How should a physicist learn Machine Learning?

In summary, the author recommends reading papers mentioned in the tensorflow pages once they start using a specific function or class, but does not have a solid background and wants to learn more. He recommends the Quick hand on introduction, Intro book, and More detailed Intro book.
  • #1
Phylosopher
139
26
TL;DR Summary
Doing my master thesis in physics using machine learning (geared toward physics not machine learning). I can not afford learning neural networks "the regular way" because lack of time. What should I start with?
Hello everyone,
I am currently a master student working on physics and neural networks. I have already started producing neural network results (I use tensorflow and keras) so I know how to program the basic things that I am required to do, the problem is that I do not understand them well.

I do not have a good solid background when it comes to the neural networks themselves, and as a master student working on his thesis, I can not/I do not want to write my thesis blindly. I want to truly understand the behavior of my program.

Problem is, I can not afford reading a 600 page book on ML! My idea was to read the papers that are mentioned in the tensorflow pages once I use a specific function or class(Example: Here). But as I said, I do not have a solid background to read them properly.
So, my question is, what resources you think I should read first before I delve deeper in these papers? I need the bare minimum so I can accelerate my learning.

Things I have bookmarked so far that I think are useful:: Quick hand on introduction, Intro book, More detailed Intro book ... Do you think these are good starting points? Do you have better suggestions?
 
Technology news on Phys.org
  • #2
Quick hand on introduction - I can thoroughly recommend this, it should help you develop a deep intuition for the process and could be all you need.

Intro book - I'm not familiar with this but looking at the ToC it might be a good fit.

More detailed Intro book - this seems to have a much wider scope than you need but might be good context.
 
  • Like
Likes FactChecker, jedishrfu and atyy
  • #3
This short, four video set is the best explanation of neural networks that I've seen in my three years of study - 3Blue1Brown Neural Networks.
 
  • Like
Likes jedishrfu, Greg Bernhardt, nsaspook and 3 others
  • #4
The minimum ideas of machine learning are that deep learning is essentially curve fitting or classification. A neural network simply parameterizes a very complicated function. A traditional algorithm for curve fitting is linear regression. A traditional algorithm for classification is logistic regression.

Compared to basic statistical methods, an odd thing about neural networks is that there are many more parameters than data points. There are tricks that prevent overfitting. However, why overfitting can be avoided is not yet well understood.

Also, methods of estimating errors in deep learning are not well understood. However, there are reasonable attempts.
 
  • Like
Likes Borg and Greg Bernhardt
  • #5
atyy said:
The minimum ideas of machine learning are that deep learning is essentially curve fitting or classification
That is the essential lesson. It is possible that more than one method could lead to the same or a very similar curve fit. It's possible even if those other methods are not similar to deep learning.

So if you are just using the fit, or just using a tool to create the fit, it is not essential to learn the methods used.
 
  • #6
atyy said:
The minimum ideas of machine learning are that deep learning is essentially curve fitting or classification. A neural network simply parameterizes a very complicated function. A traditional algorithm for curve fitting is linear regression. A traditional algorithm for classification is logistic regression.
Although I wouldn't want to make deep learning sound too ordinary. Even for someone very used to curve fitting, there are many aspects and capabilities of deep learning that are surprising, powerful, and fascinating.
 
  • Like
Likes pbuk

Related to How should a physicist learn Machine Learning?

1. What is the best way for a physicist to get started with learning Machine Learning?

The best way for a physicist to get started with learning Machine Learning is to first gain a basic understanding of programming languages such as Python and R, as these are commonly used in Machine Learning applications. They should also familiarize themselves with the fundamental concepts of Machine Learning, such as data preprocessing, model building, and evaluation. Online courses, books, and tutorials are great resources for learning these skills.

2. What are the key skills that a physicist should focus on when learning Machine Learning?

A physicist should focus on developing skills in programming, statistics, and mathematics. These skills are essential for understanding and implementing Machine Learning algorithms, as well as interpreting and analyzing the results. Additionally, critical thinking and problem-solving skills are also important for a physicist to excel in Machine Learning.

3. Do physicists need to have a strong background in computer science to learn Machine Learning?

While having a background in computer science can be beneficial, it is not a requirement for learning Machine Learning. Physicists can learn the necessary programming skills and mathematical concepts through online courses and resources. However, having a basic understanding of computer science principles can make the learning process smoother.

4. How can a physicist apply Machine Learning to their research or work?

Physicists can apply Machine Learning to their research or work in various ways, such as data analysis, predictive modeling, and pattern recognition. For example, they can use Machine Learning algorithms to analyze large datasets and identify patterns or trends that may be difficult to detect using traditional methods. Machine Learning can also be used to develop predictive models for complex systems or processes.

5. What are some resources that can help a physicist learn Machine Learning?

There are many resources available for physicists to learn Machine Learning, such as online courses, books, tutorials, and workshops. Some popular online platforms for learning Machine Learning include Coursera, Udemy, and edX. Additionally, attending conferences and workshops related to Machine Learning can also be beneficial for networking and gaining practical experience in the field.

Similar threads

  • Programming and Computer Science
Replies
4
Views
1K
  • Programming and Computer Science
Replies
13
Views
2K
  • Programming and Computer Science
Replies
9
Views
1K
  • Programming and Computer Science
Replies
6
Views
467
  • Programming and Computer Science
Replies
6
Views
925
  • STEM Academic Advising
Replies
5
Views
996
  • New Member Introductions
Replies
2
Views
219
  • Programming and Computer Science
Replies
13
Views
871
  • Programming and Computer Science
Replies
4
Views
2K
  • Programming and Computer Science
Replies
8
Views
2K
Back
Top