Not DeepSeek... yet

[HSE]

Hi all!

Deep Learning Network from Thomas Bleek (see Playing with Neural Network) is amazing... but no so easy to understand 

In order to play with the basics, I made some additions to original G. Friedrich / Jaymeson Trudgen (NaN) Matrix.inc object.

To begin some tests I adapted a little Perceptron. It's terrible (later: that was because Resguard) slow, mainly because Matrix object by design is for learning, and error proof. Some next variation of Matrix object will omit tests and use memory previously allocated.


Also playing a little with namespaces in ObjAsm. That bother me for years now, but still is there. Then, could be usefull for some ANN tests in the future

Because nobody use that, is missing the namespace mechanism in SetObject macro, that can be updated easily:

Code Select Expand

SetObject macro Register:req, ObjName:=<>, pInstance
  if (opattr Register) and OAT_REGISTER                ;;Is pInstance a register?
    ifb <ObjName>
      $$OName textequ <??ObjectContext>
    else
      PreParseExpr ObjName
      if $ParseObjExpr(ObjName) eq FALSE
        @Err <$Obj Error: syntax error in ObjName>
      else
        $$OName textequ <??ObjExpr>
      endif
    endif
    *  *  *

Any idea is welcome 

Regards, HSE

note: removed Resguard

[Biterider]

Hi HSE
I tired the the exe but it requires ResGuard32.dll. Once added, it runs smoothly.

I also added the few lines of code to SetObject. 
I must admit, that I forgot to check that code path. Thanks 

Regards, Biterider

[six_L]

Hi,HSE
interesting topic.
from DeepSeek.

A neural network is a computational model inspired by the structure and function of the human brain. It is a fundamental concept in artificial intelligence (AI) and machine learning (ML), used to recognize patterns, make decisions, and solve complex problems.

What is a Neural Network?
A neural network consists of interconnected layers of artificial neurons (also called nodes or units). These layers include:
1. Input Layer: Receives data (e.g., images, text, or numerical values).
2. Hidden Layers: Process the data through weighted connections and activation functions.
3. Output Layer: Produces the final result (e.g., a classification, prediction, or decision).

Each neuron takes input, applies a mathematical operation, and passes the result to the next layer. The network "learns" by adjusting the weights of these connections during training.
----------------------

What Does a Neural Network Do?
Neural networks are used for tasks such as:
1. Pattern Recognition: Identifying patterns in data (e.g., facial recognition, speech recognition).
2. Classification: Categorizing data into classes (e.g., spam detection, medical diagnosis).
3. Regression: Predicting continuous values (e.g., stock prices, weather forecasts).
4. Generation: Creating new data (e.g., text generation, image synthesis).
5. Decision-Making: Solving complex problems (e.g., game playing, autonomous driving).
----------------------

How Does It Work?
1. Training: The network is trained on labeled data using algorithms like backpropagation to minimize errors.
2. Inference: Once trained, the network can make predictions or decisions on new, unseen data.
----------------------

Types of Neural Networks
1. Feedforward Neural Networks (FNN): Basic architecture where data flows in one direction.
2. Convolutional Neural Networks (CNN): Specialized for image and video processing.
3. Recurrent Neural Networks (RNN): Designed for sequential data like time series or text.
4. Generative Adversarial Networks (GAN): Used for generating realistic data (e.g., images, videos).
5. Transformer Networks: Excel in natural language processing (e.g., GPT, BERT).
----------------------

Applications
- Computer Vision: Object detection, facial recognition.
- Natural Language Processing (NLP): Language translation, chatbots.
- Healthcare: Disease diagnosis, drug discovery.
- Finance: Fraud detection, stock prediction.
- Autonomous Systems: Self-driving cars, robotics.

In summary, a neural network is a powerful tool for solving complex problems by mimicking the way the human brain processes information. It is a cornerstone of modern AI and ML.

[HSE]

Hi Biterider!

but it requires ResGuard32.dll.

I have forgot that totally. Resguard is what slow down execution 

Once added, it runs smoothly.

I have Resguard32.dll in C:\Windows\SysWOW64, then is accesible from everywhere   

I also added the few lines of code to SetObject.

Fantastic 

Thanks, HSE.

[HSE]

Hi SixL!

inspired by the structure and function of the human brain

Not really, just a happy name. Only it is a numerical algorithm.

is a powerful tool for solving complex problems by mimicking the way the human brain processes information

This is directly a lie. AI is lying about its own definition   

Must we be worried? Not now, because is not intentional. Just process what is accesible. If people lie, it lie.

Regards, HSE.

[six_L]

Hi,HSE

This is directly a lie. AI is lying about its own definition

what is a lie?
if he knew it, but he said don't know; if he did not know, he said know.
if he has it, but he said he has not it; if he has not it, but he said he has it.

so we could only assume that the "DeepSeek" level is not high enough.

Regards.
six_L