Skip to main content
Share Your Experience: Take the 2024 Developer Survey

Questions tagged [generalization]

For questions related to the concept of generalization in computational learning theory and machine learning.

Filter by
Sorted by
Tagged with
16 votes
1 answer
376 views

Will parameter sweeping on one split of data followed by cross validation discover the right hyperparameters?

Let's call our dataset splits train/test/evaluate. We're in a situation where we require months of data. So we prefer to use the evaluation dataset as infrequently as possible to avoid polluting our ...
Philipp Cannons's user avatar
14 votes
3 answers
1k views

How does noise affect generalization?

Does increasing the noise in data help to improve the learning ability of a network? Does it make any difference or does it depend on the problem being solved? How is it affect the generalization ...
kenorb's user avatar
  • 10.5k
14 votes
1 answer
580 views

What are the state-of-the-art results on the generalization ability of deep learning methods?

I've read a few classic papers on different architectures of deep CNNs used to solve varied image-related problems. I'm aware there's some paradox in how deep networks generalize well despite ...
Shirish's user avatar
  • 393
10 votes
5 answers
1k views

Why can neural networks generalize at all?

Neural networks are incredibly good at learning functions. We know by the universal approximation theorem that, theoretically, they can take the form of almost any function - and in practice, they ...
Nico A's user avatar
  • 201
6 votes
1 answer
2k views

When exactly is a model considered over-parameterized?

When exactly is a model considered over-parameterized? There are some recent researches in Deep Learning about the role of over-parameterization toward generalization, so it would be nice if I can ...
Phúc Lê's user avatar
  • 181
5 votes
2 answers
3k views

Is pooling a kind of dropout?

If I got well the idea of dropout, it allows improving the sparsity of the information that comes from one layer to another by setting some weights to zero. On the other hand, pooling, let's say max-...
nsaura's user avatar
  • 258
5 votes
3 answers
10k views

How can the generalization error be estimated?

How would you estimate the generalization error? What are the methods of achieving this?
kenorb's user avatar
  • 10.5k
5 votes
1 answer
104 views

What are the techniques for detecting and preventing overfitting?

I'm worrying that my neural network has become too complex. I don't want to end up with half of the neural network doing nothing but just take up space and resources. So, what are the techniques for ...
kenorb's user avatar
  • 10.5k
5 votes
1 answer
531 views

How can my Q-learning agent trained to solve a specific maze generalize to other mazes?

I implemented Q-learning to solve a specific maze. However, it doesn't solve other mazes. How could my Q-learning agent be able to generalize to other mazes?
lrosique's user avatar
4 votes
1 answer
830 views

Does DQN generalise to unseen states in the case of discrete state-spaces?

In my understanding, DQN is useful because it utilises a neural network as a q-value function approximator, which, after the training, can generalise to unseen states. I understand how that would work ...
Redox's user avatar
  • 43
4 votes
1 answer
110 views

Is there a way to ensure that my model is able to recognize an unseen example?

My question is more theoretical than practical. Let's say that I am training my cat classifier with a dataset that I feel is pretty representative of cat images in general. But then a new breed of cat ...
mdurrant's user avatar
3 votes
2 answers
146 views

When exactly am I overfitting -- contradicting metrics

I am training an object detection machine learning pipeline. Among the many metrics provided out of the box by tensorflow object detection API, I look at total_loss and DetectionBoxes_Precision/mAP@....
user1091141's user avatar
3 votes
1 answer
271 views

Is there a notion of generalization in unsupervised learning?

I've been learning a little bit about generalization theory, and in particular, the PAC (and PAC-Bayes) approach to thinking about this problem. So, I started to wonder if there is an analogous ...
Marcel's user avatar
  • 133
2 votes
1 answer
894 views

How can we get a differentiable neural network to count things?

Imagine I have images with apples in them. I want to train a neural network which can count the number of apples in each image. BUT, I don't want to use a detector, then count the number of bounding ...
Alexander Soare's user avatar
2 votes
3 answers
3k views

Is the inductive bias always a useful bias for generalisation?

Is it true that a bias is said to be inductive iff it is useful in generalising the data? Or does inductive bias can also refer to the assumptions that may cause a decrease in performance? Suppose I ...
hanugm's user avatar
  • 3,890
2 votes
1 answer
136 views

Is there any research on models that make predictions by also taking into account the previous predictions?

With the recent revelation of severe limitations in some AI domains, such as self-driving cars, I notice that neural networks behave with the same sort of errors as in simpler models, i.e. they may be ...
user4779's user avatar
  • 203
2 votes
1 answer
145 views

How does replacing states with latent representations help RL agents?

I have seen many papers using autoencoders to replace images (states) with latent representations. Some of those methods have shown higher rewards using such techniques. However, I do not understand ...
desert_ranger's user avatar
2 votes
2 answers
87 views

How general is generalization?

I am sorry but I have to explain my question using an example, I do not know how to ask it in proper scientific terms. Let's assume, I have trained a deep learning model on classifying hand gestures, ...
GKozinski's user avatar
  • 1,270
2 votes
1 answer
670 views

Does a bigger neural network learn "worse" representations than a small neural network when the amount of data isn't enough?

