Machine learning in trading: theory, models, practice and algo-trading - page 1706

 

A little hack, maybe someone will find it useful.

The task: to draw a lot of circles of different size and color on the graph, and not to die from coding and not to go crazy.

Solution:
The wrong one is to use the ellipse object invented for this purpose, with three anchor points, fill and so on.
The correct one is to draw a line of zero length, the thickness of the line will be the diameter of the circle, the color will be the fill, and the center will be the anchor point.

 
ИИ нашел самый мощный антибиотик. У него ушло на это несколько часов
ИИ нашел самый мощный антибиотик. У него ушло на это несколько часов
  • 2020.02.21
  • Хайтек+
  • zen.yandex.ru
С помощью специального алгоритма машинного обучения американским ученым удалось выявить вещество, способное справиться со многими лекарственно-устойчивыми бактериями, включая возбудитель туберкулеза. Соединение назвали халицин - в честь ИИ из фильма "Космическая одиссея 2001 года". Оно уже доказало свою эффективность, вылечив лабораторных мышей...
 
It looks like a fake. How did they gather and describe 107 million chemical compounds for AI? What a load of crap... How without empirical sampling could the AI select the effective variants?

We had to tailor a special chemical compound for each resistant organism. There were dozens of organisms, 107 million compounds. And a selection with no hands-on experience. Exactly a fake...
 
Konow's tag:
Sounds like a fake. How did they collect and describe 107 million chemical compounds for AI? That doesn't make any sense... How without empirical sampling could the AI select the effective choices?

We had to tailor a special chemical compound for each resistant organism. There were dozens of organisms, 107 million compounds. And a selection with no hands-on experience. That's a fake...
The point of AI is exactly that it doesn't matter how big the data area is, it's important whether it's finite or infinite. If the data field is finite, then it is within the power of AI and the size of this field is not important. Well, that's just... for the record...
 
Mihail Marchukajtes:
The trick of AI is exactly in the fact that it doesn't matter how big the data range is, it is important whether it is finite or infinite. If the data range is finite then it's possible to do this using AI and the size of this range doesn't matter. Well, that's... just saying...
Well, how can you determine the effectiveness of a chemical compound against a microorganism without experimental testing?
Like, a rough statistical calculation of the probability of the destructive effect of chemical compound X against bacterial genome Y suddenly hits a 10 in a few hours?
Why not find a compound against a coronovirus in this way? Why, then, do we need months to study its testing?
 
Rettag Konow:
Sounds like a fake. How did they collect and describe 107 million chemical compounds for AI? That doesn't make any sense... How could the AI select the effective choices without empirical sampling?

We had to tailor a special chemical compound for each resistant organism. There were dozens of organisms, 107 million compounds. And a selection with no hands-on experience. That's a fake...

The article on zen is a translation of a popular article from the guardian. There you can figure out where to look for a scientific publication about this study (it is a peer-reviewed scientific journal in biology). You read it and, with enough education in the field, decide if it is true.

Don't be like Hegel who "proved" that the number of planets in the solar system must be seven)

 
Aleksey Nikolayev:

The article on zen is a translation of a popular article from the Guardian. There you can see where to look for a scientific publication about this study (it is a peer-reviewed scientific journal in biology). You read it and, with enough education in the field, decide if it is true.

Do not be like Hegel "proved" that the number of planets in the solar system should be equal to seven)

Do not try to "stun" Hegel. Everyone makes mistakes. Forgivable, for a man of the 18th century. Well, and you, Mr. Nikolaev who? What do you think you are? Your "works" are already studied by the Academy of Sciences?))
 
ReTeg Konow:
Okay, so how can you tell if a chemical compound is effective against a microorganism without experimental testing?
Like, a rough statistical calculation of the probability of the destructive effect of chemical compound X against bacterial genome Y suddenly hits a 10 in a few hours?
Why not find a compound against a coronovirus in this way? Why, then, do we need months to study its testing?

The point here is kind of the following:

AI tell us which chemical compound we need.

Guys try this one, it's more likely.

Then practical testing is already underway. By doing so, the AI helps reduce the search time. This is called simulation, when they first build a model of a ship and test it in a bathtub, and then build it in real size after knowing its behavior parameters. As an example...

 
Many German and Soviet ideologues and philosophers (including Marx, Engels and Lenin) referred in their works to Hegel and his dialectical concept. A constellation of scientists of the Soviet Academy of Sciences published a book of commentaries on Hegel's "Science of Logic," and Nikolaev simply stated that it was all crap. And he will consider not to be like those who declare something without sufficient education in the field.) Although, in this, I agree.
 
Konow's tag:
Don't try to "stun" Hegel. Everyone makes mistakes. Forgivable, for a man of the 18th century. Well, and you, Mr. Nikolaev who? What do you represent yourself? Your "works" are already studied by the Academy of Sciences?))

ad hominem is used when there is nothing substantive to contradict)