Lee Bressler Delves into Deep Learning and A.I. Revolution

Equity fund portfolio manager Lee Bressler explores deep learning technology and how hardware-accelerated computing is influencing industries and processes across the board.

Deep learning and artificial intelligence are forecast to transform the fortunes of businesses and organizations across almost all sectors and industries, according to Lee Bressler, an equity fund portfolio manager.  The true power of machine learning and AI can only be entirely realized once the complete surrounding ecosystem is fully grasped.

“Immensely powerful computers and servers require low-latency connectivity and fast storage,” Lee Bressler explains, adding that it’s possible to enhance deep learning frameworks by implementing software capable of automating data science tasks, for example. The basic idea of machine learning, he says, is to effectively train a multi-layered neural network — a computer system that is modeled on the human brain and associated nervous system.

“Studies have shown,” Lee Bressler reveals, “that neural networks can reproduce essentially any function imaginable, although the ability to accurately configure them remains fundamental in this instance.”

What’s called for is the power to solve a myriad of small problems in a massive number of equally small steps, as opposed to attempting to achieve the same result via a lesser number of much larger steps, according to Lee Bressler.

“By employing this method, we are able to train neural networks to complete and solve tasks that were previously impossible,” he adds, “although, from a computational standpoint, this can remain expensive.” Costs are beginning to come down as new solutions surrounding parallel programming are being developed, the portfolio manager explains.

“Computers now are faster and more powerful than ever before,” he continues, “and today, the required data sets are large enough to achieve what would’ve been practically impossible just a few years ago.”

Via the implementation of convolutional neural networks, wherein each artificial neuron is carefully and precisely layered, new advances in machine learning are continually being made. “Sub-sampling and pooling layers are being exercised in order to facilitate the current deep learning and AI revolution,” points out Lee Bressler, which—he says—is actively countering previously seen poor performance surrounding machine learning technology.

“Deep learning and artificial intelligence both have an increasing relevance across almost every imaginable industry sector,” Lee Bressler adds, wrapping up, “as we continue to see the technology advance, offering ever-faster results and far greater depth of insight from the data we’re providing.”