All Questions
Tagged with machine-learningneural-networks
645 questions
0votes
0answers
15views
Why do my DNN convergence graphs behave differently on linear vs. dB scales?
I'm working on a deep neural network (DNN) and using the Adam optimizer to train it by learning parameters through backpropagation. My goal is to minimize the objective function. I’ve plotted the ...
- 1
2votes
1answer
661views
How can the exact same model give different confusion matrices for the test dataset and the entire dataset?
I have recently implemented a simple artificial neural network with 1 hidden layer. I split my data using train_test_split and I end up with the following confusion matrix in my test set. ...
2votes
1answer
87views
Learning curve behaviors across double descent regimes
I am learning about double descent phenomenon from here: https://www.di.ens.fr/~fbach/learning_theory_class/lecture9.pdf I was asking myself: When training a system, how can we know in which regime ...
- 255
2votes
1answer
59views
Are we really misunderstanding VC theory as arXiv:2205.15549 suggests?
arXiv:2205.15549 claims that the machine learning community misunderstood VC (Vapnik–Chervonenkis) theory and VC-theoretical understandings are sufficient to understand the double descent phenomenon. ...
1vote
1answer
69views
How can I create a custom RL environment for a city layout?
I'm pretty much new to RL and I'm making a project in RL for urban planning. I'm building an agent that either builds a house or a road. I'm currently designing the aspects of it. But I don't know if ...
0votes
0answers
19views
standard deviation of input in neural network
I have trained my model, and when I do prediction given the new input data, the standard deviation of all new input dataset increases a bit compared to the training one. we are expecting the results ...
0votes
1answer
79views
Are there any theoretical machine learning papers that have significantly helped practitioners?
21M deciding whether or not to specialize in theoretical ML for their math PhD. Specifically, I am interested in i) trying to understand curious phenomena in neural networks and transformers, such as ...
0votes
0answers
24views
Theoretical Machine Learning model for Mathematical Olympiad
Given a mathematical Olympiad problem there is only a relatively small list of "tricks" / tools / techniques used to solve it. These include things Fermat's little theorem, circle inversion, ...
- 111
0votes
1answer
39views
Understanding Generalization Error in Empirical vs True Risk Illustration
I am trying to understand the concept of generalization error based on the attached illustration that contrasts empirical risk (𝑅_hat) with true risk (𝑅) Two regions are marked in the diagram: Red-...
- 11
4votes
1answer
238views
Is there something like Cook's Theorem for Machine Learning?
Cook's Theorem* says that any problem in NP can be reduced to a SAT problem. *Cook, Stephen A. “The Complexity of Theorem-Proving Procedures.” In Proceedings of the Third Annual ACM Symposium on ...
- 515
1vote
1answer
64views
Are multiplicative weights essential in neural networks
In a classical fully-connected neural network, the linear transformation to go from one neuron of a layer to a neuron in the next layer can be written like $x^{(2)}_1 = w_1^{(1)} \times x^{(1)}_1 + ...
2votes
1answer
123views
What is the disjoint gene in NEAT?
I, like some sources, believed that disjoint genes are those that are "locked" between identical connections (https://www.researchgate.net/figure/NEAT-crossover-operation-example-Although-...
0votes
1answer
28views
Does SharpNEAT support RNN connections? [closed]
Do SharpNEAT or its versions as UnityNEAT support RNN connections? When I tried to use it, it clearly doesn't give this option.
1vote
1answer
51views
Confusion of point 3.2 in NEAT(Tracking Genes through Historical Markings) documentation
Rereading point 3.2 of “Efficient Evolution of Neural Networks through Complexification”(NEAT) by Kenneth O. Stanley, there remains a misunderstanding: how are we so sure that the 8th innovation was ...
7votes
2answers
2kviews
Has NEAT changed in 20 years?
This algorithm is more than 20 years old, so has it changed? During this time, the prevailing ones in text generation and classification (LSTM -> Transformer) have changed, AlexNET, ResNET have ...
