In this photo taken on January 13, 2011, IBM's Watson computer defeated program champions Ken Jennings (left) and Brad Rutter (right) on the “Jeopardy!†quiz show (Photo by Associated Press/Syz Wenig )
In 2004, when Charles Lickel was eating dinner with his colleagues, he found that all his customers suddenly rushed to the bar. Out of curiosity, he followed behind and was ready to find out. It turned out that they were all witnesses to Ken Jennings' historic six-month undefeated moment on the “Jeopardy!†quiz!
He was very shocked. Paul Horn, director of the IBM Research Center at the time, was urging Lickel to come up with ideas for the company's next "big challenge," trying to see if it could solve the extremely difficult "Blue Giant" tradition. The last time it attracted widespread attention was in 1996 when IBM's "dark blue" computer defeated chess master Garry Kasparov.
The rest, as we all know, has become history. Seven years later, in 2011, IBM's Watson defeated Jennings and another champion of "Jeopardy!" Brad Rutter. Today, Watson is no longer just a clever question-answering machine. For IBM, it is already a potentially huge business opportunity. CEO Ginni Rometty hopes that it can become the core of the Blue Giant's future plans. However, there are still many challenges before achieving the goals.
A brief history of artificial intelligenceThe field of artificial intelligence began at the Dartmouth Conference in 1956, and optimism rose, and it is believed that machines will replace humans within 20 years. However, everything did not happen. By the 1970s, funds were gradually depleted, and technological development also entered a period known as the "AI winter."
Slowly, certain areas gradually made progress. By 1992, the heat of artificial intelligence had recovered. The U.S. government began hosting a series of conferences and presented many challenges for the Q&A or "QA" system. IBM participated in these meetings and began to take a lead on certain technologies.
First, the researchers passed a test rule system, similar to Doug Lenat's Cyc project, to answer questions based on information provided by human experts, just like the encyclopedia. However, they soon discovered that the scale of such systems cannot exceed a certain value.
As a result, they began to explore various technologies to come closer to the human brain to process information and make decisions. For example, deep semantic analysis techniques break sentences into language elements, and support vector machines can mine large amounts of data, learn from them and get usage conclusions.
However, these are independent projects made by different teams. What Lickel saw at the bar that night was an opportunity to organize this into a coherent system. “The Jeopardy challenge has allowed us to focus on all the different things that are going on and focus on seeing if we can solve a really big problem,†said Eric Brown. He is a member of the project team and is currently the head of the Watson department.
Solve the Jeopardy Grand Challenge"Jeopardy!" presents a unique challenge for artificial intelligence systems. First, it covers a very wide range of topics, so it is not enough to just train the system to operate in a single area. The clues are also very complex languages, including puns and literary backgrounds, which often make the real issues obscure.
E.g:
Question 1: "Difficult moments" (pun: “Hard times†is also the name of Dickens's work), indeed! On February 7, 1812, a huge earthquake occurred in New Delhi, and the writer also shocked England (puns: "struck England" refers to Dickens born in the United Kingdom).
Question 2: According to CS Lewis, it is between the eastern part of the Eastern Ocean and the northern part of the Shribble river.
To answer the first question correctly, "Who is Charles Dickens? (Charles Dickens, Charles Dickens, British 19th-century critical novelist)", you must realize that "struck England" refers to the date of birth, "Hard times" refers to one of Dickens's works.
To answer the second question correctly, "What is Narnia? (Note: Narnia is the kingdom of Narnia in the British writer CS Lewis's Children's Literature classic The Chronicles of Narnia.) You must realize that this is a fictional The geographical name does not refer to an actual location.
Other aspects of this game are also more difficult. For example, the wrong answer will be punished, so not only should there be a feasible answer, but at the same time it is right to measure how confident it is. There are also time constraints that need to be able to respond within a few seconds.
So, from the 13 researchers that began in 2007 to the final 25 staff members, the Watson team designed and built a process that can handle the process and quickly analyze the data architecture to achieve a competitive level. In the next four years, they not only needed to solve complex technical problems, but also changed the way they collaborated with each other.
"In order to develop Watson, we have to adopt flexible technologies," Brown told me. "This is a brand new thing for us as researchers and scientists. We don't just build a system, we must also develop hundreds of algorithms. Individuals must become experts in different fields, and everyone must bring different methods to solve problems. Then we must establish other systems to measure the different opinions generated within the system.
Despite the difficulties, Watson not only won Jeopardy! It used three rounds to beat the human player. In response to his final reaction, Jennings wrote, "I want to say, welcome to our new computer dominator."
Establishing a business model for WatsonAfter winning the "Jeopardy!" victory, IBM pushed Watson to the market. The first commercial application was to work with the Memorial Sloan Kettering Cancer Research Center and Concord Corporation to design an advisory system for its medical staff. Since then, the system has been deployed through Watson Healthundefined to many top medical institutions such as Cleveland Clinic and MD Anderson Cancer Center.
However, the company really sees a huge opportunity to use Watson as a service. Other companies and developers can develop their own applications through the API. “We saw Watson as a smart engine and built solutions for our partners to better serve their customers,†Jonas Nwuke, IBM Watson platform manager told me.
So far, the project has attracted more than 550 partners, including Satisfi who uses Watson to help customers browse retail stores, Online Shopping Assistant Fluid, and Smart Travel Guide Wayblazer. Developers access Watson's APIs for meter charges, so they pay only for the services they use, which is ideal for project launches.
The difference between these applications and the regular recommendation engine is twofold. First, they can analyze unstructured data such as product descriptions and user reviews. Second, they can learn user preferences. So, for example, if a hotel gets poor rankings because of guests complaining about children clamoring, but I'm looking for places where children can be arranged, Watson can help me make the right choices.
Growing WatsonLike every precocious genius boy, Watson now needs to find his place in the world. It will not completely replace the role of human experts, as many suspect. Because of some things, machines can almost never be done. Like real compassion needs to be based on understanding, interacting with people and establishing effective working relationships.
Despite this, its potential is undeniable. Think of an ordinary doctor and how efficient it would be if he had an assistant like Watson. Even before patients enter the room, it can analyze their personal medical history, which is often hundreds of pages. It can then compare it with the records of 700,000 academic papers published each year and potentially millions of other patients.
All of this, without question, goes beyond the capabilities of human doctors. Watson usually only takes a few minutes to prepare each exam. Therefore, it would be of great benefit to consult Watson with the case. At the same time, once the doctor provides feedback that Watson's advice is useful, the system will continue to learn because it also applies to tourists, shoppers and others.
Therefore, although robots cannot be our rulers, they do have the potential to become extremely valuable collaborators. “One of the ultimate goals of the system is to overcome human prejudice,†Brown told me. “So, Watson can not only give the answer, but in some cases, it can also raise questions to the traditional human wisdom.â€
The potential for such cooperation is enormous. Doctors can consult the system based on various intentions and purposes, and immediately obtain all the human medical knowledge, so that they can spend more time really caring for the patient. They can become therapists again, not just technicians.
I want to say, welcome to our new robot partner.
Via Forbes
Common Mode Coil For Automotive Product,Magntic Inductor Coil,Emi Common Mode Choke,Hemi Coil Pack Conversion
IHUA INDUSTRIES CO.,LTD. , https://www.ihua-magnetics.com