What is IBM Watson?

Do you know the AI ​​” Watson ” developed by IBM ?

Watson won a quiz show in the United States in 2011, and in 2016, it identified the disease name of a leukemia patient and saved the life of one person. .

Today, Watson is active in many areas of society. In this article, we will introduce a wide range of Watson’s history, mechanisms, and use cases.

What is IBM Watson?

Watson is an intelligent system developed by IBM. Its name comes from IBM founder Thomas J. Watson.

Watson’s specialty is question answering, which uses human natural language to answer questions, and decision-making support, which handles huge amounts of data and presents necessary information to humans.

Watson is sometimes introduced as artificial intelligence (AI), but it is also sometimes strictly defined as “augmented intelligence”.

What is Augmented Intelligence? Difference from AI

Extended Intelligence (EI) is an intelligent system that emphasizes connections and relationships with humans in individuals and society.

The conventional view of artificial intelligence (AI) often emphasizes the different aspects of intelligent systems and machines from humans, and often talks about it in a confrontational structure such as humans vs. machines.

The term “augmented intelligence” does not represent such a view, but instead expresses the idea that the view that “intelligent systems and machines are meant to extend our knowledge and our lives” is appropriate.

Augmented intelligence brings us desired outcomes and progressive knowledge. Watson aims to operate with such a purpose.

History of development

IBM has long been very interested in the “grand challenge” in the field of artificial intelligence. A “grand challenge” is a project that, even if it is useless in itself, is expected to bring about technological progress for society as a whole by attracting a great deal of attention from the public through its realization. That’s it.

One of IBM’s goals in developing Watson was to win the American quiz show Jeopardy!

Made famous by an American quiz show

In February 2011, Watson defeated two previous Jeopardy! quiz show champions. This quiz show not only asks for knowledge in various fields, but also plays on words and tests your sense of inspiration.

For example, the answer to the question ” Colorful fourteenth century plague that became a hit play by Arthur Miller.” Black Death of a Salesman “. Computers are generally not good at answering these kinds of questions. A human sense of humor is difficult to process mechanically.

However, Watson can select questions to answer according to the degree of conviction and strategically aim for victory in the program. Watson ended up beating all-time champions in simulations about 70% of the time.

What IBM Watson Can Do

Watson is generally introduced as a type of AI, but IBM refers to Watson as a ” cognitive computing system.”

Three Characteristics of Cognitive Computing Systems

General AI aims at technological development and innovation, but the main purpose of cognitive computing systems is to “assist human ability.” This is similar to the idea of ​​”augmented intelligence” mentioned earlier.

Cognitive computing systems are particularly interested in three technologies:

natural language processing

The language that humans use on a daily basis is called “natural language”.

The processing of human natural language by a computer, and the simulation of the meaning that humans have for natural language on a computer, such a process is called natural language processing.

Hypothesis presentation

Processing the obtained information and presenting hypotheses based on it is a process that is highly valued, for example, in the field of business.

Cognitive computing systems do this with massive amounts of data. As a result, it is possible to present new hypotheses that are not influenced by human prejudices (bias) and hypotheses with few errors.

study

The learning process is the core technology of AI and cognitive computing systems. New research is being conducted on learning mechanisms and algorithms every day.

 IBM Watson APIs

Watson has various Japanese APIs (Application Programming Interfaces). Here are some excerpts from them.

question answering system

It is an API that “This is what Watson says”. Natural language processing is performed on the entered question and the answer is presented.

Real-time translation function

“Watson Language Translator” can automatically detect the source language and translate news, conversations, patent documents, etc. according to the application.

Analysis of personality and emotions

“Watson Tone Analyzer” is an API that can detect emotional and stylistic tones from text .

Just specify the type of content first, such as tweets or emails, then copy and paste the text you want to analyze and it will detect the tone of that text.

Image/speech recognition

“Watson Visual Recognition” is an API for image recognition. The feature of this API is that anyone can easily learn image data with easy-to-understand GUI operations.

“Watson Speech to Text” is an API for speech recognition. This speech recognition API can transcribe multilingual speech to text. You can also add special vocabulary and phrases.

This API can be used in situations where machine operation and communication through voice is required, such as transcription of meeting minutes, voice operation of machines, and real-time audio subtitle display.

speech synthesis

“Watson Text to Speech” is an API that converts text to speech. It supports various languages ​​and voices.

The text-to-speech API allows you to interact with the other party based on the text. For example, by introducing Watson to a call center and providing voice guidance for appropriate responses, it will be possible to significantly reduce the waiting time for customer calls.

How to use IBM Watson

IBM Watson can be used by individuals as well as businesses. Watson is known for its low implementation effort and cost.

environment 

I would like to introduce an environment called “ IBM Watson Studio”.

Watson Studio is a full-featured integrated development environment for building AI on any cloud. With an easy-to-understand API, even people who are not familiar with AI can easily use Watson’s functions.

In addition, since various technologies used in conventional AI development can be used on Watson Studio, it can be said that Watson Studio is an attractive environment for experienced people and experts.

For example, you can use AI frameworks such as PyTorch, TensorFlow, and scikit -learn , development environments that are compatible with AI development such as Jupyter notebooks, and programming languages ​​such as Python , R, and Scala.

What is the installation cost?

Watson Studio offers various pricing plans. It should be noted that the individual introductory plan is available for free.

Low hurdles for introduction

In general, it is said that a large data set ( big data ) and specialized knowledge for its learning are required to introduce AI. A service called ‘Trained Watson’ announced by IBM has made the hurdles to introduce Watson considerably lower.

IBM works with users and companies across a wide range of industries and occupations to develop a portfolio of industry-specific pre-trained Watsons. By using this, you can easily introduce Watson without accumulating a large amount of learning data or having specialized machine learning skills.

 IBM Watson case study

Let’s take a look at a real-world use case for Watson.

Accurately identify the type of leukemia

The news was reported in 2016 that ‘Watson hit the type of leukemia and proposed an appropriate treatment method’.

The University of Tokyo and IBM were conducting research to make Watson learn cancer research papers and use them for diagnosis.

In that study, Watson correctly identified a disease in 10 minutes based on the input of genetic information for a patient who had a poor course of treatment . He then suggested changing the treatment method, and the clinical team who received the proposal administered another anticancer drug, which improved the patient’s condition.

This news sparked a lot of discussion about the potential of AI to contribute to medicine and other areas of society.

Widespread use cases in Japan

Watson is currently being used in a variety of fields.

As mentioned earlier, Watson was developed under the design of a cognitive computing system, so its main use cases are aimed at extending human capabilities or contributing to human society. .

Customer support

Watson can be used in the field of customer support by using it as a customer service AI and chatbot. IBM offers a dedicated customer-facing plan called Watson Assistant.

Watson’s customer support service is highly versatile and can be used for a variety of purposes. For example, it is a service that identifies the cause of trouble by answering questions and voice guidance at call centers.

As a specific example of using Watson, we will introduce the virtual assistant “Makana-chan” provided by JAL. Makana-chan is a service that specializes in assisting trips to Hawaii, and through conversation, users can receive customized travel information and plans.

 Safety diagnosis

In July 2021, it was reported that Mitsui Chemicals and IBM will operate Watson-powered industrial accident hazard extraction AI at factories. Based on information on past occupational accidents and troubles, it is possible to inquire about cases with high risk correlation and quickly extract similar cases regarding the work that employees are trying to do .

Business efficiency

Meiji Yasuda Life has introduced Watson in its “Customer Voice System,” a system that centrally manages customer inquiries and complaints.

According to the announcement in March 2020, by having Watson classify and analyze big data such as text data managed by the “Customer Voice System”, we have realized operational efficiency improvement of about 1500 hours per year.

Watson has an easy-to-use UI (user interface) and AI-based work support, so people without specialized knowledge (end users) can use Watson themselves.

in conclusion

Watson ‘s main goals were to “assist human ability” and “contribute to society.”

Through the development of Watson, IBM is interested not only in the technological development of AI, but also in what kind of benefits AI gives us, and how we can coexist with AI in society. I am sending.

The use of AI in companies and society is likely to be a field that will receive more and more attention in the future.