The History of the Evolution of Artificial Intelligence

The series of history of technology will mix past, present and future in this new chapter. Let's count on history of evolution of artificial intelligence (IA) from the first research, through revolutions and coming in the possibilities for the future.

Inspiration and Representations in Pop Culture

This theme is so inspiring that Hollywood never stopped talking about him. Since film mudo "Metrópolis" de 1927, temos produções com robôs, computadores e programas que agem para o nosso bem ou em busca da nossa destruição. Rapidinho dá pra citar também "Blade Runner", "Exterminador do Futuro", "Eu, Robô", que é baseado na obra essencial de Isaac Asimov, e "2001: Uma Odisseia no Espaço" com o ameaçador HAL 9000.

Ideas related to Artificial intelligence are long before the emergence of technology that made all this possible. The human being always wanted a machine to do the job of acting and also think like it.

The First Steps of the AI

Studies from various areas began to go this way specifically during World War II. In 1943, the author I.M.A.C. presented an article that speaks for the first time of neural networks, artificial reasoning structures in the form of a mathematical model that mimic our nervous system.

Another important article of this time is Claude Shannon's work in 1950 on how to program a machine to play chess using simple and more efficient position calculations. That same year, the legendary Alan Turing developed a way to assess whether a machine can go through a human in a written conversation, which is read by an evaluator. That's the famous Turing test.

Initial Advances in AI

In 1951, the SNARC was born, a calculator of mathematical operations that simulated synapses, the connections between neurons. The person responsible for this was Marvin Minsky, who was a student of Turing.

Em 1952, Arthur Samuel cria um jogo de damas, o IBM 701, que consegue melhorar por conta própria e se torna um desafio para jogadores amadores. Tudo isso que falamos até agora é bem relevante, mas veio antes do período considerado o pontapé inicial da Artificial intelligence.

IA Marco Zero

The real zero mark was in 1956 at the so-called Dartmouth Conference. This meeting brought together large names of the area, such as John McCarthy, who named the field of artificial intelligence research. At this conference, the maximum sector has been defined: each aspect of learning or other form of intelligence can be described so accurately that a machine can be created to simulate this.

Technological Advances and Challenges

From there, the conference participants began to draw the ideas of the role. The possibilities were so exciting that private and government agencies invested heavily in the area, including ARPA, United States Advanced Project Research Agency, the same place where the internet was born.

The Perceptor and the LISP Programming Language

In 1957, Frank Rosenblatt presented Perceptor, an algorithm that classifies results, starting with a machine called Mark I Perceptor. In 1958, the LISP programming language emerged, which becomes standard in artificial intelligence systems and today inspires an entire family of languages.

Em 1959, foi utilizado pela primeira vez o termo "machine learning", descrevendo um sistema que dá aos computadores a habilidade de aprender alguma função sem serem programados diretamente para isso.

The First Chatbots and Robots

In 1964, he had the world's first chatbot, ELIZA, which automatically chatted and imitated a psychoanalyst using keyword-based answers. In 1969, Shakey was shown, the first robot that united mobility, speech and some autonomy of action.

The field of natural language processing was one of the most promising of the time, with various applications such as translators, language generation, speech recognition, among others. However, while there was much expectation, innovations did not happen as quickly as expected.

The Winter of Artificial Intelligence

Entre os anos 70 e 80, vivemos um período sombrio conhecido como "inverno da inteligência artificial". Essa era foi marcada por poucas novidades e cortes nos investimentos, resultando em baixa atenção ao setor. A área precisava se reinventar, e um dos campos que ajudou nisso foi o de sistemas especialistas.

First proposed by Édouard Feigenbaum in the early 1980s, the specialist systems are software that perform complex and specific activities of a field, playing the role of humans, but with much faster reasoning and a wider knowledge base.

Developments in Japan and the Fifth Computer Generation

Japan entered history with the so-called fifth generation of computers, an attempt to invest in the area of technology and modernize the entire industry between 1982 and the 1990s. This put Japan on the map of the new advances and accelerated sectors such as microprocessors and supercomputers.

However, the investment was controlled and the adoption of a programming language called Prolog did not have much membership. A second small winter of the AI happened in the mid-1990s, but was soon overcome.

Internet Explosion and AI Advances

The second half of the 1990s was marked by the explosion of the commercial internet. The networks took advantage of AI to develop navigation and indexing systems, resulting in programs that searched the network automatically and rated results.

Um marco importante foi em 1997, quando a máquina "Deep Blue" derrotou o campeão mundial de xadrez Garry Kasparov, provando que estávamos no caminho certo.

Robots and Automation

In 2002, Roomba became the first self-cleaning assistant robot, showing the efficiency of robots compared to previous decades. In 2005, Boston Dynamics introduced robots capable of moving through difficult terrain, with applications in various industries.

Developments Recent

Since 2005, artificial intelligence has also begun to be studied for application in autonomous cars. This area is complex as the platform needs to be connected to various vehicle sensors and traffic. DARPA has held an annual competition called the Grand Challenge since 2004, awarding autonomous car projects.

Em 2011, a Apple lançou o assistente virtual Siri, seguido pela Alexa da Amazon, que se popularizou rapidamente. Nesse mesmo ano, a IBM trouxe de volta os holofotes com o Watson, um supercomputador que venceu competidores no game show "Jeopardy!".

Education and Technological Advancements

After an experiment at Stanford University, the need for a technology-driven university arose, leading to the creation of Udacity. In 2012, Google consolidated its development technologies with deep learning, allowing algorithms to recognize objects in videos.

The Future of Artificial Intelligence

In recent years, AI has conquered several applications in the human environment. In 2014, the chatbot Eugene Goostman managed to convince jurors that he was human. In 2016, IA AlphaGo won the world champion of Go, showing that machine learning has evolved to an impressive level.

Building an artificial intelligence involves a number of algorithms and tools that dictate their behavior. Nowadays, AI is present in virtually all processes we perform in electronics, from streaming services to automatic responses in emails.

Conclusion

The trajectory of artificial intelligence so far is impressive and essential for our day to day. However, there is still much work ahead. Technologies such as Google Duplex, which talks via voice to schedule commitments, have a great potential to evolve. Adversary-generating networks are being developed to create content, and companies like Adobe will benefit greatly from it.

If you are interested in this theme, Udacity, the first artificial intelligence school in Brazil, offers specialized courses in the area for you to become an expert on the subject and build a reputable curriculum in the market.