Decoding Human Intelligence: Stanford’s Latest AI Research Questions Innate Number Sense – A Learned Skill or a Natural Gift?

The ability to decipher any quantity is called Number sense. Number sense is key in mathematical cognition. Various activities, such as organizing large amounts into small groups and categorizing numerical quantities like numbers, are performed by our nervous system with ease but the emergence of these number sense is unknown. How numerical representations emerge in the human brain needs to be better understood. 

Stanford Human-Centered Artificial Intelligence (HAI) researchers claim that biologically inspired neural architecture can be used to understand the emergence of number sense. Using the neural architecture of cortical layers V1, V2, and V3 combined with intraparietal sulcus (IPS), the changes in neural representations can be understood. Analogous to the human brain’s visual cortex; V1, V2, V3, and IPS are visual processing streams in the Deep neural network. With deep neural networks at both the single unit and distributed population levels, neural coding of quantity emergence with learning can be investigated. 

Researchers at HAI find that due to the statistical property of images in deep neural networks, visual numerosity arises, and quantity-sensitive neurons emerge spontaneously in convolution neural networks, which were trained to categorize objects in standardized ImageNet datasets. Instead of using convolution neural networks, they used a number-DNN (nDNN) model with a biologically more plausible architecture. 

Most of the real-life Images consist of non-symbolic stimuli. They are mapped to quantity representations through numerosity training and interpreted. Researchers found that spontaneously tuned neurons change with numerosity training and lead to hierarchy. Similar to the procedures used in the brain for image studying, researchers implemented the representational similarity analysis to assess how distributed representations of numerical quantities emerge across the information processes.

Researchers at HAI experimented on numerical skills in children as they are often described as mapping non-symbolic representations to abstract symbolic representations. These are critical for the development of numerical problem-solving skills. These number sense and symbolic number processing capabilities rely on separate neural systems. Apart from the differences, they found that children often tend to learn small numbers by mapping them to non-symbolic representations and large numbers through counting and arithmetic principles. Studies also show that neural representational similarity between symbolic and non-symbolic quantities predicted arithmetic skills in children as parietal, frontal cortices, and hippocampus are positively correlated with arithmetic skills.

Most of the studies on neuropsychology are performed on animals to obtain data in understanding the emergence of cognitive reasoning. But animal brains have their limitations. It is unclear whether the way of understanding is actually the same as humans. The solution lies in research similar to HAI as it has important implications for understanding the development of cognitively meaningful number sense and learning of numerosity representations in children by training deep neural networks to perform activities like cognitive and mathematical reasoning.


Check out the Paper and Stanford Article. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to join our 26k+ ML SubRedditDiscord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more.

🚀 Check Out 900+ AI Tools in AI Tools Club

Arshad is an intern at MarktechPost. He is currently pursuing his Int. MSc Physics from the Indian Institute of Technology Kharagpur. Understanding things to the fundamental level leads to new discoveries which lead to advancement in technology. He is passionate about understanding the nature fundamentally with the help of tools like mathematical models, ML models and AI.

[Announcing Gretel Navigator] Create, edit, and augment tabular data with the first compound AI system trusted by EY, Databricks, Google, and Microsoft