
Jill Watson, an assistant professor of online courses at the Georgia Institute of Technology, before she was “exposed,” had helped more than 300 American graduate students to work on program design projects. Her competence included a rather wide range of tasks - from sending questions and reminders about the deadlines for completing assignments - to actively participating in the discussion of sensitive issues in specialized student forums where students exchange views and knowledge related to term papers and other projects.
Since the communication within the framework of the experiment was exclusively online, none of the students could see Jill personally, however, due to the way of communication and the slang used, most of the student audience represented her as a sociable friendly girl with a sense of humor for more than twenty years actively working on her doctoral dissertation. And practically none of the respondents until the last moment had imagined that they were communicating with a robotic complex. Meanwhile, a hint of the origin of Jill was hidden in her last name - Watson, a familiar to us all analytical platform developed by IBM and refined in accordance with the peculiarities of the learning process in the laboratory of the Georgia Institute.
According to the journalistic survey, there were quite a few students who expressed a desire to vote for Jill as one of the best course assistants. Here, the institute administration had to reveal to graduate students the true origin of the assistant, Jill Watson. When the truth became the property of the student community, some of the graduate students nevertheless agreed that there were some doubts about Jill - she answered the questions in the messages too quickly and accurately, regardless of their complexity.
')
The popularity of online education, which provides millions of interested applicants with the opportunity to get high-quality, and most importantly, affordable education, is growing every year, as the audience is wishing to learn online without leaving home. But what is good for students is for teachers at a certain stage that it becomes a problem in connection with the avalanche of questions asked by students and requiring prompt response. It is estimated that on average for a semester, an assistant teacher has to answer more than 10,000 (!) Questions in text messages. And here the transfer of routine functions to the robot is seen as the best way out, which was convincingly proved by the example of Jill Watson.
Of course, it’s still difficult to call the "robots" of Miss Watson an ideal, as well as alternative robotic solutions that deal with similar tasks. Nevertheless, the specialists of the analytical department of the institute calculated that Jill was able to answer 40% of routine questions, while leaving the competence of the “living” teacher to previously ambiguous questions with philosophical overtones and those that were associated with an individual personal assessment.
IBM is aware of the development of a training robot based on their software, but the company’s specialists themselves did not participate in the work of the technologists at the Technological Institute. Developing software algorithms for Jill, the institute's experts analyzed a total of about 40,000 posts on the Piazza forum and trained the system to classify questions. At the same time, as the creators of Jill Watson mention, their system is significantly ahead of the “development” of the existing training chat bots and generates a response only when it is not less than 97% “confident” of its truth and correctness of classification.
The experiment in a slightly modified format will continue next year. This time, students will be warned that one of the robot assistants, but no more. The final verdict they still have to make their own.
If the approved technology takes root in the educational sphere, the effect of its application will be difficult to overestimate. Already today it is clear that these are, first and foremost, new opportunities for the exponential expansion of the online audience and more than affordable prices.
Given the rapid development of AI systems, primarily related to the processing and analysis of large amounts of information, as well as the need to form statistical conclusions based on them, it is not difficult to suggest a further expansion of their application. Already today, Watson's intelligence is involved in a number of cancer centers in the process of cancer diagnostics, where reconciliation of the enormous amount of archival information with that provided for the diagnosis of a specific patient is required. And already today, diagnostic results prepared using AI are often more accurate than those obtained in the traditional way, using the knowledge and experience of narrow specialists.
Another promising area where the enormous potential of AI systems is visible is the judicial system. An example would be such a complex area as case law, where, in order to make a court decision, in addition to the dry rules of the legislative framework, one must be guided by a huge array of accumulated experience. Using the capabilities of AI in this case will completely exclude the biasedness of the court verdict, often associated with the corruption component or personal attitude to the accused by the judge or the panel of judges.
Opening access to the capabilities of the Watson platform, IBM thereby launched a non-stop process of expanding the presence of AI in various areas of modern human life. One of the latest and most remarkable examples was
Connie , a robot concierge who successfully coped with instructing visitors arriving at the Hilton Hotel in Virginia (USA).

That's all, with you there was a simple service for choosing sophisticated
Dronk.Ru equipment. Do not forget to subscribe to our
blog , there will be many more interesting ...
LetyShops cashback service
sponsor . Return money for any purchases on the Internet. Read more about what a cashback service is in our article.
We choose a cashback service for the 6th anniversary of Aliexpress.