of IBM Research but now is a part of Big Blue’s Watson service. One customer is ESPN Fantasy Football, which is using it to analyze football data like blog posts, podcasts and tweets to offer tips to Fantasy Football players.

Show CommentsClose Comments

Leave a comment

News ORS © 2020. All Rights Reserved.