What to feed to the input of the neural network? Your ideas... - page 64
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
one of the only valid things I think AI neural network can do is to detect certain candlestick patterns that we didn't or can't already identify or else for cracking other algos (if it can). There is nothing good in AI for predicting the future, quantitative indicators and rule based algos will always do the best job for that..not to mention - AI has proven itself to be weak at performing correct math.
stop confusing a database with a knowledge base
I'M BEGGING YOU!!!!!
BLOOD FROM YOUR EYES!!!
Before talking about the applied kind (algorithm) of learning, and even the theory of learning, I wanted to draw attention to the full picture of the local sabj (learning). From and to.
Based on some answers,
Learning is agradual mastering of a way of interacting with the world, where learning becomes not so much a process of accumulating knowledge as a constant adaptation, understanding of patterns and the ability to apply them in practice in a new context.
OK, we have generalised the concept.
But if analysis is the breaking down of the whole into parts, how can we study learning without studying the object of learning - information. Intellectual labour simulators (AI, LLMs) also work with it and nothing else. Information is first of all in the virtual field.
If virtuality is a non-material entity, then before studying the phenomena of virtuality without reference to real biology/robotics and in principle to the real world, we need to study the minimal unit of virtual space when analysing it. What can it be?
In reality - particles, atoms, molecules. Indeed, to understand the fundamental units of virtuality, we need to define what can be equivalent to "particles", "atoms" and "molecules" in virtual space.
Now it will be obvious information, don't get aggravated:
1. obviously, thebit as the minimal unit:
At the lowest level of any information, be it text, image, or sound, is the bit, an elementary unit of information that takes one of two values: 0 or 1. It is the basic "digital" unit with which all computing systems operate. In the context of the virtual world, the bit can be considered to be what all more complex structures are made of: programmes, objects, interactions.
But if a bit is just an elementary switch, then we need to move on to understand how virtual reality is created from these bits.
2.blocks of information (informationobjects):
Information can be objects of different levels of complexity, e.g.:
Thus, at a higher level, the minimum unit may not be a bit, but a certaininformation object - a minimum portion of data that has meaning in a particular context.
3. informationconnection or relation:
Information by itself, be it a bit or a byte, has no value unless it is in some kind of relationship with other data. This leads us to the following thought: the minimal unit of virtuality may be arelation or connection between elements of information. In this case, virtual space is not just a set of data, but a network of relationships and interactions. As in reality atoms interact through forces, so in virtual reality information objects interact through relations.
4.unit of meaning - semantic element:
We can go further and assume that the minimum unit of virtual space is thesemantic element - the minimum unit of meaning that a system or a person can perceive. This can be a word, a symbol, an action or any other phenomenon that has meaning in a particular context.
If virtual space is information, then meaning and context become key. Without meaning, information loses its significance. Therefore, we can consider a minimal unit as apiece of information that carries meaning, whether it is an image, a sound, text, or even a command in programme code.
5.unit of interaction:A virtual space is not just a data field, it is an active interaction between components. Therefore, another minimal unit can be anact of interaction, whether it is data processing,signal transmission, or a change of state in the system. Interaction is what makes virtuality "alive" and dynamic.
6. elementarystructure- pattern orpattern:Another thought might be that the minimal unit of virtuality is not just a single piece of data, but apattern - a repeating structure that serves as the basis for more complex phenomena. For example, in programming, these might be patterns of code or algorithms that can be thought of as the "atoms" of the software world. In this sense, a pattern is the minimal structure from which the entire virtual world is built.
Conclusion:
The minimal unit of virtual space can be different depending on the level from which we look at virtuality:
Looking at virtual space as a field of information, each of these units is important, but the key is perhaps how they interact and create meaningful structures.
Now forex:
Given that there is only one object in our information environment, the price chart, then each price, each quote, is the minimal unit of information. Each "particle" of the chart, combining into "atoms" of the chart and further into "molecules", form a different "substance" - patterns. These patterns in their turn then form something more.
But if in reality the substance has the same property at any moment of time, then in the virtual field we have such a key phenomenon ascontext.
Just like the text, which is read by Chat, defining the context from just 2 words, sets of quotes in idea should also have a "context" . As an example: when today the MA crossing upwards is a buy. And tomorrow the MA crossing upwards is a sell.
So, how would the structure/architecture of the engine look like in order to
1. Study aggregates of bits: information objects, semantic elements, relations and interactions, patterns and templates.
2. Learn an unknown language, like deciphering an ancient script
And learning itself: what would thislearninglook like
There are all sorts of MLPs, RNNs, LSTMs and so on: they just feed numbers into them. But numbers for such systems work when it is a reflection of real reality (stationarity), so in a purely virtual object like a price chart they do not work "point-blank". Numbers are fed to the input and they are further twisted with these numbers. And nobody knows what these numbers mean. They just decided to make a window. And submit it.
Learning is something more fundamental, complex, consisting of several "engines"/departments, which should probably be connected to each other, transfer something to each other, save/change/delete/correct. Read the context.
1. What does memorisation and "going through the options" have to do with it? You don't need to make up something I didn't claim and therefore couldn't impose on anyone.))
You wrote above:
Above you wrote:
Tell me how you assess quality when retaining such information: 3*3=9, which you claim only becomes memorisation after assessing quality.You sound out the answers to all the choices in the multiplication table and count the number of correct answers. When you reach 100% correct answers - congratulations!, you have learnt the multiplication table.
Retention differs from memorisation in that it does not require the formation of multiple connections. A unit of information when saving takes up a unit of space on the carrier, while memorisation requires the formation of links between units of saved information, and this is a process that requires evaluation.
Everyone went to school and remembers what it was like. Not everyone could memorise the multiplication table from the first time, they were often mistaken.))
Roughly speaking, memorisation is essentially a compression of information and takes up less space on a medium than retention and is usually an iterative process.Sound out the answers to all the choices in the multiplication table and count the number of correct answers. When you reach 100% correct answers - congratulations!, you have learnt the multiplication table.
Evaluate only the entry 3*3=9. You write down only 1 line at a time. To memorise it, how should you evaluate it?
Sound out the answers to all the choices in the multiplication table and count the number of correct answers. When you reach 100% correct answers - congratulations!, you have learnt the multiplication table.
Retention differs from memorisation in that it does not require the formation of multiple connections. A unit of information when saving takes up a unit of space on the carrier, while memorisation requires the formation of links between units of saved information, and this is a process that requires evaluation.
Everyone went to school and remembers what it was like. Not everyone could memorise the multiplication table from the first time, they were often wrong)).
Perhaps a better example is Pascal's triangle. Everyone remembers the multiplication table at least up to 10 from their toddler years....
about Pascal's triangle - also at school, but they learn the rules of construction. And all the information - units on the axes and the rule of recurrence addition ("add the number on the left with the number on the top, see the coefficient").
Evaluate only the entry 3*3=9. You only write down 1 line at a time. To memorise it, how should you evaluate it?
In this case, the answer will either be correct or not. The maximum possible number of correct answers is 1. If you reach this number, you have memorised it.