Deep Learning is reaching its best year of progress. It also becomes the buzzword for the recent year by any stakeholders. Because of that, there’s a lot of resources that exist, such as courses, papers, communities, and many more.
And amazingly, most of the resources are given at no cost, and it’s publicly accessible on the internet. Therefore, Deep Learning becomes available for communities around the world regardless of what’s their background and where they represent.
As those factors make Deep Learning so famous this time, some effects will affect to us. One of them is affecting our mental health. With that vast amount of resources, it leads us to become overwhelmed.
Feel overwhelmed is a feeling where you become frustrated because of some factors. If we don’t treat it really great, it will lead us to become unmotivated to do many things.
Yoshua Bengio, one of the winners of 2018 Turing Award and the famous deep learning researcher, recently publishes an article on his blog about rethinking how the publication in machine learning would like to be. It resembles to a manifesto called Slow Science. I’ve quoted it, and what is said is like this,
We do need time to think. We do need time to digest. We do need time to misunderstand each other, especially when fostering lost dialogue between humanities and natural sciences. We cannot continuously tell you what our science means; what it will be good for; because we simply don’t know yet. Science needs time.
— Bear with us, while we think.
Based on the quotes above, it shows that, as a scientist, we should slow down ourself to achieve things. What he really wants is that the field should have more conversations to it. Therefore, it could produce a breakthrough, or it could be new inspirations.
The concept itself is not just for the scientist, but also for us, as a learner, to slow us down to achieve more from the field.
Now, let me ask you these questions,
- How does your feeling right now with the vast amount of Deep Learning resources?
- How much you’ve recently read the new Deep Learning research papers?
- What kind of things you’ve already learned about Deep Learning?
Probably you will answer that you’ve learned a lot, and also you feel okay while learning Deep Learning. Are you sure with that? Maybe you will answer it yes or no. But if you already open so many resources and don’t feel that you’ve learned any single things from it, you should take a break.
We may want to learn so many things, and we want to achieve many things, but don’t rush on it. We have to slow ourself down, or it could lead us to become unmotivated.
Tips To Avoid Overwhelmed
Now, the problem is how we can learn Deep Learning without getting overwhelmed?
Here are my tips on how to learn without getting overwhelmed.
Find your learning style
People can learn all of the things. But to learn it, probably you will use a different style for learning. There are many kinds of learning styles. They are,
- Visual learner
- This type learns by using any graphics or visualizations to understand things. The example of this type is the people that prefer taking notes using a mind map.
- Auditory learner
- This type learns by listening to lectures and having discussions to understand concepts.
- Read and write learner
- This type basically learns by reading books and taking notes.
- Kinesthetic learner
- This type basically learns by doing. For example, by implementing code to some neural network to gain an understanding of how neural network works.
So, based on those types, which one suits you? You don’t have to choose one of them. You could combine those to make your learning more effectively. I’m the type of kinesthetic and reading learner because I would prefer to learn from reading books and then implement those.
You know yourself. Therefore, you have to choose which one suits you.
Stick with one course only until it’s finished
Right after you’ve known your learning style, the next step is to determine which resource to use. There are many resources to learn deep learning, such as Stanford University, MIT, Deep Mind, and many more. You could use any kind of those, but one rule that you must hold is that,
USE. ONE. RESOURCE. ONLY!
Each course has its own uniqueness, but mostly it gives you the same concept. Therefore, you should stick on it, and you have to learn until you’ve finished the course.
If you feel that you don’t understand the course, you can search other resource as your additional reference, and don’t use it as your primary resource. Just use with the first one for your main resource.
Dedicate 1–2 days to learn or do something else
Learning sometimes makes you bored. To make sure that you always get motivated, you should do something else that could matter to you or something that you love. Or probably you are an employee for a company, so you can not choose when to learn new things.
You can apply OpenAI’s Learning Day to solve this. Learning Day is a day for learning new things that are not your expertise. For example, you are a data analyst, and you are interested in reinforcement learning. Therefore, you can allocate one full day to learn that. Probably you can schedule it on weekdays or weekend.
By doing it, you will not get bored, and you can increase your skill in another field.
#deep-learning #mental-health #education #deep learning
