Concept and content
Artificial intelligence is a sub sector of computer science, its research involves not only this field, but also brain science, biology, neurophysiology, psychology, linguistics, logic, cognitive (thinking) science, behavioral science, mathematics, information theory, cybernetics, system theory and so on. In another word, artificial intelligence is a marginal discipline which is dedicated to develop functions related to human intelligence, for example, the ability to reason, to learn and to solve problems.
Main research areas
The research field of Artificial Intelligence focuses on six aspects: natural language processing, knowledge representation, pattern recognition, expert system, machine learning and robotics. it should be highlighted here that driven by big data, machine learning has gained new development opportunities since it provides a mountain of training data and verification data which are essential for this experiment. It can be said that nowadays big data is the basis of machine learning.
Natural Language Processing is an area of interaction between artificial intelligence, computer science, artificial intelligence, linguistics, mathematics and human language. It studies various theories and methods for effective communication between computers and humans in natural languages. In order to realize this communication, it is necessary to train the machines and make them not only understand the meanings of natural languages but also express themselves using human languages.
The widespread existence of this phenomenon makes it necessary to eliminate it by means of a large amount of knowledge and reasoning, however, current natural language processing technology has not yet reached this level. First of all, the grammar input of NLP technology so far is limited to the analysis of an isolated sentence due to lack of systematic research on the constraints and influences of the context of conversation. In addition, humans understand a sentence not only based on its grammar, but also a lot of relevant knowledge, including life experiences and expertise, which cannot be stored in the computer. Therefore, the language understanding system can only be established within a limited range of vocabulary, sentence patterns and specific topics; it will only be able to expand the scope appropriately after huge upgrades for the storage capacity and operating speeds of the computer.
From the current theoretical and technical status quo, high-quality natural language processing systems still requires more efforts, however, some practical systems with a certain level of processing capabilities have already been invented or even commercialized. Typical examples include natural language interfaces for multilingual systems, full-text information retrieval systems, machine translation systems, automatic summarization systems and so on.
Secondly on robotics, it is very important to firstly clarify is that robots and artificial intelligence are not the same. In fact these two areas are almost completely separate.
Traditional robots are not intelligent, not able to perceive objects and must work in a closed space due to the possibility that the machine can cause harm to humans. However, the combination of AI technology and robotics will make a turning point in this field. Although robots nowadays can only do limited things, the maturity of Artificial Intelligence will push robots into a new stage of development.
If successfully combined with computer vision and Artificial Intelligence, robots will be able to achieve self-determination and therefore to predict what is going to happen and act accordingly based on potential changes in the future. For example, once a person approaches a robot, the robot detects the situation and immediately stops actions to avoid injury. Collaborative robots have become a major trend in this field. In the future intelligent robots will need to work with humans and as a consequence this ability to determine and act is very important. In fact, Artificial Intelligence can provide robots with new capabilities in many other aspects. In terms of perceptual detection for example, robots will be able to understand human language and recognize faces through various sensors such as sight and hearing. The development of these functions further extends the application scenarios of robots.
Machine learning studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills, and reorganize existing knowledge structures to continually improve their performance. This is the core of artificial intelligence.
Pattern Recognition refers to the procedure of processing and analyzing various forms of (characteristic, literal and logical) information so as to characterize things or phenomena, i.e. to describe, identify, classify and interpret. It is an important part of information science and artificial intelligence.
The expert system is one of the three major models of AI and also the only one that does not require computing. It is an intelligent computer program system, which contains a large number of knowledge from human experts in a certain field and is able to deal with problems in the field by simulating the expertise and methods like human experts.
That being said, the expert system is a program system with expertise and experience. It uses artificial intelligence technology and computer technology to make inferences and judgments based on the knowledge and experience provided by one or more experts in a field. However, human knowledge is infinite and the way of thinking is diverse. To truly realize the simulation of human thinking is a very difficult task that requires development of many other disciplines.
In this information society, machines have gradually penetrated into every corner of our lives, completely changing the way we live, work and play. Examples include the smallest voice assistants like Siri to behavioral algorithms, search algorithms or automated cars and airplanes. Although the above achievements are enough to surprise us, this kind of Artificial Intelligence technology is still in its infancy. It is not necessarily the expected the case that many people claim or understand due to the fact that many AI applications now are only made of response algorithms based on predefined multifaceted inputs or user behavior.
Generally speaking, a true Artificial Intelligence system is one that can learn by itself, that can establish connections and gain knowledge without relying on predefined behavioral algorithms. True Artificial Intelligence can improve current iterations, become smarter and more sensitive and more importantly is able to enhance its own learning capabilities.
Today’s so-called Artificial Intelligence system is just advanced machine learning software with a wide range of behavioral algorithms that can adjust itself to users’ preferences and dislikes. While these machines are very useful, they are not becoming smarter or more intelligent but are only improving their skills and usability based on large data sets.
Computer vision refers to the ability of a computer to recognize objects, scenes, and activities from images. For example, doctors may use clinical expert systems to diagnose the symptoms of patients; the police uses computer software to identify portraits stored in databases to identify offenders’ faces; and of course our most used technology: license plate recognition.
Speech recognition is to convert speech into words and then identify, recognize and process the words. Main applications of this technology include medical dictation, voice writing, PC voice control and telephone customer service. The technical principle is to first process the sound, use the motion function to frame it and then conduct acoustic feature extractions on the waveforms after the frame. And finally combine the phonemes into words.
Natural language processing, like computer vision technology, combines various technologies that help achieve the goal to enable natural language communication between humans and the computer. Applications include machine translation systems and human-machine dialogue systems.
Deep learning is an important application area in the field of Artificial Intelligence. It plays a vital role in games such as chess, poker and go. For example, machines with deep learning capability can think about a large number of possible positions based on heuristic knowledge and calculate the optimal move.
Virtual Personal Assistant: The involvement of Artificial Intelligence is vital in the use of virtual personal assistants too. When the voice wakes up the virtual personal assistant, artificial intelligence equipment collects your command information, uses it to further identify your voice and provides you with personalized results.
Smart cars are now trying to replace people to operate the vehicle and therefore achieve automatic driving experience. Google’s project on self-driving cars and Tesla’s “autopilot” feature are two of the latest examples. Autopilot technology is undoubtedly based on Artificial Intelligence and is currently growing at an extremely fast pace. It is considered evident when Intel earlier this year acquired Mobileye, an auto-driving car company in Israeli.
Intelligent robots are capable of performing tasks given by humans. They have sensors that detect information from the real world in terms of light, heat, temperature, motion, sound, collision and pressure. They are equipped with efficient processors, multiple sensors and huge memory to demonstrate its intelligence and to learn from the mistakes so as to adapt to the new environment. At present intelligent robots can be seen everywhere in life, for example, sweeping robots and companion robots; these robots are inseparable from the support of Artificial Intelligence technology, whether they are chatting with people, automatically positioning navigation or conducting security monitoring.
The recommendation engine technology is to actively discover the current or potential needs of the users and push information to their browsing pages based on the results from algorithm analysis and processing which are generated from user behaviors and attributes.