Will computer systems ever match or surpass human-level intelligence — and, if that’s the case, how? When the Affiliation for the Development of Synthetic Intelligence (AAAI), primarily based in Washington DC, requested its members earlier this yr whether or not neural networks — the present star of artificial-intelligence programs — alone will likely be sufficient to hit this purpose, the overwhelming majority stated no. As an alternative, most stated, a heavy dose of an older form of AI will likely be wanted to get these programs as much as par: symbolic AI.
Typically known as ‘good old style AI’, symbolic AI is predicated on formal guidelines and an encoding of the logical relationships between ideas. Arithmetic is symbolic, for instance, as are ‘if–then’ statements and pc coding languages comparable to Python, together with movement charts or Venn diagrams that map how, say, cats, mammals and animals are conceptually associated. Many years in the past, symbolic programs had been an early front-runner within the AI effort. Nevertheless, within the early 2010s, they had been vastly outpaced by more-flexible neural networks. These machine-learning fashions excel at studying from huge quantities of information, and underlie massive language fashions (LLMs), in addition to chatbots comparable to ChatGPT.
Now, nevertheless, the computer-science neighborhood is pushing arduous for a greater and bolder melding of the outdated and the brand new. ‘Neurosymbolic AI’ has grow to be the most popular buzzword on the town. Brandon Colelough, a pc scientist on the College of Maryland in School Park, has charted the meteoric rise of the idea in educational papers. These reveal a spike of curiosity in neurosymbolic AI that began in round 2021 and exhibits no signal of slowing down.
On supporting science journalism
In the event you’re having fun with this text, contemplate supporting our award-winning journalism by subscribing. By buying a subscription you might be serving to to make sure the way forward for impactful tales concerning the discoveries and concepts shaping our world right this moment.
Loads of researchers are heralding the pattern as an escape from what they see as an unhealthy monopoly of neural networks in AI analysis, and count on the shift to ship smarter and extra dependable AI.
A greater melding of those two methods may result in synthetic basic intelligence (AGI): AI that may cause and generalize its data from one state of affairs to a different in addition to people do. It may also be helpful for high-risk purposes, comparable to navy or medical decision-making, says Colelough. As a result of symbolic AI is clear and comprehensible to people, he says, it doesn’t undergo from the ‘black field’ syndrome that may make neural networks arduous to belief.
There are already good examples of neurosymbolic AI, together with Google DeepMind’s AlphaGeometry, a system reported final yr that may reliably clear up maths Olympiad issues — questions geared toward proficient secondary-school college students. However understanding how greatest to mix neural networks and symbolic AI into an all-purpose system is a formidable problem.
“You’re actually architecting this type of two-headed beast,” says pc scientist William Regli, additionally on the College of Maryland.
Confrontation
In 2019, pc scientist Richard Sutton posted a brief essay entitled ‘The bitter lesson’ on his weblog (see go.nature.com/4paxykf). In it, he argued that, because the Fifties, folks have repeatedly assumed that one of the best ways to make clever computer systems is to feed them with all of the insights that people have arrived at concerning the guidelines of the world, in fields from physics to social behaviour. The bitter capsule to swallow, wrote Sutton, is that point and time once more, symbolic strategies have been outdone by programs that use a ton of uncooked information and scaled-up computational energy to leverage ‘search and studying’. Early chess-playing computer systems, for instance, that had been skilled on human-devised methods had been outperformed by people who had been merely fed numerous recreation information.
This lesson has been broadly quoted by proponents of neural networks to assist the concept making these programs ever-bigger is the most effective path to AGI. However many researchers argue that the essay overstates its case and downplays the essential half that symbolic programs can and do play in AI. For instance, the most effective chess program right this moment, Stockfish, pairs a neural community with a symbolic tree of allowable strikes.
Neural nets and symbolic algorithms each have execs and cons. Neural networks are made up of layers of nodes with weighted connections which are adjusted throughout coaching to acknowledge patterns and be taught from information. They’re quick and artistic, however they’re additionally sure to make issues up and may’t reliably reply questions past the scope of their coaching information.
Symbolic programs, in the meantime, wrestle to embody ‘messy’ ideas, comparable to human language, that contain huge rule databases which are tough to construct and gradual to go looking. However their workings are clear, and they’re good at reasoning, utilizing logic to use their basic data to contemporary conditions.
When put to make use of in the true world, neural networks that lack symbolic data make traditional errors: picture turbines would possibly draw folks with six fingers on every hand as a result of they haven’t learnt the overall idea that palms usually have 5; video turbines wrestle to make a ball bounce round a scene as a result of they haven’t learnt that gravity pulls issues downwards. Some researchers blame such errors on an absence of information or computing energy, however others say that the errors illustrate neural networks’ basic incapacity to generalize data and cause logically.
Many argue that including symbolism to neural nets is perhaps the most effective — even the one — solution to inject logical reasoning into AI. The worldwide know-how agency IBM, for instance, is backing neurosymbolic strategies as a path to AGI. However others stay sceptical: Yann LeCun, one of many fathers of recent AI and chief AI scientist at tech large Meta, has stated that neurosymbolic approaches are “incompatible” with neural-network studying.
Sutton, who’s on the College of Alberta in Edmonton, Canada, and received the 2024 Turing prize, the equal of the Nobel prize for pc science, holds agency to his unique argument: “The bitter lesson nonetheless applies to right this moment’s AI,” he advised Nature. This means, he says, that “including a symbolic, extra manually crafted ingredient might be a mistake.”
Gary Marcus, an AI entrepreneur, author and cognitive scientist primarily based in Vancouver, Canada, and one of the vital vocal advocates of neurosymbolic AI, tends to border this distinction of opinions as a philosophical battle that’s now being settled in his favour.
Others, comparable to roboticist Leslie Kaelbling on the Massachusetts Institute of Know-how (MIT) in Cambridge, say that arguments over which view is correct are a distraction, and that individuals ought to simply get on with no matter works. “I’m a magpie. I’ll do something that makes my robots higher.”
Combine and match
Past the truth that neurosymbolic AI goals to meld the advantages of neural nets with the advantages of symbolism, its definition is blurry. Neurosymbolic AI encompasses “a really massive universe,” says Marcus, “of which we’ve explored solely a tiny bit.”
There are lots of broad approaches, which individuals have tried to categorize in numerous methods. One choice highlighted by many is using symbolic strategies to enhance neural nets. AlphaGeometry is arguably one of the vital refined examples of this technique: it trains a neural web on an artificial information set of maths issues produced utilizing a symbolic pc language, making the options simpler to verify and making certain fewer errors. It combines the 2 elegantly, says Colelough. In one other instance, ‘logic tensor networks’ present a solution to encode symbolic logic for neural networks. Statements could be assigned a fuzzy-truth worth: a quantity someplace between 1 (true) and 0 (false). This supplies a framework of guidelines to assist the system cause.
One other broad strategy does what some would say is the reverse, utilizing neural nets to finesse symbolic algorithms. One drawback with symbolic data databases is that they’re usually so massive that they take a really very long time to go looking: the ‘tree’ of all potential strikes in a recreation of Go, for instance, incorporates about 10170 positions, which is unfeasibly massive to crunch by means of. Neural networks could be skilled to foretell probably the most promising subset of strikes, permitting the system to chop down how a lot of the ‘tree’ it has to go looking, and thus dashing up the time it takes to choose the most effective transfer. That’s what Google’s AlphaGo did when it famously outperformed a Go grandmaster.
Another concept is to insert symbolics into the center of an LLM’s workflow, in the identical means as consulting a calculator would possibly assist individual to unravel a maths puzzle. Utilizing rules-based programs throughout essential reasoning steps will help to maintain LLMs from going off-track, many argue. Initiatives together with the Program-Aided Language (PAL) mannequin, for instance, use an LLM to transform natural-language duties into Python code, use that symbolic code to unravel the issue, after which interpret that answer again into pure language with an LLM.
Jiayuan Mao, an AI researcher who has simply accomplished her PhD at MIT with Kaelbling and is on her solution to the College of Pennsylvania in Philadelphia, has had success in utilizing neurosymbolic AI to make robotic coaching extra environment friendly. Her technique is to make use of a neural community to acknowledge objects (comparable to a purple rubber ball or a inexperienced glass dice) in a visible subject after which use a symbolic algorithm to cause by means of relational questions on these objects (comparable to ‘is the rubber object behind the inexperienced object?’). A pure neural community would want 700,000 examples in its coaching information to realize 99% accuracy on this process, she says. However by including symbolic strategies, she wants simply 10% of that quantity. “Even in the event you use 1%, you’ll be able to nonetheless get 92% accuracy, which is sort of spectacular,” she says. The same neurosymbolic system she created trounced a neural-network-based system at guiding a robotic that encountered unfamiliar objects whereas washing dishes or making tea.
