Questions tagged [machine-learning]
Questions about computer algorithms that automatically discover patterns in data and make good decisions based on them.
1,080
questions
5
votes
1answer
142 views
What algorithm do SVMs use to minimize their objective function?
Support Vector Machines turn machine learning linear classification tasks into a linear optimization problems.
$$ \text{minimize } J(\theta,\theta_0) = \frac1n \sum_1^n \text{HingeLoss}(\theta,\...
0
votes
0answers
21 views
why divide norm of w in svm
So I get this through the math behind the SVM (Support Vector Machine), and I get this formula
$$(w^T)(x_1-x_2) = 2.$$
We then divide both side with norm of $w$ then we get the new formula
$$ \frac{w^...
0
votes
0answers
5 views
Which model to apply on such panel data with so may rows but for each unique id rows are 6-8 rows per unique id?
I am new to such panel data where I have multiple observation for same ID in different Quarter and I am not sure what kind of machine learning algorithm I can apply.
I have data from Q1-18 till Q4-...
0
votes
0answers
43 views
deep learning - is loss value should be divided by `seq_length`
I'm new to deep learning and I'm looking at train.py here: https://github.com/huanghao-code/VisRNN_ICLR_2016_Text/blob/master/train.py.
I run this code and the loss ...
0
votes
0answers
6 views
How is the dimensionality of the volume in ConvNets determined for a general case?
In ConvNets, I understand how the dimensionality of a flat image changes after convolving it with a single filter. For example, if you convolve a P x P x 1 image (assume no padding) with a filter with ...
0
votes
0answers
23 views
Looking for references for real-world scenarios of data-poisoning attack on labels while doing supervised learning
Consider the following mathematical model of training a neural net : Suppose $f_{w} : \mathbb{R}^n \rightarrow \mathbb{R}$ is a neural net whose weights are $w$. Suppose during the training the ...
0
votes
1answer
21 views
Is any of the assumptions hold true for the mentioned?
Suppose the use of linear regression. The result of the MSE is 120.5(mean squared error) for the train-set. Wev'e reached the minimum of training the data.
Is it possible that by applying Lasso(L1 ...
1
vote
0answers
11 views
What is spatial information in neural network? [duplicate]
The architecture of the CNN model contains several convolutional layers, non-linear activations, batch normalization, and pooling layers. The initial layers learn the low-level concepts such as edges ...
0
votes
1answer
28 views
Is it possible to do face recognition with just the eyes?
Assuming the input photo is focused on a person's face, if the person is wearing a surgical mask, most face recognition software fail to identify the subject's face.
Most facial landmark models are ...
0
votes
0answers
16 views
Kohonen, 1 dimension, SOM, puzzle
We consider training one-dimensional open map of Kohonen with neurons in one-dimensional input space. We assume it is completed the phase of the device and the weights wi, i = 1,2, ..., N, are placed ...
1
vote
0answers
22 views
Even distribution of weighted sums across a period of time
I want to automatically&optimally split and distribute a number in several buckets as evenly as possible across a period of time.
Example:
5000 across 12 months, in 6 buckets, with weights 15%, ...
1
vote
1answer
51 views
How does Gradient Descent treat multiple features?
As far as I know, when you reach the step, in a gradient descent algorithm, to calculate step_size, you calculate ...
1
vote
0answers
32 views
Reinforcement learning and Graph Neural Networks: Issue with entropy [closed]
I am currently working on an experiment to link reinforcement learning with graph neural networks.
This is my architecture:
Feature Extraction with GCN:
there is a fully meshed topology with ...
0
votes
1answer
13 views
What exactly is label noise?
I've been doing research on precollege summer programs, and one ongoing project that has come up is "Improving Label Noise Robustness with data Augmentation and Semi Supervised Learning". So,...
1
vote
1answer
20 views
Is transfer learning applied on only similar datasets only?
I am trying to make a CNN model on different brands of logos . Firstly , I wrote a CNN from scratch and trained it on which I got 70% accuracy, I have total 40 classes and each class has 100 images . ...
1
vote
0answers
46 views
Best algorithm/model to establish relevance between events utilizing mixed data type (Tags, Time, x_coordinate, y_coordinate)? [closed]
I'm building a relevance ranking system for incidents occurrence and prevention. My goal is to use four attributes to establish relevance: tag (About 500 tags), x_coordinate, y_coordinate and time. ...
