A.I

Discussion in 'Biology & Genetics' started by Ethernos 1997, Aug 31, 2020.

  1. Ethernos 1997 Registered Senior Member

    Messages:
    230
    I came across this awsome tech a.i. learning in which a person describes how a.i. can beat pro gamers after learning over and over again. I thought to myself if protein folding were made to be a game can a.i. solve it too. What are your views on this?
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Michael 345 New year. PRESENT is 70 years old Valued Senior Member

    Messages:
    9,709
    Interesting

    However, to me, it appears that with games the purpose is to win. Games have a end point

    A powerful learning AI would be able to store a larger number of end points for a particular game than a human gamer. Not a given but likely

    Protein folding computer simulations are being used, although it seems not in a learning game situation

    Nowadays, researchers predict the structure of a protein by inputting the amino acid sequence into a computer. The advanced technology and modeling software allow scientists and researchers to form a predicted structure. However, the structure is not accurate, as there is always a small degree of errors present. Nevertheless, this can speed up discovery of new medications since the digital structure can be manipulated.

    https://en.m.wikibooks.org/wiki/Structural_Biochemistry/Proteins/Protein_Folding_Problem

    I may be off base but these situations appear to being worked backwards

    Scientists KNOW the result, just require to understand how the end point was reached ie reverse the folding and find what initiated each fold

    The problem stated in the extract is
    • the structure is not accurate and
    • there is always a small degree of errors present
    Can these be overcome by a learning AI? or just putting in more precise details?

    Sorry don't know

    Again interesting. It would be much more than interesting to plug in the parameters and target protein and find a learning AI finds more than one fold configuration produces same target protein

    Please Register or Log in to view the hidden image!

     
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. Ethernos 1997 Registered Senior Member

    Messages:
    230
    T
    Q for the answers
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Michael 345 New year. PRESENT is 70 years old Valued Senior Member

    Messages:
    9,709
    ? ? ?

    Please Register or Log in to view the hidden image!

     
  8. C C Consular Corps - "the backbone of diplomacy" Valued Senior Member

    Messages:
    2,079
    Sure, much if not all study territory is vulnerable to it over time, as AI progresses. Machine learning is producing solutions, discoveries, and successful autonomous robot behavior without need of traditional human background conceptions and laws/principles/methods.

    Via searching for patterns in big data and trial and error interactions with the environment, after being assigned goals. Deriving its own unconventional or "alien" understandings and behavior approaches in the form of statistical probability constructs. The drawback being that it is very difficult in traditional contexts for researchers to grasp/analyze how such machine learning schemes, predictions and remedies actually work and succeed, and thereby where any potential flaws and errors might rest in them.

    Robot chemist that thinks in 10 dimensions makes new discovery
    https://eurekalert.org/pub_releases/2020...070620.php

    INTRO: Researchers at the University of Liverpool have built an intelligent mobile robot scientist that can work 24-7, carrying out experiments by itself. The robot scientist, the first of its kind, makes its own decisions about which chemistry experiments to perform next, and has already discovered a new catalyst. It has humanoid dimensions and works in a standard laboratory, using instruments much like a human researcher does.
     
    Ethernos 1997 likes this.

Share This Page