Log 2

The book, Machine Learning: Concepts, Tools and Data Visulization is written by the professor of Medical IT at Eulji University, Min-soo Kang, and Assistant Research Engineer Eun-Soo Choi. The book offers the information regarding Machine Learning to two main targeted audience, those who have no particular application in mind but want to start on artificial intelligence, and those who want to start to use AI on their fields of profession. For the first group of people, tha author gives the general introduction of AI on Machine learning, its algorithms, mathematics, logics. As for the second group of people, he offers practical matters  like coding that can be applied to their fields.  It appeals to similar groups of people who have no prior knowledge about AI that want to step their feet into the pioneer of the fourth industrial revolution, the artificial intelligence field, but with different targets and approaches and are of each other's great interest. 

I chose the book because even though it looks difficult,but after browsing through its preface, I believe it's a challenging book that I can learn a lot from! Plus, the Machine Learning itself interests me. I have always wanted to know the concepts about it, how we can use it, and most importantly, to know how we eactly made "it" and what are the difficulties we encountered. In other words, knowing what it is thoroughly is what I am interested in, to learn more about what we see as "magic" and how this "magic" was actually produced and what the thing is made of.  There are some main points regarding the book, the introduction to and the Articial intelligence itself, the definition, history, and classification, it also talks about the general concepts about Machine Learning, along with some case study. I would like to summarize the main information I learned from the book in the following.

If I want to learn Artificial Intelligence, the most important thing woulb be the Artificial intelligence itself, briefly, what is AI? This is the thought that always resonates in my head. Luckily enough, the book enhances my perception of Artificial Intelligence from the foundamentals I had, and it refined the fundamentals and built my understanding about the "machine" higher. This thing is larger, greater than a machine, is what insight I gained after reading the book, it's more human-like, way more than we think. It is equipped with the intellectual ability of human, which can be summarized in judgement, behavior, and cognition. The main idea is "thought", the thoguht, expereicne humans have, created artificially, that's artificial intelligence. To be more specific, is to make computers with minds, I would like to say this in a more abstract idea, to make them have what we call heart in humans, to let machines do problem soliving, decision making automatically by actual experiments to let it think, act, and learn from experience as human babies. But there are also difficulties doing so, I find that it's just not that easy to tell AI to just focus on being human, it actually needs a lot of approach because it's hard to tell a machine to think and act based on what it has heard or seen. It's not easy to make informal knowledge understandable to models because the logical notation needs a formal one, especially it's way harder when you aren't 100% sure about the knowledge you are giving to the computer, it's tough to put things you are not sure in a formal term that the logical notation of the system requires. Therefore, I learn from the book that by formnatting non-formatted items into formatted items so they can be handled by the machine's logical system. By formatting this, I got the idea that how AI can now think like human, it's because from the logical sytstem and the formatted information we put in, it can draw fits own conclusion and decide what should be done next, this is how we can not completely "transform" the computational system into human, but at least we can allow them to mimic human and think as the thought pattern of human. There's an interesting statement about discussing about whether this machine is human-like or not. "It is the study of how to make computers do things at which, now, people are better." I think it's quite an interesting statement becuase this is the act of letting machines perform functions that require human intelligence when we do it. So it is definitely not better compared to us, but through the act of keep behaving in a human way, I believe it can acquire the human intelligence and make people think it's another person talking to them when it's just a ronot that has gained human intelligence. Briefly, it's about the methodology and how feasible it is to implement human cognition, reasoning, and learning in a way that's defined clearly so the models can understand. Machine learning is atually similar, or mixed with a thing called "Data Mining", because it actually uses the same methods. However, there are actaully differences that we can tell each of them apart. If the thing is on training data, then it's main focus would be on prediction, while if it discovers attributes, the special pattern of some meaningful data within a large scale of unorganized data, it's data mining. Machine Leaning, or what we call ML in abbreviation somehow sparks interest on me, because I think it would be really cool if we can predict things based on our prior experience, for instance, will it rain tomorrow? How much of a traffic congestion we would likely to encounter if we go for work at 8 A.M., and what about 7 A.M.? So it's quite amusing because the books also talks about Supervised Learning, it's a method used in ML,it gives labels to data sets, so the data sets will only give back the results of the labels. It's useful when dealing with datas that are too much to handle and it can thus bring a better, calssified result. The book has given me a lot of insights and basic introduction about ML, also some really cool and useful figures that make me understand more about what it's talking about in texts. There are still a lot of concepts and derivatives of ML, such as deep learning, and how to visulize the data we get by ML, also how to actually get the data by some godo to use software made for the purpose of collecting data and predict it, I highly hope that I can get to know more information about these things regarding AI and the computer science field as well.