0
votes
0answers
15 views
Are Learning Classifier Systems considered (learning) agents?
Does a LCS model a learning, rule-based agent or does a LCS just implement the agent program (mapping of inputs to actions)?
According to Russel & Norvig :
agent = architecture + agent program
...
1
vote
1answer
46 views
What is the Number of epochs with no improvement after which training will be stopped.?
I am trying to make a Convolutional neural network. Training the images of different brands of Logos. Have 100 images per class and there are 40 classes. I have trained the model now want to check ...
0
votes
1answer
33 views
How does CNN deal with rotation invariant pictures?
I am trying to make a CNN model . Training the image . Want to know that When we apply kernel on image and take out the features of images. That features are rotation invariant or we have to apply ...
0
votes
0answers
12 views
How do I choose the right model for production?
I am trying to build a classification algorithm having 28 classes. These classes consists of Logo of companies like adidas , Nike etc. I have very low dataset below than 100 images and greater than 70 ...
1
vote
1answer
28 views
Which pretrained model will be best for my dataset?
I am trying to build a classification algorithm having 28 classes. These classes consists of Logo of companies like adidas , Nike etc. I have very low dataset below than 100 images and greater than 70 ...
0
votes
0answers
30 views
Human brain controlled computer
It is already possible to let a computer drive the human brain to some limit like for example in the medical world, but will it be possible to let the human brain control a computer system by directly ...
0
votes
0answers
19 views
Recommender system - Binary rating adding a condition
I want to implement a recommender system based on collaborative filtering. Each user will have a set of visualized items (0-1's) and based on that similitude the recommendation will be generated.
In ...
0
votes
1answer
64 views
If i use other peoples code for my final year project does that come under Plagiarism?
I want to make a facial recognition project but my coding ability is very limited so I am going to be relying on other peoples code but if i use their code even if i change the variables and make ...
1
vote
1answer
17 views
Is there an online/offline tool that can perform K-means/median, given an initial centroid from the user?
The types of problems I am trying to solve are as follows:
Given a set of co-ordinates such as: (1,2), (3,3), (6,2), (7,1), a value of k such as k=3 and an initial set of centroids such as c1=(2,2) ...
2
votes
0answers
51 views
any hope to solve np hard problem using deep learning? [duplicate]
I know some basic machine learning and deep learning. Now a days deep learning solve many types of problem. I working working optimization problem like np, np hard problem. Is there any hope to solve ...
0
votes
0answers
23 views
Why do Google Deepmind images always look the same?
There is a certain sameness to google deep mind images. Why do they always look like a bad tool album cover?
0
votes
0answers
17 views
Has GPT-3 put web designers out of work yet?
https://prototypr.io/post/gpt-3-design-hype/
Has GPT-3 had any effect on the how many humans are hired to do web design yet?
0
votes
0answers
17 views
Designing a neural network with LSTM and feedforward NN combination
Currently, I'm designing a neural network that works with reinforcement learning. In summary, the agent takes in information about itself and nearby agents and, in conjunction with global world ...
0
votes
0answers
16 views
What does Online Transfer Learning Mean?
The last sentence of A review on transfer learning in EEG signal analysis states that:
Furthermore, online learning needs to be developed to ensure the
performance of transfer learning in practical ...
3
votes
0answers
58 views
Axis aligned rectangles: why is A an ERM in the case of infinite domain?
I'm working on a problem 2.3a in Shalev-Shwartz/Ben-David's Machine learning textbook, which states:
An axis aligned rectangle classifier in the plane is a classifier that assigns 1 to a point if and ...
0
votes
0answers
14 views
What are Volterra classifiers
I want to learn about Volterra classifiers in-depth to understand the following papers
Papers
https://openaccess.thecvf.com/content_cvpr_2014/papers/Kang_Convolutional_Neural_Networks_2014_CVPR_paper....
1
vote
0answers
10 views
Can you use a bunch of old laptops and a VNC to make a cluster computer? [closed]
I want to use a bunch of old laptops and connect them to each other through either usb or Ethernet and act as one multi cpu computer. Then I want to use a vnc to access that computer either through my ...
2
votes
1answer
68 views
Can you suggest a topic for a Bachelor Thesis in Mathematics that is related to Machine Learning? [closed]
Context
I am a final year Bachelor of Mathematics student and next semester I will write my Bachelor thesis.
