On Machine Learning
From earliest of times, people have been using tools which simplified and enriched their lives. At first humans used natural tool such as stones to manufacture necessary items, and animals as a source of food, transportation and clothing. Then people learned to design more sophisticated tools such as engines, planes and cars. So the tools were progressing from the simplest to more advanced by the means on human intellect. In twentieth century people invented computers which could perform millions of operations in a fractions of second, this allowed people to scale things up in an unprecedented levels. Currently, people are trying to use these computers to produce the intelligence, a mighty feat if it is to be accomplished. But the new tool has a potential of altering our very lifestyle. If earlier tools were intended to quicken thing up, the tool of artificial intelligence has a potential of turning things upside down. For this reason, I believe, we need to be aware of general developments in the field of artificial intelligence and machine learning. On the one hand, this field is a source of challenging problems, one the other hand it has a potential of making our lives so much better, if used wisely.
A field of computer science (CS) for me is science of scaling things up. For instance, as it becomes hard to check if given number is prime, so we need to use computers for large scale tests. From practical side, it might be easy to store and manipulate with a single data entry, but manipulating millions of data points requires architecture and methods from computer science. Most of the computer science appears to be hard coded, in a sense that structures and methods are fixed. This contrasts our experience, where our way of thinking is much softer or nearer to heuristics, I would say. Machine learning is a way to soften up the computations. Machine learning algorithms extract patterns from the data. The data might come in different shapes and forms, what matters is certain relationships between its elements. In this regard, humans can be considered as walking data processors. We observe multitude of data coming from all our sensory organs, and extract some patterns necessary to take actions.
To make our discourse more specific, let us take an example of driving. A driving involves several senses such as vision, hearing and feeling, maybe. A driver needs to process visual information from the road, watch out for honks of the cars behind, and make sure that the car is driving as usual. If one ponders about driving experience, then one could realize that it is not that complex and it is amenable for simulation by algorithms. Indeed, we are witnessing more companies producing self-driving cars. Here comes a bad news. If all vehicles become self-driving what should professional drivers do. Here comes a civic responsibility of people. A society needs to find means to accept ex-drivers and show them the ways of carrying on. On the other hand, it is hard to imagine driving job to be exciting. Those people who quit driving might enjoy working in some exciting areas, which offer wider and deeper variety of tasks. But for this to happen those drivers and people overall need to learn how to speak to machines, as our lives is going to be more interconnected.
For these reasons, namely intellectual curiosity and civic responsibility, I want to learn some machine learning. Currently, a field of machine learning is extremely vast so that it is infeasible to exhaust it. However, one could start be learning techniques which passed a test of time, such as linear and logistic regression, tree methods and naive Bayes methods. On top of that there are many exciting things happening at the boundary of the current knowledge. Most of these activities seem to be related to neural networks, or deep neural networks, which sounds fancier. I hope to learn some of machine learning and share my experiences in this blog.
A field of computer science (CS) for me is science of scaling things up. For instance, as it becomes hard to check if given number is prime, so we need to use computers for large scale tests. From practical side, it might be easy to store and manipulate with a single data entry, but manipulating millions of data points requires architecture and methods from computer science. Most of the computer science appears to be hard coded, in a sense that structures and methods are fixed. This contrasts our experience, where our way of thinking is much softer or nearer to heuristics, I would say. Machine learning is a way to soften up the computations. Machine learning algorithms extract patterns from the data. The data might come in different shapes and forms, what matters is certain relationships between its elements. In this regard, humans can be considered as walking data processors. We observe multitude of data coming from all our sensory organs, and extract some patterns necessary to take actions.
To make our discourse more specific, let us take an example of driving. A driving involves several senses such as vision, hearing and feeling, maybe. A driver needs to process visual information from the road, watch out for honks of the cars behind, and make sure that the car is driving as usual. If one ponders about driving experience, then one could realize that it is not that complex and it is amenable for simulation by algorithms. Indeed, we are witnessing more companies producing self-driving cars. Here comes a bad news. If all vehicles become self-driving what should professional drivers do. Here comes a civic responsibility of people. A society needs to find means to accept ex-drivers and show them the ways of carrying on. On the other hand, it is hard to imagine driving job to be exciting. Those people who quit driving might enjoy working in some exciting areas, which offer wider and deeper variety of tasks. But for this to happen those drivers and people overall need to learn how to speak to machines, as our lives is going to be more interconnected.
For these reasons, namely intellectual curiosity and civic responsibility, I want to learn some machine learning. Currently, a field of machine learning is extremely vast so that it is infeasible to exhaust it. However, one could start be learning techniques which passed a test of time, such as linear and logistic regression, tree methods and naive Bayes methods. On top of that there are many exciting things happening at the boundary of the current knowledge. Most of these activities seem to be related to neural networks, or deep neural networks, which sounds fancier. I hope to learn some of machine learning and share my experiences in this blog.
Comments
Post a Comment