"New Neural" is an Open Source neural network engine project for the MetaTrader 5 platform. - page 50

[TheXpert]

Urain:

A layer is a union of neurons not dependent on the same iteration and having the same type.

How's that? How are you going to squeeze the juice out of a GPU without a vector representation? That's one of the things I'm considering.

[Mykola Demko]

TheXpert:
How do you do that? How are you going to squeeze the juice out of the GPU without vector representation? That's one of the things I'm taking into account.

Why without vector? The data arrays are external, it only remains to prescribe the correspondence of what data do with what data.

You have data about what neurons are in one layer, and pass to GPU calculation of vector of this layer. and so on through the layers.

SZY ah yes in neuron model that I drew, you need to introduce the concept of data to store intermediate calculations (well, to work with GPU they should be external too).

[Mykola Demko]

joo:

1) Why?

2) Why?

1) because in my formulation the layer can contain different types of neurons, and this is atypical tasks that cannot be transferred to the GPU

2) because in Andrew's formulation one neuron can be a layer, and that threatens to underload the GPU

ZY generally choose the lesser of two evils, underloading the GPU is not as terrible as the potential inability to use the GPU.

[Mykola Demko]

It's too bad that mql doesn't have pointers to data, otherwise we could just aggregate data from vectors into neurons directly.

HH wait, why do we need aggregation, if we can just pass the index of data vector instead. it will be the same binding with direct access.

HINT create a data reference object in the neuron instead of a data object.

[TheXpert]

Urain:

2) because in Andrei's formulation one neuron can be a layer, and this threatens to underload the GPU

If it can be merged... Then it should be merged.

[Mykola Demko]

TheXpert:
If it can be merged... Then it should be merged.

That's the problem if it can, and it can depend on two conditions, independence at one iteration and sameness with other layer neurons (i.e. how it processes data internally).

[Mykola Demko]

I have scraps of thoughts on how to organize a growing neural gas with such an engine, but they have not yet formed into words.

Only one thesis: network reinitialization will be necessary. The initialization itself should be available to the learning algorithm.

[Vladimir]

TheXpert:

The project manager can be gpwr. Partially can me.

Thank you for your confidence, but it seems to me that I would not fit as a leader of this project. I will explain why.

1. Programming is not my specialty. I do it as a hobby. Many of the nuances of coordinating a group of programmers is unknown to me.
2. For the last 15 years I managed several projects at work. I've found that the project under my supervision is successfully implemented in an atmosphere where participants are obligated to implement my commands (dictatorship) for fear of losing their jobs and paychecks. In a democracy such as the one here, where project members are not paid and there is nothing to keep them on the project team, the dictatorship method is not appropriate, and progress is very difficult without them.
3. The leader himself must be interested in the project. My interest so far is as an observer. As I mentioned before, I'm not interested in classic neural networks any more, especially the issue of their coding in MQL5. My interest now lies in the methods of transformation of input information. When I finish my self-training in these methods and experimenting with them on market quotes, my interest will move to something else.
4. The management of this project (as well as any other project) requires time and dedication, which I do not have at the moment (I already have a project at work).

From all of the above, my recommendation is this.

1. MCs should be well paid project participants so that there is a financial interest in participating in this project and its successful completion. Since I live in the USA, my pay standards are higher (which is probably why the MCs have never asked me to write an article, even when I offered myself as a candidate). For example, I believe this project will be successful if each participant is offered at least $10000 at the end of the project.
2. The leader of the project should be a representative of the MC. He should also set the requirements, deadlines, and their payment. Many open-source codes are created in universities, by students whose supervisors use diploma as motivation. You can't do without it.

[TheXpert]

gpwr:

PM doesn't have to be a programmer. It's a shame that you refuse.

[Reaper]

Good morning.

I saw that you were looking for a name for the project.

Even NeuRobot/NeuroBot/NeuRobotic -- Neurobot/Neurobotic might be an idea, in the case of neural robots.

A little childish as a name, but you get the idea of neural and financial robotics.