How the Brain Informs Data Science
By Vincent Huang (Editor)
The human brain operates on just 20 watts of power—barely enough to run a small desk fan. Yet the brain has constructed civilizations from nothing, explained the universe from the microscopic to the interstellar, and appraised the meaning of life and our existence. The more we understand the physical and chemical intricacies of our brains, the more complex and unfathomable it appears.
The brain is incredibly complex. It has the ability to store around 2.5 petabytes of data and our unique and enigmatic attributes of creativity, abstract reasoning, and neuroplasticity. Although we have vast bodies of data regarding the structure of and networks which exist in the brain, we lack a fundamental ground-up understanding of how our brains work. For starters, we know very little about how information is encoded and transferred among neural networks. Scientists have translated our genetic code, yet we are far from full comprehension of the neural code that governs our thoughts and personalities.
Despite our lack of a complete understanding of our brains, scientists can apply what we know about the structure and function of the mammalian brain to bolster our technology in computing hardware and software.
In the brain, a neural circuit is a group of neurons connected by a complex net of axons, dendrites, and synapses that carry out a specific function when activated. Neural circuits come together and interconnect, forming large scale brain networks. These large-scale networks, often linking several different parts of the brain, are responsible for advanced cognitive abilities such as those activated when solving a math problem, recalling information, or performing complicated physical activities. Although the circuits used in connectionist artificial connectionist neural networks are much less complex than their biological counterparts, recent advances in computing power and the successful emulation of basic neural networks have led to various breakthroughs in computer science and our understanding of the brain. Artificial neural networks are computing systems inspired by biological neural networks of simple animal brains. Using artificial neural networks, programs can “learn” to perform certain operations without being specifically programmed through the consideration of certain cases. This allows computers to emulate the abstract capabilities of brains in tasks such as identifying pictures, analyzing market data to determine future trends and the best investments to make, and translating between languages.
 A tangle of mouse neurons.
Beyond computer software, algorithms, and data structures, our brains have also inspired innovations in processing hardware. The SyNAPSE chip is a neuromorphic computer processor developed by the Defense Advanced Research Projects Agency that emulates the architecture and function of the mammalian brain by using Very Large Scale Integration and analog circuits to mimic the neurological architecture of the mammalian brain. This allows neuromorphic computers to adapt to localized changes (neuroplasticity) and allows computers to better learn and adapt to changes. Together, these brain-inspired technologies form Artificial Narrow Intelligence, or ANI, that exceeds human intelligence in certain specific tasks such as data analysis, language translation, or chess.
 A 4x4 array of SyNAPSE Chips on a computer’s circuit board.
Emerging technologies such as artificial neural networks, computational plasticity, and advanced sensory perception coupled with these neurologically-inspired paradigms are changing the way information is created, analyzed, and manipulated. As we continue to discover more about the incredibly complex structure and function of our brains, not only will the fields of neuroscience and physiology advance, but so will the fields of artificial intelligence and data science—bolstered by our understanding and ability to replicate the incredibly powerful and flexible paragon of efficient functionality: the human brain.