The human brain is truly astounding. Think about how it manages to process so many incoming messages so quickly in order for us to function as human beings. In this manner, we can actually compare the human brain to a computer. Computers also work on complex algorithms in order to compute an outcome and can also do so at amazing speeds.
So, how long will it take before we can get a computer to function in much the same way as the brain does? The thought of creating artificial intelligence that is capable of creativity, emotion and free will is exciting and slightly scary at the same time. But, will we be able to see this happen in our lifetime?
In this book summary readers will discover:
- How the human brain stores information
- Why the neocortex of the brain is so important
- What has already been achieved in artificial intelligence
- What our future with artificial intelligence looks like
Key lesson one: How the human brain stores information
Our brains have an incredible ability to store and recall memories. Just think about all the times you have seemingly forgotten about something but then suddenly remember it when something triggers the memory. It’s how people remember their loved ones who have passed from a simple scent.
The fact that this occurs actually tells us how the brain organizes and stores information. Simply put, it uses patterns. When we pass someone on the street, we won’t necessarily recall their face with accuracy. However, police sketch artists employ a technique to stimulate our memory. They show people variations of facial features to trigger and recover these memories. These triggers can occur at any time and can involve any of your senses.
It is for this reason that many memory techniques involve patterns. If you can recall just one bit of the pattern, the rest should also fall into place. It’s like reciting the alphabet – if most kids hear the beginning, they can complete it. However, starting from Z and moving backwards will prove difficult. It is also the reason why when we hear a song, we usually need to hear the beginning and not just a random part in the middle in order to recognize it. The brain has the information stored in sequential order and anything out of order throws it off.
Key lesson two: Why the neocortex of the brain is so important
The majority of the brain is made up of its outermost layer known as the neocortex. It is responsible for body control, language comprehension and object recognition to name a few. The neocortex is also responsible for organising and storing information in sequence as mentioned earlier.
There is a pattern to everything we do whether it is driving to work in the morning or getting ready for bed. These patterns are stored in their sequence in cortical columns which are found in the neocortex. In total, there are approximately 500 000 cortical columns that contain 60 000 neurons each. Furthermore, each cortical column has about 100 neurons that make up pattern recognizers.
The pattern recognizers are divided into low-level and high-level recognizers. The low-level recognizers are triggered when specific pieces of a pattern are present. This information gets passed to the high-level recognizers which then put the pieces of the pattern together. Needless to say, these recognizers are constantly at work as our senses constantly receive input.
The neocortex also works in conjunction with other parts of the brain. The sensory inputs we receive travel to the sensory cortex, moving through the thalamus to end up in the neocortex. The thalamus is located in the midbrain and is responsible for the instinctive reaction to our senses. This means it decides whether the sense it received is pleasing or not. Once this occurs, the information is sent to the insula which is part of the neocortex. This happens continuously and is crucial to our functioning. Without a working thalamus, the chances of a coma or even death are extremely high.
The hippocampus also works together with the neocortex by telling it what to remember. When the neocortex receives sensory inputs, the hippocampus evaluates these inputs in order to determine whether it is new or something which it already knows. It is your hippocampus that recognizes when your partner has a new haircut or when you meet someone for the first time.
The neocortex also works along with the cerebellum to control our movements. Most of our movements are determined by the neocortex but the cerebellum is responsible for our basic functions. These are those movements that are done instinctually like catching something which has been thrown in your direction.
Lastly, the neocortex is important in humans because it where creativity and emotions originate. The insula of the neocortex has long neurons that connect parts of the neocortex and are called spindle cells. These spindle cells create feelings of love, sexual desire, anger and sadness. Creativity, however, is in direct correlation to the size of your neocortex. The bigger your neocortex, the larger your creative potential. Also, if you find ways to strengthen your neocortex, you could also become more creative.
Thus, the neocortex is at the very centre of what we do and how we do it. This fact is what makes it so important to those considering artificial intelligence. The thought of creating a nonbiological neocortex is the one that is central to creating artificial intelligence.
Key lesson three: What has already been achieved in artificial intelligence
The first studies of artificial intelligence began in the 1930s. It was also the early days of computing. Researchers had to figure out a way to develop a powerful processor and how to constantly feed it information in order for it to function as an artificial mind.
A huge breakthrough to this occurred in the 1980s when a method called the hierarchical hidden Markov model or HHMM was implemented. This model was a mathematical solution based on probabilities and enabled software to learn and evolve. Software that uses the HHMM model learns through pattern recognition, much like the neocortex.
When the model is presented with a specific dataset, it will make a prediction about what will come next in the data series based on its past learning experiences. If this prediction is correct, another prediction will be triggered in the sequence. This model has proved to be quite effective and is still in use today. It is popular in voice recognition software and is used every day when people ask Apple’s Siri a question.
However, as impressive as Siri can be, you can tell she’s not exactly close to being human. The best example of the possibilities of artificial intelligence comes from IBM’s Watson. Watson was able to read and comprehend all of Wikipedia along with 200 million other pages of information. Watson was even able to defeat other contestants on Jeopardy! – completely understanding the puns of the clues on the game show and answering correctly.
Watson was able to do this by taking all the possible answers to the question and statistically analysing this to choose the most likely answer.
Now the question remains whether Watson will be able to answer the complex problems of the world using the same reasoning process. With this possibility also comes the questions about consciousness and free will. Like every sci-fi movie dealing with artificial intelligence has taught us, do we have to be concerned about these machines developing free will? However, the concept of free will is also debatable. Research into the matter shows that the brain decides an action before we are aware of the decision we are about to make. If this is true, free will is actually a feeling more than an ability we possess.
Key lesson four: What our future with artificial intelligence looks like
Given everything that we know about the brain and the level of sophistication that computers now have, the creation of true artificial intelligence is not that far away. Technology is constantly evolving and this means that the requirements of artificial intelligence are closer than ever to be met.
Kurzweil describes the predictable growth of information technology as the law of accelerating returns. If technology follows this predicted trajectory, the simulation of the human brain will come in the next decade. An estimation of the processing power needed for a nonbiological neocortex is 1016 calculations per second. Japan has already produced a supercomputer with these capabilities. Kurzweil goes on further to predict that by 2029, examples of artificial intelligence will emerge in earnest with them taking their place in society by the 2030s. This may seem like wishful thinking to some but it must be pointed out that Kurzweil has a history of correct predictions – 98 per cent accuracy to be specific.
The key takeaway from How to Create a Mind is:
The knowledge that researchers have today regarding the human brain and how it works is an important step in creating true artificial intelligence. We have come a long way since the 1930s and the beginning of research into computers and artificial intelligence. Technology is evolving fast, and soon enough, it will be at the level that is needed to create a digital neocortex. For now, the research continues but every step made, is a step closer to having artificial intelligence a part of our daily lives.
How can I implement the lessons learned in How to Create a Mind:
Now that you know how the neocortex works, why not use that to your advantage. You can train your brain to recognize patterns as a memory technique. You can also work with people who are more creative than you to strengthen your neocortex and thus improve your own creativity. Collaboration is a very effective means to learn and will contribute to your self-practice.