The wave of artificial intelligence is sweeping the world, and many words are always lingering in our ears: artificial intelligence (Artificial Intelligence), machine learning (Machine Learning), deep learning (Deep Learning). Many people always have a vague understanding of the meaning of these high-frequency words and the relationship behind them, so what kind of connection is there between them? Let's take a look below:
Artificial Intelligence: From Concept to Prosperity
In 1956, several computer scientists gathered at the Dartmouth Conference and proposed the concept of "artificial intelligence", dreaming of using computers that were just emerging at the time to construct complex machines with the same essential properties as human intelligence. Since then, artificial intelligence has been lingering in people's minds and is slowly incubating in scientific research laboratories. In the decades that followed, artificial intelligence has been reversing its polarities, or been called a prophecy of the dazzling future of human civilization, or thrown into the garbage heap as the fantasies of technological lunatics. Until 2012, both voices existed simultaneously.
After 2012, thanks to the increase in the amount of data, the improvement of computing power and the emergence of new machine learning algorithms (deep learning), artificial intelligence began to explode. According to the "Global AI Talent Report" recently released by LinkedIn, as of the first quarter of 2017, the number of global AI (artificial intelligence) technical talents based on the LinkedIn platform exceeded 1.9 million, and only the domestic AI talent gap reached more than 5 million. .
The research field of artificial intelligence is also expanding. Figure 2 shows various branches of artificial intelligence research, including expert systems, machine learning, evolutionary computing, fuzzy logic, computer vision, natural language processing, and recommendation systems.
However, the current scientific research work is concentrated on the weak artificial intelligence part, and it is very hopeful that a major breakthrough will be made in the near future. Most of the artificial intelligence in the movies are depicting strong artificial intelligence, and this part is difficult to achieve in the current real world. (Artificial intelligence is usually divided into weak artificial intelligence and strong artificial intelligence. The former enables machines to have the ability to observe and perceive, and can achieve a certain degree of understanding and reasoning, while strong artificial intelligence allows machines to acquire adaptive capabilities and solve some problems that were not previously problems encountered).
Weak artificial intelligence is expected to make breakthroughs, how is it achieved, and where does "intelligence" come from? This is largely thanks to one approach to artificial intelligence – machine learning.
Machine Learning: An Approach to Artificial Intelligence
At its most basic, machine learning uses algorithms to parse data, learn from it, and then make decisions and predictions about real-world events. Unlike traditional software programs that are hard-coded to solve specific tasks, machine learning is "trained" with a large amount of data, using various algorithms to learn from the data how to complete the task.
For a simple example, when we browse the online store, information about product recommendations often appears. This is the mall to identify which products you are really interested in and willing to buy based on your previous shopping records and lengthy collection lists. Such a decision-making model can help the mall provide customers with recommendations and encourage product consumption.
Machine learning comes directly from the early field of artificial intelligence, and traditional algorithms include decision trees, clustering, Bayesian classification, support vector machines, EM, Adaboost, and more. In terms of learning methods, machine learning algorithms can be divided into supervised learning (such as classification problems), unsupervised learning (such as clustering problems), semi-supervised learning, ensemble learning, deep learning and reinforcement learning.
The application of traditional machine learning algorithms in the fields of fingerprint recognition, Haar-based face detection, and HoG feature-based object detection has basically reached the commercialization requirements or the commercialization level of specific scenarios, but each step is extremely difficult until The emergence of deep learning algorithms.
Deep Learning: A Technique for Implementing Machine Learning
Deep learning is not an independent learning method, and it also uses supervised and unsupervised learning methods to train deep neural networks. However, due to the rapid development of this field in recent years, some unique learning methods have been proposed one after another (such as residual network), so more and more people regard it as a learning method alone.
The original deep learning is a learning process that uses deep neural networks to solve feature representation. Deep neural network itself is not a new concept, it can be roughly understood as a neural network structure containing multiple hidden layers. In order to improve the training effect of deep neural network, people make corresponding adjustments to the connection method and activation function of neurons. In fact, there are many ideas that have been proposed in the early years, but due to the insufficient amount of training data and backward computing power at that time, the final effect was not satisfactory.
Deep learning is destructive for a variety of tasks, making seemingly all machine-assisted functions possible. Driverless cars, preventative healthcare, and even better movie recommendations are all on the horizon, or on the horizon.
Differences and connections between the three
Machine learning is a method to achieve artificial intelligence, and deep learning is a technology to achieve machine learning. We use the simplest method - concentric circles to visually show the relationship between the three.
At present, there is a false and more common sense in the industry that " deep learning may eventually make all other machine learning algorithms obsolete ". This awareness is mainly due to the fact that the current application of deep learning in the fields of computer vision and natural language processing far exceeds that of traditional machine learning methods, and the media has made exaggerated reports on deep learning.
Deep learning, as the hottest machine learning method at present, does not mean that it is the end of machine learning. At least the following problems exist:
1. The deep learning model needs a lot of training data to show the magical effect, but in real life, there are often problems with small samples. At this time, the deep learning method cannot be used, and the traditional machine learning method can handle it;
2. In some fields, traditional and simple machine learning methods can be well solved, and there is no need to use complex deep learning methods;
3. The idea of deep learning comes from the inspiration of the human brain, but it is by no means a simulation of the human brain. For example, after showing a three- or four-year-old child a bicycle, even if the bicycle looks completely different, the child will Nine times out of ten, it can be judged that it is a bicycle. That is to say, the human learning process often does not require large-scale training data, and the current deep learning method is obviously not a simulation of the human brain.
When Yoshua Bengio, a deep learning tycoon, answered a similar question on Quora, there was a passage that was particularly good. Here is a quote to answer the above question:
Science is NOT a battle, it is a collaboration. We all build on each other's ideas. Science is an act of love, not war. Love for the beauty in the world that surrounds us and love to share and build something together. That makes science a highly satisfying activity, emotionally speaking!
The general meaning of this passage is that science is not war but cooperation. The development of any discipline has never been a one-way road to darkness, but peers learn from each other, learn from each other, learn from others' strengths, and complement each other, standing on the shoulders of giants. Forward. The same is true of machine learning research. It is a cult if you die, and openness and tolerance are the right way.
Combined with the development of machine learning since 2000, I am deeply touched by Bengio's words. Entering the 21st century, looking at the development of machine learning, the research hotspots can be simply summarized as manifold learning from 2000 to 2006, sparse learning from 2006 to 2011, and deep learning from 2012 to the present. Which machine learning algorithm will become a hot topic in the future? Wu Enda, one of the three giants of deep learning, once said, "After deep learning, transfer learning will lead the next wave of machine learning technology." But in the end, who's to say what the next big thing in machine learning will be.
Content source Zhihu problem, link: https://www.zhihu.com/question/57770020/answer/249708509