We dey live inside three-dimensional space.
Inside dis space, we dey see three-dimensional space with only wetin our eyes show us, wey be like flat pictures wey get only two dimensions.
Dis one mean say we get like three-dimensional picture of space inside our head, and we dey take the two-dimensional picture wey our eye see, come turn am back to fit dat three-dimensional picture for our mind.
I dey believe say if we use dis idea, man fit come dey see four-dimensional space. Even if we no fit make four-dimensional space or four-dimensional things for real life,
we fit make dem for computer. If we take wetin we make for computer (four-dimensional space) come turn am to flat picture (two-dimensional), man fit see and understand am.
Then, as man dey learn how dis four-dimensional space and things dey behave and look, dem go fit come build one four-dimensional space for inside their mind.
But dis one just be possibility, and to learn am go take plenty time.
Apart from dat, even if person come sabi see four-dimensional space, e no go get plenty place to use am.
How AI Go See Four Dimensions
For di other side, AI fit do di same thing. In fact, AI fit even use dis four-dimensional space sense well-well.
For example, if AI get four-dimensional space sense, e go fit draw and understand four-dimensional graphs.
Man eye only fit see and understand flat two-dimensional pictures well. So, even if dem draw three-dimensional graph and we try understand am with our mind, some parts go dey hidden, we no go see am.
Even with three-dimensional graphs, plenty parts dey hide, and for four-dimensional graphs, even more data go dey hidden.
Even if you rotate di graph to see di hidden parts, dat one no go help you understand di data fully and quick-quick at once.
But AI no need to follow di rule of two-dimensional flat pictures. Dem fit give AI like three-dimensional or four-dimensional eye for computer and train am.
If dem do so, AI go fit see three-dimensional and four-dimensional graphs finish, straight up, no data go hide, and e no go need to rotate am.
Wetin more, dis one no just stop for four dimensions; logically, dem fit increase di dimensions as much as dem want, five, ten, twenty, and even pass dat.
How To Understand Plenty-Dimension Graphs
If you fit see graphs well-well, e go make you fit check trends wey dey across plenty dimensions. You fit also compare how big or small things be and understand di proportion quick-quick.
Wetin more, e go allow you check how data dey behave, like data wey dey similar or resemble each other. E fit even make you see patterns and rules.
Dis one go make you understand data well-well pass just matching multi-dimensional data patterns, wey current AI sabi do well.
For example, even if some parts get di same pattern inside different-different combinations of dimensions, e go hard to find dem if you just do simple multi-dimensional pattern matching.
But, if you look data with plenty-dimension eye, shapes wey resemble go show face sharp-sharp, even if dem dey for different combinations of dimensions.
Wetin more, apart from just using di dimensions wey dey for di data you put in, you fit explore dimensional structures wey go make you understand di data well. You fit do dis by making specific axes big or small, or by turning dem to logarithms, or by mapping plenty axes to another set of axes with di same number, without making di dimensions less.
So, when you train multi-dimensional vision, e go open way to understand data structures well-well, wey bin dey hard for both man and old AI. Dis one fit lead to finding new things and laws.
Making New Discoveries Faster
When you fit understand plenty-dimension data straight, without making am small-small go less dimensions, e get serious power.
For example, dem invent di heliocentric theory so dat dem fit put di things dem observe for sky into maths formulas wey easy to understand. Di geocentric theory, wey talk say sun dey go round earth, no fit make di observation data enter formulas wey easy to understand, so dat one bring out di heliocentric theory.
But, if dem fit understand di astronomical observation data straight up, without making its dimensions small, e possible say dem for don discover laws like heliocentric theory long time ago.
Like dat, scientific inventions like di theory of relativity and quantum mechanics for don happen fast-fast if dem fit understand multi-dimensional data well-well for its original dimensions.
Dis one mean say AI wey sabi multi-dimensional things straight fit make new discoveries happen faster, and dat fit lead to finding plenty theories and laws wey man never even know about.
Final Talk
AI wey dem train to understand plenty-dimension spaces like say e dey born inside dem fit use its power to see plenty dimensions, wey human beings no fit do, to quick-quick expand how science and school things dey work.
Paradigms dey tend to just add up, instead of just changing. Even if dem invent new paradigms, we no must follow all of dem.
Of course, AI go fit explain complex, plenty-dimension paradigms by turning dem to less dimensions for us to understand easy.
Even with dat, e fit be say human beings no go fit fully understand paradigms wey too plenty for dimensions. And we no go fit understand all di plenty paradigms wey don expand.
If dat happen, we fit come dey live inside products and systems wey dey work well, even if we no fully understand how dem take work.
As engineer, I no really want imagine dat kind situation, but for many people, e fit no too different from how things dey now.