The Emergence of Deep Learning in Bio-Mechanics
Dr. C. Sowmya Dhanalakshmi
Professor and HOD - ACADEMICS, Department of Mechanical Engineering, at SNS College of Technology, Coimbatore
#snsinstitutions #snsdesignthinkers #designthinking #snsct
Artificial intelligence emerged as a serious topic of research in the mid-1950s. At the time, researchers aimed to replicate human intelligence within the period of a research career. Hopes were dashed when it became evident that the algorithms and processing power available at the time were just inadequate. Some skeptics even dismissed the initiative as plain arrogance.
?
A renaissance has occurred in recent years, with software fashioned closely after networks of neurons in the brain demonstrating that AI's early promise may yet be achieved. Deep learning is a technology that use deep neural networks to learn abstract concepts and has already achieved human-level performance on some tests. Brain science encouraged the development of artificial neural networks, which created virtual neurons using software or hardware.
?
Early researchers in this AI subfield, known as connectionism, hypothesized that neural networks could learn complex tasks by gradually altering the connections between neurons, so that patterns of neural activity captured the content of its input, such as an image or snippet of dialogue. As these networks were exposed to more examples, the learning process would continue by adjusting synaptic strengths among the connected neurons to obtain more accurate representations of, say, sunset views.