The strategy to solve the Machine Learning & Deep Learning problems
- The strategy to solve the Machine Learning & Deep Learning problems
- What I liked about this book
- Overall Review
- Reference
The strategy to solve the Machine Learning & Deep Learning problems
It was honor to be nominated to review one of the books that’s going to be published by the GoldenRabbit as a Deep Learning Engineer.
The field of AI(Artificial Intelligence) has been growing so fast that some of misbeliefs do exist, but people are willing to learn them because the idea of deep learning and machine learning is cool. Don’t get me wrong, I may sound a little bit skeptical about deep learning or machine learning field, but I am actually not. It’s more of the people who deliver the wrong idea and conception to people.
What I liked about this book
Let’s get back to the review part of this book. Since this is written in koreank, some expression might not make a sense! In this book, the author talks about how machine learning or deep learning problem can be easily approached by the Kaggle, which i felt very positive and practical about it. For the beginners, they don’t realize how important the data consist of or the deep learning itself is data-driven, which means that there are some importance of curating(collecting) the data as mentioned by Andrew Ng. In reality, it’s very hard to collect the data and process them, but people who practice the ML/DL design in Kaggle(Kaggler) already made the data, even more lots of company opens the competition for kaggler, which means the beginner don’t have to make or collect the data.
As a ML/DL engineer or data scientist, it’s crucial to know how the data are formed, that is why we need to analyze the data and visualize how the data are distributed. This process is called the EDA(Exploratory Data Analysis). This book talks about the visualization as shown below
[Figure] This figure was edited because of the intellectual property rights.
Also, this book uses all the resources from the most popular example from Kaggle such as the London bike sharing data, Porto Seguro’s Safe Driver, and Predict Future Sales, which they are mostly solvable by the machine learning algorithm. Furthermore, the book helps reader to expose actual kaggle competition such as the Aerial Cactus Identification, Cassava Leaf Disease Classification, and RSNA Pneumonia Detection.
Even though it’s hard to write a book considering who’s going to be the reader of this book. This book was well-written in overall. There are some misconceptions and improvements needed, but i felt that the auther did a descent job on explaining a machine learning concepts.
Overall Review
This book was great to people who’s going to start implementing the ML / DL. But we all know that we have to understand the concepts behind all application. Knowing that DL and ML is not a tool for everyone, it’s a tool for specific problem.
