On Neuromorphic Chips: Advanced AI, Deep Learning, and Neural Network Computer Architecture

Neuromorphic Chips to Become Affordable

In the August 2020 edition of Communications of the ACM, West Linn, OR, USA-based author and journalist, Samuel Greengard, wrote Neuromorphic Chips Take Shape. The available data Greengard had is the claim that a Caltech (The California Institute of Technology) professor invented this technology, neuromorphic chips. Greengard wrote, “The concept of a brain-like computing architecture, conceived in the late 1980s by California Institute of Technology professor Carver Mead, is suddenly taking shape” (Greengard, Aug 2020, p. 9). Since the date, today, is around forty years later, this technology has remained dormant, but not in the research, thus so recent advancements probably allowed neuromorphic chips to become affordable (With LTE capabilities, see current pricing of a standard productivity PC, an i5 Microsoft Surface Pro: (link). With 3G capabilities, alternatively see the current pricing of a legacy, but GPU intensive PlayStation Vita: (link)).

Safely Surpassing the von Neumann Bottleneck

With maximum efficiency regarded, Greengard mentioned that this contentious model is aiming against the ineffective von Neumann bottleneck: this demands a processor enter an idle state, but it awaits memory data (see a fast external SSD: link) or other component data inclusively moving; however, this is an intense computation, and it plateaus exponentially worse problems. These exponentially worsening problems not inclusively are AI (see 2001: A Space Odyssey, a popular film regarding advanced data technologies including artificial intelligence: link), so thus this is a deep learning issue addition, but machine learning is also remaining a power-intensive issue, unfortunately.

This May be the Next Crucial Step Data Science Demands to Survive

On Greengard’s column, Chris Eliasmith, a Systems Design Engineering and Philosophy professor of the University of Waterloo located in Ontario, Canada, stated neuromorphic parallelism is a refreshing standard (see a book on parallelism because I read this as an undergraduate, but I do not recommend any particular book: link). Greengard wrote that Eliasmith said, “Neuromorphic chips introduce a level of parallelism that doesn’t exist in today’s hardware, including GPUs and most AI accelerators” (Greengard, Aug 2020, p. 9). According to Greengard, contemporary deep learning systems are dependent against software measurements to play oversimplified neuromorphic systems: they use standardized FPGAs (On field-programmable gate arrays, see this computer organization architecture book that I read because I was an undergraduate Computer Science major: link), CPUs (on Central Processing Units, see this text (this is the same Computer Organization and Architecture book): link), and GPUs (On Graphics Processing Units, see this text: link). As Greengard mentions, this technology as a single deliverable chip might be as effective as researchers have begun developing: touch-sensing prosthetics, anti-stroke or anti-Alzheimer’s brain implants, or self-healing electronic skin; or vision technology...other possibilities are predicting an earthquake or, I believe, an economic recession. According to Greengard, this technology shall be knowable as an embedded processor during the next year or two years on commercial markets.

With no many any longer populating the Garden of Eden, invest in Mental Stability.

In Scripture (NIV: link), there is a record of Adam being taken by ABBA, but then Adam being put in the Garden of Eden: Genesis 2:15. With that event done, then Adam and Eve were driven from the Garden of Eden (Genesis 3:24), investments became a recurring long-term proof that persons were willing to work. In the time of the Ecclesiates being written, this was probably proven by Solomon, kinsman of David, who invested in wisdom literature (Ecclesiastes 1:1), but this was a condemnation of Solomon because Solomon married seven hundred women and had his ways with three hundred concubines, worshipping himself through fertility rituals foreign to Israel (1 Kings 11). However, for others, Biblical wisdom literature has remained relevant over the generations since Solomon passed away because of this: mental illness doctrine like the APA’s (American Psychological Association) has remained unscientific to this day. The unscientific part was recorded in Genesis because Adam died for eating the fruit of the tree of knowledge of good and evil (Genesis 5:5), but more recently, in the APA’s case, not having a biological standard for typical or atypical antipsychotic medications, the related law enforcement conduct, and the court cases surrounding them then the culture supporting this entity. I wish Greengard or anyone he cited would have expressed interest for investing in R & D (see this Software Engineering text because the first phase of a Software Engineering project ought to involve Research and Development as project management does: link)  programming regarding the applicability of neuromorphic chips on mental illness cases, but none did, here. Hopefully, by the time neuromorphic chips are on the market, this shall come to pass, or maybe I shall have this opportunity, myself. 

Image by Gerd Altmann from Pixabay

References

Greengard, S. (2020). Neuromorphic chips take shape. Communications of the ACM, 63(8), 9-11. doi:10.1145/3403960