Creating new mathematical conjectures and theorems needs a complex approach which requires three factors that are:
- Intuition,
- Skill, and
- Formation of complicated but logical steps.
At DeepMind, a UK-based artificial intelligence laboratory, researchers in collaboration with mathematicians at the University of Oxford, UK, and University of Sydney, Australia, respectively. The researchers over there have made an important breakthrough by using machine learning to highlight the mathematical connections that human counterparts miss.
Into the technology behind DeepMind
In fascination with the way humans usually used to think and human-based intelligence has long caught the image of computer scientists. Human intelligence has en-sharpened the digital modern world, thus allow us to learn, create, communicate and develop by our own self-awareness.
Since 2010, researchers and developers at the DeepMind team have been trying to solve intelligence-based problems, developing problem-solving systems that are an Artificial General Intelligence (AGI).
In order to perform, DeepMind takes an interdisciplinary approach that commits machine learning and neuroscience, philosophy, mathematics, engineering, simulation, and computing infrastructure together.
The company has already made significant breakthroughs with its machine learning and AI systems, for example, the AlphaGo program, which was the first AI to beat a human professional Go player.
Thinking DeepMaths
The work developed by the DeepMind team says that mathematicians can benefit from machine learning tools to sharpen up and enhance up their intuition where complex mathematical objects and their relationships are highly concerned.
Initially, the project was focused on identifying mathematical conjectures and theorems that DeepMind’s technology could deal with, but ultimately it is all dependent upon probability as opposed to absolute certainty.
However, when dealing with large sets of information, the researchers tried to apply their own intuition that the AI could detect the signal relationships between mathematical objects. Afterward, the mathematicians could then apply their own conjecture to the relationships to make them an absolute certainty.
Tied up in Knots
Machine learning requires several amounts of data in order to complete the task efficiently and effectively. So the researchers tied knots as their starting point, calculating invariants.
DeepMind’s AI software was assumed to work on two separate components of knot theory; algebraic and geometric. The team then used the program to seek relationships between straightforward and complex correlations as well as subtle and unintuitive ones.
The leads presenting the most promising data were then directly handed over to human mathematicians for analysis and refinement.
The DeepMind team believes that mathematics can release the benefits from methodology and technology as an effective mechanism that could see the widespread application of machine learning in mathematics. Thus, this strengthens the relationship between methodology and technology.
