Numerosity processing in the miniature brain of insects- efficient processing enables bioinspired Neural Networks
Abstract
This project is concerned with bioinspired approaches to construct new design principles for efficient Artificial Intelligence (AI) solutions. The honeybee is an accessible model that has demonstrated novel solutions for information processing that translate to impactful bioinspired applications. This project uses psychophysics experiments with free-flying honeybees to understand how the insect brain can solve complex numerosity problems including arithmetic, manipulation, and a mental rotation of numbers along a number line, and the relationship between spatial and number-based information processing in brains. The project compares how a recently established capacity in insects to enable addition and subtraction using operators, and more complex problem-solving incorporating multiplication, can be implemented by simple neural structures. In parallel, new research on mental number lines reveals how numerosity information is ordered and mentally translated within brains (from insects through to humans). The psychophysics data will be used to construct Neural Networks (NN) that map to recently identified anatomical array structures within the insect brain that are known to be responsible for the learning acquisition of relational size information.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 16, 2024
- Source ID
- FA23862314063
Entities
People
- Adrian G Dyer
Organizations
- Air Force Office of Scientific Research
- Monash University
- United States Air Force