In the first part on the history of Artificial Intelligence we explored the origins of this technology and the challenges it had to face to establish itself as a field of research. Today we will focus on the rebirth after the first winter and on the extraordinary progress that has been made from the 1980s to the present day, also delving into Machine Learning, Deep Learning and modern applications of AI.
Rebirth and Second Winter (1980-1990)
In the 1980s, artificial intelligence finally emerged from the so-called “first winter” and experienced a rebirth thanks to the introduction of expert systems. Expert systems were programs that used specific knowledge to solve complex problems in narrow domains.
Expert systems
The most famous expert system is called MYCIN, and was developed at Stanford University for the diagnosis and treatment of infectious diseases. Expert systems use a series of “If-then” logical rules, such as “If the patient has this symptom, then this condition should be checked”, in order to imitate the decision-making process that a doctor would make. Through a series of questions, MYCIN collects clinical information, formulates a diagnosis and suggests a treatment, also including probabilistic reasoning to manage uncertainty. MYCIN, despite its success, was never adopted on a large scale, but it laid the foundation for the rest of the history of artificial intelligence, having a significant impact on modern clinical decision support systems.
The second winter
At the end of the 1980s, artificial intelligence, in its history, found itself facing new difficulties. It turned out that expert systems were very expensive to develop and maintain, because creating and updating the rules required a lot of time on the part of human experts. Furthermore, codified knowledge tended to quickly become obsolete.
The lack of significant progress in this field, combined with the fact that the hardware and software of the time were not advanced enough, contributed to curbing enthusiasm for artificial intelligence. Research funding was reduced and companies shifted their resources to other sectors. This period is known as “the second winter,” and is a time of slowdown and disillusionment in the history of artificial intelligence.
The Machine Learning Revolution (1990-2010)
In the 1990s and 2000s, artificial intelligence encountered a second rebirth in its history, mainly due to the progress made in Machine Learning technology.
Learning Algorithms and Big Data
Thanks to Machine Learning, machines could start learning from data, rather than having to rely on a set of rigid predefined rules. Thanks also to the greater computing power of machines and the availability of a large amount of data, machine learning algorithms that are more sophisticated and effective than their predecessors have begun to flourish.
At this time, techniques such as Support Vector Machines (SVMs) and decision trees are being discovered as key tools of artificial intelligence. SVMs, for example, have been used successfully for classification and regression tasks, and have
Successes in Voice and Visual Recognition
Technological advances in artificial intelligence have enabled the creation of voice and visual recognition technologies. Google, for example, began in this period to exploit ML algorithms to improve the speed of transcriptions and decisions. Facebook implemented facial recognition thanks to artificial intelligence, which was able to identify people in photographs with increasingly better precision.
The Deep Learning Revolution (2010-Present)
Since 2010, we have seen another great revolution: that of Deep Learning. Deep Learning can be considered a subcategory of Machine Learning, and uses conventional deep neural networks.
Convolutional Neural Networks and Recurrent Networks
Deep neural networks have the ability to analyze and interpret very large amounts of complex data. Conventional neural networks, for their part, have had a great impact in the field of computer vision, in fact, thanks to them, object recognition and image processing have become more precise and faster.
Recurrent neural networks, on the other hand, improved efficiency in the fields of speech recognition and natural language processing, so that machines could understand and generate text more naturally.
AI in the Commercial and Daily Life Sectors
Such and many advances in the history of artificial intelligence have made its integration into numerous commercial products and services possible. Today, artificial intelligence is easily found in search engines, voice assistants and streaming and e-commerce platforms, in the form of recommendation systems.
But that’s not all: artificial intelligence, with its rich history, has also found applications in the real world, for example in self-driving cars. Such vehicles use a combination of Machine Learning, Deep learning and other technology to navigate the road and make decisions in real time, and represent one of the best examples of how artificial intelligence can improve our daily lives.
Conclusions
This evolution continues to push the boundaries of what AI can do, with implications that extend far beyond the technological realm, impacting society in increasingly profound and pervasive ways. Looking to the future, we expect the story of artificial intelligence to continue, continuing to integrate more and more into our daily lives and transforming the world in ways that are still unimaginable to us.