How will Artificial Intelligence (AI) change the way scientists hypothesize., conduct their research, choose their methods, work with their collected data, interpret their data, and draw conclusions from their data?
Dennis
Dennis Mazur
This question should trigger a few graduate student projects. AI means many things to many people. So, I will word my answer carefully.
The use of massive shared data in machine readable form and analysis of selected data by computer based systems under human guidance should play a major role in the early stages of research. If you think of an interesting hypothesis and could have quick feedback on it, you will make better decisions on how much of your life you spend on that hypothesis. A number of "advances in AI" have been based, in fact, on using large collections of data. There are pretty impressive collections of scientific data accessible to researchers over the Web. For some detail, visit
https://www.nature.com/sdata/policies/repositories
However, it is difficult to visualize scientific data being used easily in the absence of good skills in using domain specific data and having tools to handle domain specific formats. This creates opportunities for inter-disciplinary researchers to implement hypothesis testers that can be run over the Internet and give quick and dirty feedback on a given hypothesis.
Will this make graduate students redundant? No, I believe that graduate students are smart enough to move one notch up the research ladder instead of slogging for months on their advisors' half-baked ideas! Leave that to AI!
As I was not thinking about answering some questions in Research Gate 10 years back, I can not imagine the conditions of 2027 today.
Development of AI may force us to accept many changes in many aspects of life. Research pattern may not be outside the list.
hi?
There are pretty impressive collections of scientific data accessible to researchers over the Web. For some detail, visit
https://www.nature.com/sdata/policies/repositories
However, it is difficult to visualize scientific data being used easily in the absence of good skills in using domain specific data and having tools to handle domain specific formats. This creates opportunities for inter-disciplinary researchers to implement hypothesis testers that can be run over the Internet and give quick and dirty feedback on a given hypothesis.
Obviously many things like the ones mentioned above would be automated, which would facilitate on a large scale the work of a researcher.
How will AI change the way scientists hypothesize., conduct their research, choose their methods, work with their collected data, interpret their data, and draw conclusions from their data?
This is a question already for 60 years. At the moment, AI has not real breakthrough in research, in particular in theoretical research.
AI is hot in market, because big data can be used to develop AI products and can make money dramatically.
More generally, ICT, AI, big data analytics, and digital transformationhave changed our research, research methods, and data collection methods.
This can be considered as big data driven AI. However, AI should research "Intelligence without data" Its impact on our research methods, research and development will be more significant. For detail, please see my article with the same title.
Prof. Dr Zhaohao Sun
2017-12-12
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence