symbol grounding problem
In plain wordsThe puzzle of how words or symbols inside a machine can mean anything by themselves, rather than only pointing to more symbols — like a dictionary that defines every word using other words.
The problem of how the symbols a formal system shuffles could carry meaning for the system itself, instead of borrowing all their meaning from the people who interpret them. Stevan Harnad likened it to learning Chinese with nothing to go on but a dictionary written entirely in Chinese: each symbol leads only to more symbols, so meaning is never reached. His proposed remedy was hybrid: anchor basic symbols in the sensorimotor capacity to recognize and categorize what they stand for, then compose higher-level symbols out of these grounded ones. The problem is central to debates over whether language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code. trained only on text can mean anything, since their representations are learned from relations among words; replies appeal to multimodal training, interaction with the world through toolstool use, n.: An AI model's ability to reach beyond its own knowledge by operating outside tools — searching the web, running code, using a calculator or an app — and folding the results into its work. and bodiesembodied AI, n.: AI that has a body, usually a robot and sometimes a simulated character, and learns or acts by sensing and moving in a world rather than only processing text or images., and the possibility that the structure of language itself carries information about its referents.
Adding images to training eases the symbol grounding problem for some words but not for words like justice.
Origin
Named by Stevan Harnad in 'The Symbol Grounding Problem' (Physica D, 1990), which generalized John Searle's Chinese roomChinese room, n.: A thought experiment: a person who knows no Chinese follows a rulebook to answer Chinese messages so well that outsiders think they understand. If the person doesn't understand, does a computer running a program? argument; the octopus testoctopus test, n.: A thought experiment in which a clever octopus learns to impersonate a stranded islander by eavesdropping on an undersea cable, without ever knowing what the words refer to. It asks whether text-trained AI can grasp meaning. (2020) restated it for language models.
Related
- groundinggrounding, n.: Connecting an AI's words to something real — trusted documents, a live search, images or the physical world — so that its answers rest on evidence rather than only on what it absorbed in training.
- Chinese roomChinese room, n.: A thought experiment: a person who knows no Chinese follows a rulebook to answer Chinese messages so well that outsiders think they understand. If the person doesn't understand, does a computer running a program?
- octopus testoctopus test, n.: A thought experiment in which a clever octopus learns to impersonate a stranded islander by eavesdropping on an undersea cable, without ever knowing what the words refer to. It asks whether text-trained AI can grasp meaning.
- machine understandingmachine understanding, n.: Whether, and in what sense, an AI system grasps what words, situations or problems mean — rather than just producing answers that look as though it does.
- intentionalityintentionality, n.: The way thoughts, beliefs and desires are about things: a belief is about rain, a wish is for cake. Philosophers ask whether a machine's internal states can be about anything.
- embodied cognitionembodied cognition, n.: The idea that thinking is not something the brain does on its own: the body, its senses and movements, and its surroundings play a real part in how we think and understand.
- distributional hypothesisdistributional hypothesis, n.: The idea that words found in similar company — surrounded by similar words — tend to be similar in meaning, the insight that lets computers learn something about word meaning from large amounts of text alone.
- distributional hypothesisdistributional hypothesis, n.: The idea that words found in similar company — surrounded by similar words — tend to be similar in meaning, the insight that lets computers learn something about word meaning from large amounts of text alone.
Sources
- Harnad, S. (1990). The Symbol Grounding Problem. Physica D 42(1–3), 335–346.
Mentioned in
Chinese roomChinese room, n.: A thought experiment: a person who knows no Chinese follows a rulebook to answer Chinese messages so well that outsiders think they understand. If the person doesn't understand, does a computer running a program?, computational theory of mindcomputational theory of mind, n.: The idea that thinking is a kind of computation: the mind processes information by following rules, much as a computer runs a program, and the brain is the machine doing it., embodied AIembodied AI, n.: AI that has a body, usually a robot and sometimes a simulated character, and learns or acts by sensing and moving in a world rather than only processing text or images., embodied cognitionembodied cognition, n.: The idea that thinking is not something the brain does on its own: the body, its senses and movements, and its surroundings play a real part in how we think and understand., groundinggrounding, n.: Connecting an AI's words to something real — trusted documents, a live search, images or the physical world — so that its answers rest on evidence rather than only on what it absorbed in training., John SearleJohn Searle, prop. n.: An American philosopher who devised the Chinese room argument, which holds that a computer running a program can manipulate symbols without understanding anything., machine understandingmachine understanding, n.: Whether, and in what sense, an AI system grasps what words, situations or problems mean — rather than just producing answers that look as though it does., philosophy of artificial intelligencephilosophy of artificial intelligence, n.: The branch of philosophy that asks what AI is and could be: whether machines can think, understand or be conscious, what intelligence is, and what we owe to and should expect from intelligent machines., physical symbol system hypothesisphysical symbol system hypothesis, n.: The 1976 claim by two AI pioneers that a machine which stores and rearranges symbols according to rules has everything needed for general intelligence, and that anything generally intelligent must be such a machine., distributional hypothesisdistributional hypothesis, n.: The idea that words found in similar company — surrounded by similar words — tend to be similar in meaning, the insight that lets computers learn something about word meaning from large amounts of text alone., intentionalityintentionality, n.: The way thoughts, beliefs and desires are about things: a belief is about rain, a wish is for cake. Philosophers ask whether a machine's internal states can be about anything., knowledge argumentknowledge argument, n.: A thought experiment about Mary, a scientist who knows every physical fact about color but has lived all her life in a black-and-white room. When she first sees red, does she learn something new?, octopus testoctopus test, n.: A thought experiment in which a clever octopus learns to impersonate a stranded islander by eavesdropping on an undersea cable, without ever knowing what the words refer to. It asks whether text-trained AI can grasp meaning., visual groundingvisual grounding, n.: Linking words to the exact part of an image they refer to, for example finding 'the man in the red hat on the left' and pointing to him.