My interests are in Machine Learning (ML) and I will do a master in ML next year. More ...
0
votes
0answers
9 views
Is there an accurate way to convert frequencies of the Stupid Backoff language model to probabilities?
We introduce a similar but simpler scheme, named Stupid Backoff, that does not generate normalized probabilities. The main difference is that we don't apply any discounting and instead directly use ...
0
votes
0answers
11 views
How to do training & inference with denoising autoencoders?
Following the development in the Goodfellow book:
https://www.deeplearningbook.org/contents/autoencoders.html (sec 14.5)
I have 2 questions:
To me the text seems to imply that we are doing SGD only ...
2
votes
1answer
14 views
How to interpret parametric formulation of information bottleneck?
I'm reading this paper on latent representations with the information bottleneck https://arxiv.org/pdf/1804.06216.pdf and in section three, the authors write that the parametric formulation of the ...
1
vote
0answers
12 views
Is relation extraction considered a subtask of information extraction?
I'm currently trying to investigate the relationship between relation extraction (RE) and event extraction (EE). Doing more reading on the two tasks has caused me to question my initial belief that RE ...
0
votes
1answer
13 views
Conditional probability in Expectation Maximization (EM)
I've got the following equation:
$p(j = 1 | x, \theta) = \frac{p(j=1,x | \theta)}{p(x | \theta)}$
Why does it hold? Or maybe, how do I use Bayes Theorem in this case, i.e. if we do not only have $p(j =...
0
votes
0answers
37 views
Why is Machine Learning taught in Computer Science Courses but not in Mathematics?
I always wonder why machine learning is native to computer science rather than its more close fields like statistics and mathematics? In machine learning we CS engineers learn stats and maths and ...
0
votes
0answers
27 views
Questions on AlphaZero Implementation
So I've been implementing AlphaZero for Chess from scratch and there were a few things the papers mentioned that I'm not sure how to implement. I'll reference both the original AlphaGoZero paper and ...
0
votes
0answers
27 views
Stuck on Ejaz and Islam's paper about Masked Face Recognition
I'm relatively new in machine learning and I am trying to put together my undergraduate thesis on masked face recognition.
I've read Ejaz and Islam's paper (available at https://www.researchgate.net/...
0
votes
0answers
36 views
PAC Learnability of axis aligned rectangles
I came across a question while going through ' Understanding Machine Learning' by Shai Ben-David.
For the class H of axis-aligned n-dimensional rectangles in R^n i.e H = {[a1, b1] × · · · × [an, bn] :...
0
votes
0answers
20 views
Applying Bayes rule in the context of reinforcement learning
I was watching this video on reinforcement learning. At 1:28, it says following:
$$Pr(s'|a,z,s)=\frac{Pr(z|s',a,s)Pr(s'|a,s)}{Pr(z|a,s)}$$
I was unable to get how this was obtained. I pondered a bit ...
0
votes
1answer
28 views
Trying to figure out the best way to approach this problem with Computer Vision
Newbie to CV here so sorry of this is basic. Here's the deal, I have a program that I run many times. and each run I produce a screenshot. I need to compare screenshots from N-1 and N runs and make ...
0
votes
0answers
6 views
Citation Graph prediction
If I have reference section of a paper, with how much probability can I predict one of the authors of that paper ?
Any recommendations on suitable algorithm to build a model are appreciated.
Also how ...
0
votes
0answers
7 views
Prophet Model in Batch Implementation
I am newby at Prophet model. I was wondering Prophet uses stochastic or batch approach while in training. If it uses stochastic that means there is nothing to do convert and compute mini batch ...
0
votes
1answer
28 views
SMO, Random forest and Bayes net algorithms: why does Random forest perform better?
I analyzed a dataset using those 3 different algorithms.
As I can see, Random forest performs better in most cases.
My dataset is composed of 4000 instances of two classes (class A 2000 elements, ...
0
votes
0answers
26 views
Papers on depth separations in neural networks
I am new to the domain of Machine learning. I have been asked to present a paper related to the mathematics behind the depth separations in Neural Networks (by Itay Safran, Ronen Eldan and Ohad Shamir)...
0
votes
2answers
35 views
Conv Net dimensions misunderstanding?
I was asked this question:
Given an image with shape [1,28,28], what will be the shape of the output of a convolution layer with 10 5x5 kernels (filters) without padding?
Now, are the shape dimensions ...
