Machine Learning: 5 Biggest Mistakes – I’d like to share with you 5 of the biggest mistakes I made during my Machine Learning career. My life would be so much easier if only I had known how to avoid them. So I thought, maybe I should record a movie about that, so you guys don’t have to repeat those mistakes.
Let’s get started. First mistake…”Doing everything on my own”. That’s a big one, and I can’t imagine how much further I would’ve gotten if I only I had asked for help. People always think they know best and it’s very likely you and I are not exceptions to that rule. On the one hand, it’s very important to be confident with what you do – but on the other hand, it’s crucial to not be afraid to ask for help or advice.
I would even say that asking for help is rather proof of good and stable self-esteem because someone who doesn’t believe in himself will not find the courage to admit that he needs help. Just look at how many conversations I have on LinkedIn every day. There’s always something to learn from my friends.
It’s just about how you look at your conversations and what questions you ask – not about what someone has to offer. Don’t be afraid of making use of your friends’ knowledge. Just grow with them and don’t forget to give them a hand when they need it. The second mistake: “Learning things that most likely won’t be used often in the future”. I think it’s a very common mistake made by both beginners and experts in the Data Science field.
Think twice before diving deeply into a very specialized topic. You might be lost in a sea of information for days or even weeks and most likely you won’t have the occasion to use a newly acquired skill. Let me give you an example. Let’s say you’re a Machine Learning Engineer who just got an offer to take care of an anomaly detection model for a small e-gaming company that would like to check for bots in theirs in their game’s network.
You are quite familiar with anomaly detection algorithms like k-nearest-neighbor or support vector machines. But you haven’t worked with Alibaba Cloud and their entire infrastructure is based there, so they would like to use Alibaba’s resources. You can agree because you don’t want to lose a project but it will require weeks from you to feel confident in a different environment. Instead, you could offer migration to let’s say AWS for this particular task because you’re much more familiar with this platform.
You’d be able to offer a lower price and due to experience gained on a more popular platform, you will be able to use it in the future in the following projects. Think about it when new “occasion” pops up on the horizon and let’s move to the next mistake. The third mistake is kind of not intuitive because this approach works perfectly in other fields.
Namely, it’s “Reading too many books and articles”. Yep, I’m not a big fan of learning programming from books. Even if those books are highly technical and the topic might seem theoretical. Without hands on the keyboard, you will not get the essential skills needed to deliver a project in the future. Let’s say you’d like to learn about neural networks. You can buy a book and go through it but what I would personally do after learning from my mistakes, would be to look for some practical tutorials on how to build neural networks and I’d instantly jump to my programming environment, import necessary libraries, follow steps in the tutorial and without wasting time, I would just start building it.
Listen to the Godfather of the whole modern Artificial Intelligence. My feeling is if you wanna understand a really complicated device like a brain, you should build one. I mean you can look at a car and you think you can understand cars but when you try to build a car, you suddenly discover that there’s staff that has to go under the hood, otherwise it doesn’t work. Before learning tons of theory, just think of how you can use these skills and jump right into the building-mode filling in the theory leaks in the meantime.
And remember, books are gold but not necessarily while learning on how to code. We’ve got 3 mistakes behind but show must go on, so let’s talk about the fourth one. It’s “Using low-quality learning materials”. Well, it’s quite hard to distinguish valuable sources of information and those that aren’t worth our time. I felt into hundreds, literally if not thousands of useless learning materials.
What I think might be a good distinguishing feature is a highly practical approach. Spend one to three minutes to generally inspect the course you are going to take and check whether it’s based on some real project or it contains mainly theory. In the second case, just run away as fast as possible and don’t look back. It’s a good rule of thumb and it can save lots of your time. Eventually, it’s better to have fun with your friends, than learning a theory of Quadratic Discriminant Analysis and forgetting everything right after the course.
That was the fourth mistake. Let’s go to the last one. “Not managing my time properly”. Most of the learning occurs next to your computer and we live in the 21st century which means there are more distractions than hairs on your head. If you don’t manage your time, you simply end up scrolling social media and surprise, surprise – your Facebook wall isn’t bottomless without a reason. It’s very easy to sink in the ocean of social media information.
I caught myself multiple times doing things that I wasn’t planning and just wasting time without a reason. What really helped me were two things. First of all, I started writing down the start and end time of every activity I do at the computer. It let me focus on this particular thing and recall it when something distracts me. The second thing is taking regular breaks. I try to take a 10 to 15 minutes break every 1.5h of working.
Most of the time, senseless internet browsing comes from not long enough breaks. This way we get to the last mistake that slowed down my Machine Learning development. Using this occasion I’d like to ask you, whether such a short format of the movie is enjoyable and helpful for you. I tried to put everything in a nutshell so you get the most information in the shortest amount of time but I’m not sure if everything was explained clearly enough. I’d love to hear your opinion in the comments.
reference – Machine Learning: 5 Biggest Mistakes
Web enthusiast. Thinker. Evil coffeeaholic. Food specialist. Reader. Twitter fanatic. Music maven. AI and Machine Learning!