Misplaced in translation
One of many huge challenges for symbolic AI is tips on how to encode typically slippery human data inside a language of logic and guidelines. One of many earliest makes an attempt was a mission known as Cyc, began by pc scientist Doug Lenat in 1984 and later overseen by his AI firm Cycorp, primarily based in Austin, Texas. The intent was to explicitly articulate commonsense details and guidelines of thumb, comparable to ‘a daughter is a baby’, ‘folks love their youngsters’ and ‘seeing somebody you’re keen on makes you smile’. The mission’s language, CycL, makes use of symbols (for logical operators comparable to IF, AND, OR and NOT) to precise logical relationships in order that an inference engine can simply draw conclusions, comparable to ‘seeing your little one would make you smile’.
Cyc, which now holds greater than 25 million axioms, has been utilized in quite a lot of AI efforts and has impressed different tasks, comparable to Google’s Data Graph, which holds greater than 500 billion details. In the present day, ‘data engineers’ use an identical technique to assemble human-generated details and relationships, construct specialised databases and combine them with AI.
Symbolic databases will help an AI to generalize data from one state of affairs and apply it in one other, says Kaelbling, which is a strong solution to make reasoning extra environment friendly. However there’s a trade-off in accuracy when coping with topics for which there are lots of exceptions to the ‘guidelines’ — not all folks love their youngsters, for instance, and seeing one thing you’re keen on doesn’t at all times make you smile. Symbolics needs to be included solely when it’s useful to take action, she says. “Cyc was making an attempt to show frequent sense into math. That’s virtually certainly a foul concept,” Kaelbling says.
In 2023, Marcus posted a paper with Lenat laying out what LLMs can be taught from Cyc. As a part of that work, the duo requested GPT-3, an early LLM of the sort that underpins ChatGPT, to write down CycL statements that encode the logical relationships within the sentence “Did you contact a blue object situated within the capital of France on September twenty fifth, 2022?” The response “at first amazed the Cyc crew”, they report, as a result of it generated what the proper kind of statements in the proper kind of language. However on nearer inspection, GPT-3 made many essential errors, they write, comparable to concluding that “the factor that’s touching the blue object is the date”.
“It appears prefer it’s good, it appears prefer it ought to work, however it’s completely rubbish,” says Colelough. This exhibits that it’s pointless to easily ram collectively a symbolic engine and a neural web, he says. “You then would possibly as properly simply not have the neurosymbolic system.”
What’s wanted, Colelough says, is much more analysis on AI ‘metacognition’ — how AI displays and conducts its personal considering. That may allow AI ‘conductors’ to supervise a extra refined integration of the 2 paradigms, fairly than having totally different engines merely take turns. Colelough says AlphaGeometry does this properly, however in a restricted context. If a versatile conductor that works for any area of information may very well be developed, “that might be AGI for me”, Colelough says.
There’s much more work to do. Recent {hardware} and chip architectures is perhaps wanted to run neurosymbolic AI effectively. In time, different forms of AI — possibly primarily based on neural networks, symbolic AI, each or neither — would possibly grow to be extra thrilling, comparable to quantum AI, a fledgling subject that seeks to take advantage of the properties of the quantum world to enhance AI.
For Mao, the last word purpose is to leverage neural networks’ studying skills to create guidelines, classes and paths of reasoning that people aren’t but conscious of. “The hope is that finally we will have programs that additionally invent their very own symbolic illustration and symbolic algorithms, in order that they’ll actually transcend what a human is aware of,” she says. That is perhaps like a pc discovering an as-yet-unknown mathematical or bodily idea — maybe analogous to π or the property of mass — after which encoding the brand new idea to assist to increase data. “We have to examine how computer systems can train people, not how people can train machines.”
This text is reproduced with permission and was first printed on November 25, 2025.
