My first contact with AI happened due to an school assignment in Taipei, Taiwan in 2020. At that time, I looked back on quite a few years of dreaming to invent a useful protein. Designing new molecular machines was my dream and just 2 years previously Frances H. Arnold had won the Nobel Prize designing new proteins. Few of these proteins were industrially produced for a few years already, and mixed into houshold appliences, such as laundry detergent to receive an imprived detergent. It was then, when a paper was published that used frontiers technology to solve a task that was initially stated several decades earlier, by another Nobel Prize winner. The paper did quite a jump in the leaderboard (CASP14 challenge) that was used to record progress respective the task. Initially, I nearly overlooked this paper for several reasons. Firstly, I was intimidated, doubting my ability to grasp the sophisticated statistical methods it utilized. Secondly, the demands of my PhD program left me swamped with work.
Fortunately, fate intervened during a class raffle where each student was assigned a research paper to review. I was selected to analyze exactly this paper, titled ‘Improved protein structure prediction using potentials from deep learning‘. It took me 3 days to read it, but finally I had understood what technology was used to bring about this landmark scientific advance.
The technology was called Neural Network and was represented in the article on page 2. I redraw it for my homework, however please excuse me if you do not understand it, however it is not curicial what I want to tell.