Researchers from the Hebrew University have proposed a new theory about how animals store and recover stored food. Their research, published in Scientific Reports, challenges traditional notions of animal storage behavior by suggesting a mechanism that is not based on memory.
A new study at the Hebrew University offers a new mechanism that is not based on memory for the way animals store food and find it easily when needed. Instead of relying on memory, the researchers suggest that animals use a neural mechanism similar to HASH functions in computing, which allow efficient storage and retrieval of the storage locations. This is important because it challenges long-held beliefs about animal cognition and offers a more efficient explanation for how animals can manage thousands of storage locations without overtaxing their memory systems. The proposed mechanism may have far-reaching consequences for our understanding of animal behavior, brain function and even the development of new artificial intelligence systems. By providing a simple and scalable model for information processing in the brain, this research opens new avenues for the study of cognitive processes in both animals and humans.
The researchers, Dr. Oren Purkosh and Sharon Mordechai from the Department of Cognitive Sciences and the Department of Animal Sciences explain that contrary to the long-held belief that food-dispersing animals rely on memory to find stored food items, Purkosh and Mordechai propose a static mechanism similar to the HASH functions used in computing. HASH functions in computing are algorithms that convert input data of any size into a fixed-size character string, which usually represents the data in a unique and efficient way.
The spatial cells of the hippocampus: The researchers' mathematical model aligns with the activity of the spatial cells of the hippocampus, which respond to the animals' spatial attention. The remapping ensures that these cells will be activated consistently on repeat visits to the same region but will vary between regions.
Functions HASH continuous: This remapping, along with unique cognitive maps, creates persistent HASH functions that can aid in both food storage and retrieval.
Neural network architecture: The study presents a simple neural network architecture capable of generating a unique probabilistic HASH for each animal, providing an almost unlimited capacity for encoding structured data.
The proposed framework includes a biologically plausible implementation of HASH through a neural network. The input layer encodes key environmental landmarks, while the output layer designates food storage locations. The two layers are arranged in a two-dimensional grid, with each cell corresponding to a specific location. The storage site is determined by the activity level of the output neurons, known as the storage score.
This novel approach offers a new perspective on animal behavior and cognitive processes, suggesting that animals may use non-memory-based mechanisms for complex tasks such as storage. The findings may have broader implications for understanding brain functions and developing artificial intelligence systems.
More of the topic in Hayadan:
Comments
Hash is a method of encoding information into memory. It is definitely a memory, so I don't understand the claim in the article that it is not a memory.
What the researchers have shown is that there is a biological feasibility of using a hash-like data generator.
It would have been appropriate if you would have mentioned that the article was written by Chat Gipiti.
Unintelligible article!
There is a lack of a practical explanation of how with is used in a similar way to a function and any example would help