''' ARTIFICIAL
INTELLIGENCE
ARMISTICE '''
IN 2019 - A GROUP OF RESEARCHERS LED BY VAHE TSHITOYAN, then at Lawrence Berkeley National Laboratory, in America, used an AI technique called unsupervised learning to analyse the abstracts of materials science papers........
And extract information about the properties of different materials into mathematical representation called ''word embeddings''. These place concepts into a multi-dimensional space where similar concepts are grouped together.
The system thereby gained a ''chemical intuition'' so that it could, for example, suggest materials with similar properties to another material.
The AI was then asked to suggest materials that might have thermoelectric properties [ the ability to turn a temperature difference into an electrical voltage, and vice versa], even though they were not identified as such in literature.
The ten most promising materials were selected, and experimental testing found that all ten did indeed display unusually strong thermoelectric properties.
Academic journals and laboratories revolutionized how science worked in the past. Could artificial intelligence do the same in the future?
'' BY AMPLIFYING HUMAN INTELLIGENCE -AI may cause a new Renaissance, perhaps a new phase of the Enlightenment,'' Yann LeCun, one of the godfathers of modern artificial intelligence [AI], suggested earlier this year.
AI can already make some existing scientific processes faster and more efficient but it can do more, by transforming the way science itself is done?
Such transformations have happened before. With the emergence of the scientific method in the 17th century, researchers came to trust experimental observations, and the theories they derived from them, over the received wisdom of antiquity.
This process was, crucially, supported by the advent of scientific journals, which let researchers share their findings, both to claim priority and to encourage others to replicate and build on their results. Journals created an international scientific community around a shared body of knowledge, causing a surge in discovery known today as the scientific revolution.
A further transformation began in the late 19th century, with the establishment of research laboratories - factories of innovation where ideas, people and materials could be combined on an industrial scale.
This led to a further outpouring of innovation, from chemicals and semiconductors to pharmaceuticals.
These shifts did more than just increase scientific productivity. They also transformed science itself, opening up new realms of research and discovery. How might AI do something similar, not just generating new results, but new ways to generate new results?
A promising approach is '' literature based discovery '' [LBD] which, as its name suggests, aims to make new discoveries by analysing scientific literature. The first LBD system, built by Don Swanson at the University of Chicago in the 1980s, looked for novel connections in MEDLINE, a database of medical journals.
In an early success, it put together two separate observations - that Raynaud's disease, a circulatory disorder, was related to blood viscosity - and suggested that fish oil reduced blood viscosity - and suggested that fish oil might therefore be a useful treatment. This hypothesis was then experimentally verified.
We're charging our battery : But Dr. Swanson's LBD system failed to catch on outside the AI community at the time.
Today AI systems have become far more capable at natural-language processing and have a much larger corpus of scientific literature to chew on. Interest in LBD-style approaches is now growing in other fields, notable materials science.
Exhilarated, the researchers then retrained their system, omitting papers from more recent years, and asked it to predict which new thermoelectric materials would be discovered in those later years.
The system was eight times more accurate at predicting such discoveries than would be expected by chance alone.
It could also make accurate discovery predictions using other terms, such as ''photovoltaic''. The researchers concluded that ''such language based inference methods can become an entirely new field of research at the intersection between natural-language processing and science.''
The Honour and Serving of the Latest Global Operational Research in Artificial Intelligence, Science and the Future Revolution continues. The World Students Society thanks The Economist.
With most loving and respectful dedication to The Global Founder Framers of !WOW! - and then Mankind, Students, Professors and Teachers of the world.
See You all prepare for Great Global Elections on The World Students Society : wssciw.blogspot.com and Twitter X !E-WOW! - The Ecosystem 2011 :
Good Night and God Bless
SAM Daily Times - the Voice of the Voiceless
0 comments:
Post a Comment
Grace A Comment!