IBM Watson facts for kids
 |
|
| Operators | IBM |
|---|---|
| Location | Thomas J. Watson Research Center, New York, USA |
| Architecture | 2,880 POWER7 processor threads |
| Memory | 16 terabytes of RAM |
| Speed | 80 teraFLOPS |
IBM Watson is a special computer system built by IBM. It can answer questions asked in everyday language, just like you would ask a person. It was named after Thomas J. Watson, who founded IBM.
Watson became famous in 2011 when it competed on the TV quiz show Jeopardy!. It played against two of the show's greatest champions, Brad Rutter and Ken Jennings. Watson won the competition and a prize of $1 million!
In 2013, IBM announced that Watson would start helping doctors. Its first job was to assist with decisions about treating lung cancer at a hospital in New York City.
How Watson Works
Watson is a special computer system designed to answer questions. It uses advanced technologies like natural language processing (understanding human language) and machine learning (computers learning from data). It also uses information retrieval to find facts and knowledge representation to store information.
IBM explained that Watson uses many different methods to understand questions. It finds information, creates possible answers, and then checks how strong the evidence is for each answer. Finally, it picks the best answer.
Over the years, Watson has improved a lot. It can now be used through cloud computing, which means you can access its power over the internet. It also uses newer learning methods and better computer parts.
Watson's Software
Watson uses special software called IBM DeepQA. It also uses a framework called Apache UIMA, which helps manage information. The system was built using different computer languages like Java, C++, and Prolog. It runs on a type of Linux operating system and uses Apache Hadoop for big computing tasks.
Watson's Hardware
Watson is built with many powerful POWER7 processors. It uses IBM's DeepQA technology to create ideas, gather lots of evidence, and study data. The system has 90 IBM Power 750 servers. Each server has a fast 3.5GHz processor with eight cores. In total, Watson uses 2,880 processor threads and has 16 terabytes of memory (RAM).
Watson can process a huge amount of information very quickly. It can handle 500 gigabytes per second, which is like reading a million books in one second! For the Jeopardy! game, all the information was stored in Watson's RAM. This was because using regular hard drives would be too slow to compete with human players.
Watson's Data
Watson learned from many different sources. These included encyclopedias, dictionaries, and thesauruses. It also read news articles and books. Watson used special databases like DBPedia, WordNet, and Yago. The IBM team gave Watson millions of documents to help it build its knowledge base.
How Watson Plays Games
Watson breaks down questions into keywords and phrases. Then, it looks for related information. Watson's main strength is that it can quickly run hundreds of language analysis methods at the same time. The more methods that agree on an answer, the more confident Watson is that it's correct. Once it has a few possible answers, it checks them against its huge database.
Watson vs. Human Players
Watson's way of working gives it some advantages and disadvantages against human Jeopardy! players. Watson is not always great at understanding the full meaning or context of clues. However, it can read, analyze, and learn from human language. This helps it make decisions that seem human-like.
Human players often answer faster than Watson, especially for short clues. Watson is programmed not to buzz in until it is sure of its answer. But once Watson has an answer, its reaction time on the buzzer is much faster than humans. It also isn't affected by mind games from other players.
During practice games, humans had a few seconds to think while Watson processed the clue. Watson also had to decide if it was confident enough to buzz. The computer's buzzing system was very fast. It could activate the buzzer in about eight milliseconds. This made Watson faster than humans, unless a human guessed when the signal would come. When Watson buzzed in, it spoke its answer using an electronic voice. This voice was created from recordings by an actor named Jeff Woodman.
The Jeopardy! show used different ways to tell Watson and the human players when to buzz. Humans saw a light, which took them a fraction of a second to notice. Watson got an electronic signal and could buzz almost instantly. Humans tried to guess when the light would come on, but Watson's speed was usually too much to beat.
Watson's Story
How it was Developed
After its Deep Blue computer beat the chess champion Garry Kasparov in 1997, IBM looked for a new challenge. In 2004, an IBM manager saw everyone in a restaurant watching Jeopardy! because Ken Jennings was on a winning streak. This gave him the idea for IBM to try the quiz show.
In 2005, IBM decided to take on the challenge. At first, it was hard to find researchers who thought it was possible. But David Ferrucci took on the project. An earlier system could only answer about 35% of clues correctly and took several minutes. To win Jeopardy!, Watson needed to answer in just a few seconds.
In 2006, Watson was tested with 500 past Jeopardy! clues. It only got about 15% correct. By 2008, the team had improved Watson enough for it to compete with champions. By 2010, Watson could regularly beat human players.
During the game, Watson had access to 200 million pages of information. This included the entire 2011 version of Wikipedia. But it was not connected to the internet. For each clue, Watson showed its top three most likely answers on the screen. Watson was better at buzzing in than its human opponents. However, it sometimes struggled with very short clues.
Many universities and students helped IBM develop Watson. This included teams from Rensselaer Polytechnic Institute, Carnegie Mellon University, and MIT.
Jeopardy! Competition
Getting Ready
In 2008, IBM talked to the Jeopardy! producers about Watson playing against Ken Jennings and Brad Rutter. The show agreed. There were some disagreements during planning because Watson was so different from human players. IBM worried the show's writers might make clues that were hard for Watson. To fix this, clues were picked randomly from old, unaired shows.
Jeopardy! staff also worried about Watson's buzzing speed. At first, Watson buzzed electronically. But the show asked it to press a button physically, like humans. Even with a robotic "finger," Watson was still faster. Ken Jennings said, "If you're trying to win on the show, the buzzer is all."
IBM built a practice set to look like the Jeopardy! stage. They played about 100 test games, and Watson won 65% of them.
For the TV show, Watson was shown as a glowing blue globe on a screen. This globe had "threads" of thought. Ken Jennings joked that the 42 threads were a reference to Hitchhiker's Guide to the Galaxy.
Practice Game
Before the main event, a practice game was held on January 13, 2011. Watson won a 15-question round against Ken Jennings and Brad Rutter. Watson scored $4,400, Jennings had $3,400, and Rutter had $1,200.
First Match
The first part of the match aired on February 14, 2011. Watson quickly found a "Daily Double" clue, which was part of its strategy. By the end of the first round, Watson was tied with Rutter at $5,000. Jennings had $2,000.
Watson had some interesting moments. Once, it repeated an incorrect answer that Jennings had given. This happened because Watson couldn't "hear" or understand what other contestants said. Watson also made unusual bets on Daily Doubles, like $6,435 or $1,246. These bets were based on how confident Watson was in its answer.
Watson took a big lead in the second round, called Double Jeopardy! It answered both Daily Doubles correctly.
However, in the Final Jeopardy! round, Watson was the only one to miss the clue. The category was "U.S. Cities." The clue talked about a city's largest airport named after a World War II hero and its second largest named after a World War II battle. Rutter and Jennings correctly answered "Chicago." But Watson answered, "What is Toronto?????" The question marks showed it wasn't confident. Experts later explained that Watson might have struggled because "U.S. city" wasn't directly in the question. Also, there are small towns named Toronto in the U.S. Watson's second guess was Chicago, but both were very low confidence.
The first game ended with Watson at $35,734, Rutter at $10,400, and Jennings at $4,800.
Second Match
In the second match, Jennings finally got to choose a Daily Double. Watson also answered one Daily Double incorrectly for the first time. At one point, Rutter and Jennings briefly took the lead. But in the end, Watson won the second match with a score of $77,147. Jennings had $24,000 and Rutter had $21,600.
Final Results
Watson won the competition and $1 million. Ken Jennings came in second with $300,000, and Brad Rutter was third with $200,000. IBM donated all of Watson's winnings to charity. Jennings and Rutter also donated half of their winnings.
After the game, Ken Jennings wrote, "I for one welcome our new computer overlords." He also wrote that Watson might automate jobs in the future, just like robots changed factory work. He joked that "Quiz show contestant" was the first job Watson made unnecessary.
Watson's Philosophy
Some thinkers, like John Searle, argue that Watson cannot truly "think." They say Watson only moves symbols around, but doesn't understand what those symbols mean. However, others disagree with this idea.
Match Against Congress Members
On February 28, 2011, Watson played an untelevised Jeopardy! game against members of the United States House of Representatives. Watson won this match too, with a score of $40,300 against the politicians' combined score of $30,000.
