Symbolic artificial intelligence

In <a href="/facts/Artificial_intelligence/lJGauQwX">artificial intelligence</a>, symbolic artificial intelligence (also known as classical artificial intelligence or logic-based artificial intelligence)
is the term for the collection of all methods in artificial intelligence research that are based on high-level <a href="/facts/Physical_symbol_systems_hypothesis/mcfKIcyd">symbolic</a> (human-readable) representations of problems, <a href="/facts/Formal_logic/SaI7Y3wN">logic</a> and <a href="/facts/Search_algorithm/eJuw0nyz">search</a>. Symbolic AI used tools such as <a href="/facts/Logic_programming/Xgh8lj3F">logic programming</a>, <a href="/facts/Production_(computer_science)/J7E2FXKs">production rules</a>, <a href="/facts/Semantic_nets/mt7FlP53">semantic nets</a> and <a href="/facts/Frame_(artificial_intelligence)/YnQaR34X">frames</a>, and it developed applications such as <a href="/facts/Knowledge-based_systems/uocg0bY3">knowledge-based systems</a> (in particular, <a href="/facts/Expert_systems/CUCzSEaX">expert systems</a>), <a href="/facts/Symbolic_mathematics/hjun1KAX">symbolic mathematics</a>, <a href="/facts/Automated_theorem_provers/vKGvuKFI">automated theorem provers</a>, <a href="/facts/Ontologies/SQlhl6lp">ontologies</a>, the <a href="/facts/Semantic_web/NFHo9dB7">semantic web</a>, and <a href="/facts/Automated_planning_and_scheduling/VL1TcrmD">automated planning and scheduling</a> systems. The Symbolic AI paradigm led to seminal ideas in <a href="/facts/Artificial_intelligence/lJGauQwX">search</a>, <a href="/facts/Symbolic_programming/wg3xA0Dn">symbolic programming</a> languages, <a href="/facts/Intelligent_agent/8C2Q93Vw">agents</a>, <a href="/facts/Multi-agent_systems/A85vt3UA">multi-agent systems</a>, the <a href="/facts/Semantic_web/NFHo9dB7">semantic web</a>, and the strengths and limitations of formal knowledge and <a href="/facts/Automated_reasoning/x5EdkKoU">reasoning systems</a>.
Symbolic AI was the dominant <a href="/facts/Paradigm/HXpK3X7a">paradigm</a> of AI research from the mid-1950s until the mid-1990s. Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with <a href="/facts/Artificial_general_intelligence/vbxwaq8t">artificial general intelligence</a> and considered this the ultimate goal of their field. An early boom, with early successes such as the <a href="/facts/Logic_Theorist/cCCDAyCq">Logic Theorist</a> and <a href="/facts/Arthur_Samuel_(computer_scientist)/hjG9AfX2">Samuel</a>'s <a href="/facts/Arthur_Samuel_(computer_scientist)/hjG9AfX2">Checkers Playing Program</a>, led to unrealistic expectations and promises and was followed by the first <a href="/facts/AI_winter/DLpU1wxl">AI Winter</a> as funding dried up. A second boom (1969–1986) occurred with the rise of expert systems, their promise of capturing corporate expertise, and an enthusiastic corporate embrace. That boom, and some early successes, e.g., with <a href="/facts/XCON/XfAoPuoI">XCON</a> at <a href="/facts/Digital_Equipment_Corporation/ztxs6keA">DEC</a>, was followed again by later disappointment. Problems with difficulties in knowledge acquisition, maintaining large knowledge bases, and brittleness in handling out-of-domain problems arose. Another, second, AI Winter (1988–2011) followed. Subsequently, AI researchers focused on addressing underlying problems in handling uncertainty and in knowledge acquisition. Uncertainty was addressed with formal methods such as <a href="/facts/Hidden_Markov_model/ur1zTAhP">hidden Markov models</a>, <a href="/facts/Bayesian_reasoning/Cvn7WWfG">Bayesian reasoning</a>, and <a href="/facts/Statistical_relational_learning/tFA87qG1">statistical relational learning</a>. Symbolic machine learning addressed the knowledge acquisition problem with contributions including <a href="/facts/Version_space_learning/d8SSLoQS">Version Space</a>, <a href="/facts/Leslie_Valiant/hA5okAxL">Valiant</a>'s <a href="/facts/Probably_approximately_correct_learning/vBQlvDV1">PAC learning</a>, <a href="/facts/Ross_Quinlan/skDLL2u5">Quinlan</a>'s <a href="/facts/ID3_algorithm/cey9VCT7">ID3</a> <a href="/facts/Decision-tree/XDVeGlkJ">decision-tree</a> learning, <a href="/facts/Case-based_reasoning/JkxLXCqF">case-based learning</a>, and <a href="/facts/Inductive_logic_programming/gKc9Nxq4">inductive logic programming</a> to learn relations.
<a href="/facts/Artificial_neural_network/6V1jMlkx">Neural networks</a>, a subsymbolic approach, had been pursued from early days and reemerged strongly in 2012. Early examples are <a href="/facts/Frank_Rosenblatt/fdPnkXdt">Rosenblatt</a>'s <a href="/facts/Perceptron/ArxdkAC1">perceptron</a> learning work, the <a href="/facts/Backpropagation/lCsIdKHc">backpropagation</a> work of Rumelhart, Hinton and Williams, and work in <a href="/facts/Convolutional_neural_network/kHEdnmGU">convolutional neural networks</a> by LeCun et al. in 1989. However, neural networks were not viewed as successful until about 2012: "Until Big Data became commonplace, the general consensus in the Al community was that the so-called neural-network approach was hopeless. Systems just didn't work that well, compared to other methods. ... A revolution came in 2012, when a number of people, including a team of researchers working with Hinton, worked out a way to use the power of <a href="/facts/GPUs/PTK1RQVp">GPUs</a> to enormously increase the power of neural networks." Over the next several years, <a href="/facts/Deep_learning/JLuwD3ea">deep learning</a> had spectacular success in handling vision, <a href="/facts/Speech_recognition/z7S7Pgk6">speech recognition</a>, speech synthesis, image generation, and machine translation. However, since 2020, as inherent difficulties with bias, explanation, comprehensibility, and robustness became more apparent with deep learning approaches; an increasing number of AI researchers have called for <a href="/facts/Neuro-symbolic_AI/6FTW3UkR">combining</a> the best of both the symbolic and neural network approaches and addressing areas that both approaches have difficulty with, such as <a href="/facts/Commonsense_reasoning/HfL7zQe4">common-sense reasoning</a>.

Symbolic artificial intelligence open-in-new

Symbolic artificial intelligence