Humans and certain animals appear to have an innate capacity to learn relationships between different objects or events in the world. This ability, Known as “Relational Learning,” is widely registered as critical for cognition and intelligence, as Learned Relationships are thought to allow humans and animals to Navigatee New Situations.

Researchers at Ml College in San Francisco and Columbia University Have Conducted A Study Aimed at Understanding The Biological Basis of Relational Learning by Using A Particular Type of Brace Al Network. Their work, Published in Nature neuroscienceSheds new light on the processes in the brain that would underpin real learning in humans and other organisms.

“While I was visiting columbia university, I met my co-author kenneth kay and we talked about his research,” Thomas Miconi, Co-Outhor of the Paper, Told Medical Xpress.

“He was training neural networks to do something called ‘transitive infererance,” and I didn’T know what that was at the time. The Basic Idea of ​​Transitive Infererance is Simple: ‘If A> B & B> C, then A> C.’ That’s a concept we’re all familyer with and is actually essential to a lot of our undersrstanding of the world. “

Past work indicates that when humans and some animals perform certain psychological tasks, they appear to grassp relationships between objects, even if there regarding In Tasks Known as Transitive Infererance Tasks, Thei can Figure out Ordering Relationships (IE, A is “>” Or “<" Than B, Etc. Ous comparisons ( IE, "a vs. b," "b vs. a," "b vs. c," etc.).

“In keeping with this, the ‘a,” B,’ c ‘are totally arbitrary stimuli, like odors or images, which doch do’ Give Away ‘The Relationship, “Explied Miconi. “If the Ordering Relationship is successful Learned, then Subjects Can Answer Correctly when they see ‘a vs. c’- c’ – hat’s transitative infection. Species (Such as rats, pigeons, and Monkeys) Get the correct answer on ‘a vs. c’ and other similes of stimuli never directly seen before (eg ‘b vs. f’).

Past Studies Found That when Trained on “Adjacent” Pairs of Stimuli (EG, AB, CD, etc.), Humans, Rats, Pigeons, Pigeons and Monkeys Can Learn to Correctly Gues the ORDERING RELEASHIP Efore (EG, AE, CF, etc.). The processes in the brain underling this well-Reported capability, howyver, remain poorely undersrstood.

“It was intriguing to hear about this ability and these findings, not only beCAuse of the Intuitive, Relational, and Combinatorial Nature of the Task (Who is Unconventional Among Currently Popular Tasks But also beCAuse Despite Considerable Study, We Still Do Not know how the brain learns orders in a way that automatically produced produces transitive infection, “said Miconi.

“In our discussion, one thing that made matters even in additional finding from past work: namely, that humans and monkeys (but not painons or rodents) having to bee existing knowledge of orders After encountering a small information. “

Interestingly, additional past research showed that if humans and monkeys successfully learned the order C> D, ” They will instantly know that “b> e.” This shows that their brains can re-organize previous knowledge based on new information; A process that has been termed “Knowledge re-casembly.”

“This struruck us as an additional ability worth looking into, since it is a simple yet dramatic institution of learning or acquiring knowledge,” said Miconi.

“At some point, we realized that it might be possible to get into how the brain has either of these abilities by taking the approach of an area in intelligence, in intelligence, the call Dea of ​​’Learning to Learn. “

“For an artificial system, the idea is that instalad of training the system (like a neural network) to give the correct answer for a particular set of stimuli (eg stimuli ‘a,’ b, ‘b,’ b, ”), Intead Train a system to learn by its correct answer for any new set of stimuli (eg stimuli ‘p,’ q, ‘r,’ etc.), much like like animals are tasked with doing in experiences. “

To explore the underpinnings of these Various Aspects of Relational Learning, Miconi and Kay LOKED to Emulate Relational Learning Using A Newly Developed Type of Artificial Neral Neral Network Inspiral Network Inspiral Miconi and Kay Assessed Whether this type of network was altar to learn relationships on its own, potentially mimicking the relational learning and knowledge re-case

“Maybe the most exciting part of this approach- Plausible mechanisms, “said miconi . “We thought it would be pretty convenent if machines could be part of the process to help us do this!”

The Artificial Neural Networks Utilized by the Researchers have a conventional architecture, but with a key unique feature. Specifically, the networks were augmented with an artificial version of “synaptic plastic,” which means that they could change their own change their own synaptic weights after fulfilling.

“These Networks can learn autonomously because their connections change as a result of ongoing neural activity, and this ongoing neural activity interesting activity,” ExPLANED MCONIIDII.

“The rationale for studying these networks is that their basic architecture and learning processes mimic those that of real brains. I had some existing code from from Previous work That i thought could be quickly re-arrived for this problem. By some kind of mirackle, it worked the first time, which Never Happens. “

Using some code that Miconi Developed as Part of His Previous Research, The Researchers Applied The Synaptic Plasticity-Augmented Artificial Neural Networks to Tast the Tast the Tasts Hiag S and animals.

They found that their neural networks should solve these tasks, and also consistently attained similar similer behaviors to those achieved by humans and some animals as documented in previous students.

“For example, one behavioral pattern is that performance is better for pairs of stimuli farther apart in the ordering (Eg B vs. f has higher performance compared to b vs. c). “What was also also exciting is that some of these experimentally observed behavioral patterns had never being explained in a model.”

Overall, The Recent Paper By Miconi and Kay Pin-Points Several Mechanisms that Cold underpin the Relational Learning and Knowledge Assembles of Abilities of Biological Organisms. In the future, the means that Identified Could Be Investigated Further, by Further Study of Eather Artificial Neural Networks or Humans and Animals.

“The More Specific Contribution of our work is the Elucidation of Learning Mechanisms for Transitive Infererance: In Particular, Learning Mechanisms which can explain a collection of behavior Work on Transitive Infererance, “said Miconi. “One Striking Result is that Meta-Laurning Approach Actually Found Two different Learning Mechanisms.”

The two learning mechanisms unveiled by miconi and kay vary in complexity. The first is Simpler and only allowed their neural networks to learn general relations, without re-casembling knowledge. The second is more sophistic This new pair.

“This deliberate, targeted ‘recall’ is what enables the network to perform knowledge reassembly, unlike the former, simpler one,” said Miconi.

“This is an intriguing parallel to the apparently different Learning Capacities Across Animal Species Docuted for Transitive Infection. Ference, but only primates see alle to perform this fast ‘reassembly’ of existing knowledge in response to limited novel information.

This recent study also highlights the potential of neural networks augmented with self-diarved synaptic plastic for study processes underpinning in humans and animals. The team’s methods should serve as an inspiration for future works aimed at explooring biological mechanisms using brain- Inspired Artificial Neural Networks.

“Nowadays, it is quite common to train and analyze artificial neural networks on single instals of a task, and this has been shown to be successful in discovery And Decision-Making, “said miconi.

“With plastic neural networks, this approach is extended to discovery Multiple tasks. “

The Initial Results Gathered by Miconi and Kay Cold Serve as a Basis for Future Efforts aimed at Shedding Light on the intricacies of Relational Learning. In future work, the results anticipate testing

“In the study, the system only ever performs one task –Larning the ordering relationship (‘a> b> c’),” Added miconi.

“This would be similar to an animal who has spent it life Rain a plastic network on a wide Range of Learning Tasks.

“Would such an agent be removed to generalize immediatily to a new Learning task that it didn’T see before, and what would it take it take for such an ability to emerge?”

© 2025 Science X Network