3 ms·

I enjoy machine learning, pure and applied maths. Where do I start? Any books, any summaries of the state of the art for the computationally inclined? I ignoran

by Dn_Ab 16y ago

I enjoy machine learning, pure and applied maths. Where do I start? Any books, any summaries of the state of the art for the computationally inclined? I ignorantly guess that graph theory, statistics and perhaps even group theory may provide useful perspectives? I know dynamical systems theory has been applied and hear you can code new simple organisms or something? Am I talking nonsense? heh This has always been an end goal of mine but I don't know where to begin. I figure start now and in 5 - 10 years I'll be ready to join you :)

possibilistic 16y ago
Graph theory is applicable for exploring biochemical pathways, and all manner of statistics, ML, and metaheuristics are useful. We can't "code" an organism, unfortunately. That won't be possible for some time. Venter has inserted a bacterial genome into a preexisting cell and called it "synthetic life", but this is hyperbole. He essentially just swapped out the operating system if you will. Both the genome and the cell were preexisting. This would be impossible to do for a eukaryotic cell with any of today's technology since we have extremely complicated molecular (genomic) machinery. Not to mention how much of our state information is epigenetic--precise methylation, etc. We certainly cannot create novel proteins outside of fusions between subunits. Protein folding is just way too difficult to solve. If we ever figure it out, society will change overnight. If you would like to study this, get some texts on cell physiology, molecular biology (genetics++), and maybe do some reading on higher level function. Like me you will see that your existing skill set is not immediately transferrable until you learn enough biology: you cannot effectively program in a language until you really know it innately. Biology is the same way, just much, much more sysemitized, complex and abstract. It has a very high barrier to entry (akin to learning Vim, but takes longer.)

by Dn_Ab 16y ago

possibilistic 16y ago

Graph theory is applicable for exploring biochemical pathways, and all manner of statistics, ML, and metaheuristics are useful. We can't "code" an organism, unfortunately. That won't be possible for some time. Venter has inserted a bacterial genome into a preexisting cell and called it "synthetic life", but this is hyperbole. He essentially just swapped out the operating system if you will. Both the genome and the cell were preexisting. This would be impossible to do for a eukaryotic cell with any of today's technology since we have extremely complicated molecular (genomic) machinery. Not to mention how much of our state information is epigenetic--precise methylation, etc. We certainly cannot create novel proteins outside of fusions between subunits. Protein folding is just way too difficult to solve. If we ever figure it out, society will change overnight. If you would like to study this, get some texts on cell physiology, molecular biology (genetics++), and maybe do some reading on higher level function. Like me you will see that your existing skill set is not immediately transferrable until you learn enough biology: you cannot effectively program in a language until you really know it innately. Biology is the same way, just much, much more sysemitized, complex and abstract. It has a very high barrier to entry (akin to learning Vim, but takes longer.)