IBM said that Watson's technology is a big step forward for computers. They believe it can help governments make better decisions and serve citizens.
What Watson is Used For Now and in the Future
IBM says its goal is for computers to talk and understand humans in many different areas. They want computers to understand questions and give answers that humans can easily understand and trust. IBM thinks Watson could be used for legal research, phone companies, banking, and government work.
Watson is built on IBM Power 750 servers, which are available for businesses.
In 2013, three companies started working with IBM to create apps using Watson. One app, "The North Face," helps online shoppers get advice. Another app helps people find ways to improve their health. A third app helps hospitals decide what equipment to buy.
In 2014, IBM announced it would let other software companies use Watson's technology. This means more apps and services can be built using Watson's abilities. IBM also started a $100 million fund to encourage new app development. Watson has become much faster and smaller since its Jeopardy! days. IBM wants Watson to earn $10 billion in revenue each year within ten years.
In 2017, IBM and MIT started a new research lab for artificial intelligence. They are working on projects like computer vision, understanding language, and making sure AI systems are fair and safe.
Chefs Watson
Watson even helped chefs create new poutine recipes for a food festival! It looked at what people liked in different cities and used a database of thousands of recipes to suggest new food combinations. IBM and Bon Appétit magazine also made a cooking app called Chef Watson.
Chatbots
Watson is used as a chatbot in some children's toys to help them talk and interact.
Building Codes
In 2015, an engineering company created an online service called GoFetchCode. It uses Watson to understand building codes and answer questions about them.
Teaching Assistant
IBM Watson is being used in several education projects. It has partnered with companies like Pearson Education and Apple.
With Pearson, Watson is being put into electronic textbooks. It can act as a one-on-one tutor, helping students with their reading material.
A professor at Georgia Tech used Watson to create a virtual teaching assistant named "Jill." Jill answered student questions if it was 97% sure of the correct answer. Human assistants answered the rest.
Weather Forecasting
In 2016, IBM announced it would use Watson for weather forecasting. Watson analyzes data from many weather stations and other sources to help predict the weather.
Fashion
IBM Watson helped design a dress for the fashion brand Marchesa. The dress changed color based on how people felt about it on Twitter! Watson analyzed the mood of tweets and sent signals to lights in the dress.
Tax Preparation
In 2017, the tax preparation company H&R Block started using a Watson-based program nationwide to help with taxes.
Advertising
In 2017, IBM announced that Watson would help with AI-powered advertising solutions.
Healthcare
Watson has been used in healthcare to analyze medical information. It helps doctors make decisions about diagnoses and treatments, especially for cancer. When a doctor asks Watson a question about a patient's symptoms, Watson looks at the patient's history, medical guidelines, research, and other information. Then, it gives a list of possible recommendations with a confidence score.
Even though it was designed to help with diagnoses, Watson has mostly helped with treatment options for patients who already know what they have.
In 2011, IBM partnered with Nuance Communications to develop a product using Watson's abilities to help doctors. Doctors from Columbia University and the University of Maryland helped figure out how Watson could best assist medical professionals.
In 2013, IBM and WellPoint used Watson for the first time to help with lung cancer treatment at a hospital. IBM also partnered with Cleveland Clinic to help Watson learn more about health and assist doctors.
In 2013, IBM started a project with MD Anderson Cancer Center to help fight cancer. However, this project was stopped after spending $62 million because it didn't meet its goals.
In 2016, IBM and Manipal Hospitals in India launched IBM Watson for Oncology. This helps doctors and cancer patients find personalized treatment options.
In 2017, a Japanese insurance company used IBM Watson Explorer to analyze claims and medical records. This helped them speed up payments and save money.
By 2022, IBM Watson Health was making a lot of money, but it wasn't profitable. It also faced more competition. In 2022, IBM sold its Watson Health unit to another company called Francisco Partners.
Images for kids
-
IBM Watson computer system
Error: no page names specified (help).
- Artificial intelligence
- Blue Gene
- IBM Watsonx
- Commonsense knowledge (artificial intelligence)
- Glossary of artificial intelligence
- Artificial general intelligence
- Tech companies in the New York metropolitan area
- Wolfram Alpha