Assume we have a neural network and we want to train it on a classification problem. The hidden layers of the neural network are kind of feature representations of the input data. If the neural ...
realmarv's user avatar
2 votes
2 answers
994 views

Why don't neural networks project the data into higher dimensions first, then reduce the size of each layer thereafter?

Background From my understanding (and following along with this blog post), (deep) neural networks apply transformations to the data such that the data's representation to the next layer (or ...
Kevin's user avatar
  • 133
2 votes
0 answers
233 views

Why does learning rate reduce train-test generalization gap?

In this blog post: http://www.argmin.net/2016/04/18/bottoming-out/ Prof Recht shows two plots: He says one of the reasons the plot below has a lower train-test gap is because that model was trained ...
user3180's user avatar
  • 628
2 votes
0 answers
80 views

How to Prove This Inequality, Related to Generalization Error (Not Using Rademacher Complexity)?

This is an inequality on page 36 of the Foundations of Machine Learning by Mohri, but the author only states it without proof. $$ \mathbb{P}\left[\left|R(h)-\widehat{R}_{S}(h)\right|>\epsilon\right]...
j200932's user avatar
  • 181
2 votes
0 answers
47 views

References on generalization theory and mathematical abstraction of ML concepts

I'd like to learn about generalization theory for machine learning algorithms. I'm looking for books and other references (in case books aren't available) that provide a gentle introduction to the ...
Shirish's user avatar
  • 393
1 vote
1 answer
281 views

Why is the validation loss less than the training loss, and what can be said about the effect of the learning rate?

I have the following results I am trying to make sense of. I have attached the loss curves here for reference. As you can see, the first issue is that the validation loss is lower than the training ...
chinmay's user avatar
  • 113
1 vote
2 answers
251 views

What does "the expectation is taken across different possible inputs, drawn from the distribution of inputs we expect the system to encounter" mean?

I am currently studying Deep Learning by Goodfellow, Bengio, and Courville. In chapter 5.2 Capacity, Overfitting and Underfitting, the authors say the following: Typically, when training a machine ...
The Pointer's user avatar
1 vote
1 answer
242 views

Out-of-domain generalization

Given $X$ the space of all $N \times N$-pixel images and $I=\{$airplane,clock,axe,...$\}$ a set of labels. An image classification task is generally concerned to learn a map $$F:X \rightarrow I$$ Let'...
NicAG's user avatar
  • 113
1 vote
1 answer
98 views

Why does the discrepancy measure involve a supremum over the hypothesis space?

I am referring specifically to the disc defined by Kuznetsov and Mohri in https://arxiv.org/pdf/1803.05814.pdf This is a kind of worst case path dependent generalization error. But what is the ...
safetyduck's user avatar
1 vote
0 answers
16 views

Enhancing Generalization in DRL Agents in Static Data Environments

Context: I'm working with a deep reinforcement learning (DRL) agent in a market-like environment where its actions do not affect the environment. The environment uses historical data up to a certain ...
ElonMuskofBadIdeas's user avatar
1 vote
2 answers
115 views

Master theorem about polynomial classifiers?

Does anyone know if there is a theorem or counterexample establishing whether or not for any given binary classification task in some finite (possibly large) dimensional vector space of attributes, ...
letsmakemuffinstogether's user avatar
1 vote
1 answer
389 views

How does Weight Sharing with the Generalization in Graph Neural Networks work?

I have two closely related points regarding the weight sharing and generalization of graph Neural network. For illustration purposes, I attached two images which I reference. Images are taken from the ...
Imago's user avatar
  • 111
1 vote
2 answers
79 views

Late Onset Augmentation

If I train a U-Net model for image segmentation (e.g. medical images) and start training until it converges and then add augmentation - can i expect similar results as if i train with augmentation ...
Samuel Peterson's user avatar
1 vote
0 answers
101 views

How do I determine the generalisation ability of a neural network?

I am trying to ascertain if my neural network is able to generalize or if it’s simply using memory/overfitting to solve a task. I would like my model to generalise. Currently, I train the neural ...
skywalkerdk's user avatar
0 votes
2 answers
110 views

Why does model overfitting lead to poor generalization?

If a model overfit to the training data, why does it generalize poorly? Consider the basic problem of a noisy 2d dataset where I am fitting polynomials. A good model would be a parabola and a line ...
JobHunter69's user avatar
0 votes
1 answer
60 views

How does not learning far inputs make the RNN forget far inputs?

I am totally aware of the problem of the vanishing gradient. It usually occurs with vanilla RNN, where with a long sequence of data, the gradient will vanish or explode for far input sequence, and ...
John adams's user avatar
0 votes
1 answer
234 views

Is it possible that the model is overfitting when the training and validation accuracy increase?

I am aware of similar questions that have been asked, and I have gone through many. I want to bring my case to SE to understand better what my results are. I am working with a large dataset (around ...
nachofest's user avatar
0 votes
2 answers
3k views

How is the DQN able to generalise the learning to unseen states with such a loss function?

I am trying to understand how deep Q learning (DQN) works. To my current understanding, each $Q(s, a)$ functions is estimated to be a function of a feature vector of its state $\phi$(s) and the weight ...
calveeen's user avatar
  • 1,271
0 votes
0 answers
46 views

Can you illustrate how the weights in transformer model generated from a training sentence can be generalized to an unseen test sentence?

Can you show how the weights in transformer model are generalizable?
Steven's user avatar
  • 99