“The Forgetting Machine” a reflection

This book written by Rodrigo Quian Quiroga gives the reader an idea of the fundamental mechanisms by how we have a memory. Funny enough, the author pokes fun at some “meta” moments, demonstrating that his memory for his ideas, yet to be articulated, also undergoes the same process of neuronal impulses firing at specific frequencies for given ideas. His approach to the book aims to not overwhelm with scientific information that would discourage the layman from reading further but to pique interest in the subject. After all, the author references famous movies like “Space Odyssey 2001” and “Blade Runner.” Still, in the preview of the book, I was expecting more of an artificial intelligence perspective behind the challenges behind how machines aim to memorize and then apply to novel situations. This was not included in much detail in the book (aside from what I will discuss in a future paragraph), so the target audience would perhaps be someone who has a more cursory interest in the subject.

Personally, I have trouble reading philosophical excerpts from the many famous philosophers that we have seen in history. After all, I rarely delved into the subject matter myself, and I had other pressing matters regarding the knowledge in other subjects during school. Growing up, I thought philosophy was a subject meant for experts to discuss, and for experts to listen. Here, the examples listed by the author include material by Borges, Aristotle, Plato, and Locke. As such, the reader is taken on the journey to first think biologically, then psychologically, then philosophically, and interchange each discipline together to provide a wholesome understanding of the subject. For me, I knew some fundamental psychology and biology, but philosophy made me examine the text much less efficiently. Interestingly enough, I think a person who is well-versed in philosophy could have an easier time reading this book than someone who is more versed in biology.

The title also is an interesting choice. Does the “Machine” refer to us, always forgetting, always learning? In Chapter 5 of the book, we are introduced to who Shereshevskii was. An individual who seemingly did not have the ability to forget, but had a pitying trait of not being able to *understand* the significance of such knowledge. What good is a vast amount of knowledge on any subject you can think of, be it all of the country’s capitals in the world, the incessant memorization of 100,000 digits of pi (look up Akira Haraguchi), or all of the actors who played in the Lord of the Rings movies? Shereshevskii did not have the ability to analyze as well as he could memorize. In fact, the psychologist Alexander Luria gave Shereshevskii a task to memorize a large list of numbers, to which Shereshevskii was able to do so with no problem. Yet, when Luria asked Shereshevskii about the significance of the task, Shereshevskii had not realized, until Luria had told him, that this was a list of consecutive numbers. It is almost as if the brain’s resources, energy, and efficiency cannot produce a fully competent brain that has no tangibly observed difficulty. Surely, after reading about this example, one would be curious to see what exactly is the brain’s potential for competency and flexibility.

The book also mentions some experiments done to measure neuronal activation to different stimuli, including viewing pictures of celebrities. In marketing this book, the author could not help but reference a “Jennifer Aniston neuron,” which fires in response to a picture of Jennifer Aniston. Ultimately, the same effect was seen with other celebrities, such as Mark Hamill or Halle Berry. Each picture shown to participants had celebrities looking (older, makeup, hair styling) slightly different in every sequential photo, but the same neuron that was being observed would keep firing. This led to the discussion of neurons purportedly firing in response to the concept of these celebrities. For instance, Mark Hamill would dress up as Luke Skywalker, or have a more recent photo included in another movie being shown to participants. No matter the setting, the same neuron fired in sequence with being shown Mark’s face.

Even more interesting, was the evolutionary explanation behind this. Over time, it would be quite important to recognize certain faces, whoever they may be; our family members, friends, leadership, and I guess celebrities? Of course, of great importance relating to survival, we would have to know or understand who we can trust based on past experiences. Even viewing a photograph of the spelling of Mark Hamill’s name would also yield a similar response by the neuron. Invariably, I cannot understand how I have family members who will have no issue recognizing faces but have issues with remembering the names of those celebrities. Understandably, seeing a face from a film, then meeting that person on the street would undoubtedly leave you feeling very comfortable when seeing the celebrity up close, even though you two had never met in real life! I guess this could not work with shows like “America’s Most Wanted,” “Cops,” or some of the local news.

In closing this brief piece, I would like to finish with some information about the information processing power of the brain that may shed some light on the brain’s amazing information processing power. Acknowledging that the human eye has roughly 1 million retinal ganglion cells, we can calculate exactly how efficient our human brain is in processing visual information. (Retinal ganglion cells are cells connected to our retina, that are able to respond to light in the environment through their photoreceptors) Each ganglion retinal can transmit six to thirteen bits of information per second to the brain (depending on the environment, let us just say ten for simplicity). That means that the eye is sending about ten million bits of information to the brain every second from the environment, or 10 Mbps of information. In each day, the brain receives (if you are awake for 16 hours a day), 57.6 GB of information from the environment (multiply by 16 and 3600).

Now that we have calculated the amount of information sent to the visual cortex of the brain, let us consider resolution. According to Steve Jobs, in one of his announcements, he shared that the eye is barely able to “distinguish individual pixels in a rendered image with a resolution of 300 PPI (pixels per inch), from 30 cm away.” If one stood further away, resolution diminishes but the field of view expands. If one stands closer, the resolution improves, but the field of view shrinks. Taking for example a whiteboard in my room, a 24 in by 36 in whiteboard, from 30 cm away, I am able to see 57,600,000 pixels. However, most of what we actually see is not in detail unless it is in focus through the fovea. As our fovea’s area can be calculated by (pi * 0.15²) we can multiply by 300² PPI (because PPI is pixels per inch * pixels per inch again would yield pixels per inch squared) = 6,361 pixels. After multiplying this number by three, to represent three bytes for every pixel to see color, and then multiply by thirty frames per second, reflecting a standard video camera ability to gather information at that rate, we reach a value between 0.5–0.6 MB per second through the fovea.

But since a UPenn research team, according to page 25, asserts that the eyes transmit one MB per second to the brain, there is a discrepancy between the amount of information that can theoretically be sent by all those retinal ganglion cells and the information that is actually measured to be transmitted to the brain. The difference between 0.5–0.6 MB per second, and the 1MB per second finding also needs to acknowledge the area around the fovea and its image resolution in sending signals to the brain (think peripheral vision resolution).