Fruit Fly-Inspired AI Takes on Bitcoin Trading

You can also read this news on BH NEWS: Fruit Fly-Inspired AI Takes on Bitcoin Trading

A digital experiment inspired by the neural network of a fruit fly has been set up by Alex Wormuth, a software engineer at Coinbase. The project, named Stonkfly, was launched with a $100 Bitcoin trading test and resulted in a $1 profit after its first day, demonstrating the unique implementation of neurobiology into financial transactions.

How the Experiment Functions

The system translates market data from Coinbase’s BTC USDC trading pairs into RGB visual representations. These visual inputs are then mapped to thousands of simulated sensory inputs corresponding with the fruit fly’s neural visual system. Decisions for trading are made according to this neural activity model, adapting biological strategies to assess market trends without involving a physical insect.

Stimuli and Response Mechanisms

The Stonkfly responds to profit and loss scenarios by activating dopamine neurons, specifically 15 PAM11 neurons in case of gains, and two PPL101 neurons during losses to simulate an avoidance response. This design aims to mimic a neural feedback system for reacting to market changes, yet the initial $1 gain is not seen as evidence of effective learning—potentially attributed to random chance and market movements.

Neuroscience has recently turned its focus towards projects like Stonkfly, with a previous announcement highlighting a complete neural map of the fruit fly’s brain. Scientists from Google Research and various institutions utilized AI to map 166,700 neurons, capturing the connectivity of the male fruit fly’s brain and nervous system. However, limitations exist as such simulations don’t replicate neurotransmitter processes and other intricate biological mechanisms fully.

Earlier innovations using the same MaleCNS model include integrating with games like Doom, Beat Saber, Super Mario 64, Minecraft, and Pong. Wormuth’s venture into linking this neurobiological simulation with financial trading embodies the latest frontier in applying this model outside the gaming sphere.

• The Stonkfly aims to translate market data into neural inputs replicable by the fruit fly brain’s structure.
• Initial results showed a minor profit, indicating potential yet highlighting the randomness of such early findings.
• Previous models like DOOMFLY illustrate similar attempts to bridge neurobiology with interactive platforms.

The Stonkfly experiment expands the horizon of neural network applications, yet its journey in establishing a reliable advantage in Bitcoin trading remains in preliminary stages. As this intriguing intersection of biology and economics progresses, it will be interesting to observe whether such systems can mature into consistent predictors of market behavior or remain innovative explorations without material advantage.