Skip to main content

Deep Learning Networks

The article linked below highlights deep learning, a type of machine learning that may be applied to several different types of domains within the networks space. Deep learning networks have already engaged concrete results across spaces such as cybersecurity, gaming, search engine optimization, and emotion identification, and they continue to hold the potential to become even more advanced in their technological capacities given the new innovation detailed in the Forbes article. Fujitsu Laboratories recent development allows for a deep learning network to manage the memories within its network layers more efficiently. Rather than using calculations within each network that depend on each other, Fujitsu’s innovation allows calculations that take place to occur independently of each other. This allows the optimization of memory management since calculations that are not needed in the future may be erased and more memory space can be made available for use.


Throughout our course, we’ve learned of how networks interact with each other and how nodes within networks also interact with each other. Rather than focusing on how nodes within a network might match up with each other, however, this article takes our course concepts to a level of greater depth by emphasizes how machine learning can function within a network to capitalize on memory space. A deep learning network specifically is made up of layers stacked on top of each other, and the processing that takes place occurs within each node. The nodes on each layer may have weighted connections to the nodes on layers above or below. As information passes through a network, each node processes the data and passes it alone to the nodes that occupy the following layer. In relation to the nodes within networks, we’ve learned about throughout the course, this article adds an interesting dimension we haven’t considered in terms of memory space and machine learning.



Leave a Reply

Blogging Calendar

October 2016