Human-Machine Intelligence

Exploring the very elements of who we are and what we can do


Content

  1. The Human Brain

  2. Thoughts

  3. Emotions

  4. Decision Making

  5. Machine Intelligence

  6. Artificial Neural Net

  7. Bridging Human and Machine Intelligence

  8. Conclusion


The Human Brain


The human brain is an amazing biological organ that dictates a human's life altogether right from the time the person is born to death and everything in between. The human brain is one of the most complex systems in our universe that hides within itself a lot of secrets. Secrets that have boggled scientists of many years – like what are emotions, what are thoughts and ideas and how can they be studied and quantified.


To study the working of any system, we need to build it.


"what I cannot create I do not understand" - Richard Feynman

The idea of creating an artificial mind was lingering in the hearts of scientists for a long time. A mind that could think like humans and create like humans. It dates back to ancient Greece, where philosophers and mythical stories have paved its way to what we call Artificial Intelligence.

“I believe that whatever the universe is about to see in the future already exists in our past and present. Nothing is new. It's all hidden in the Womb of the Universe.”

Many authors still believe the same and one amongst them is Pamela McCormick author of Machines who think. She describes in her book that old fictional stories like Frankenstein which was published in the early 1800s involved a creature who was a hideous sapient created in an unorthodox scientific experiment — an artificial human. The fictional authors had already prophesied and saw dreams and visions of the world we are living in today. One of the interesting things to note is that most of these scientific ideas and notions come from fiction and imagination. And it is what has driven the human race to evolve and survive.


But can machines develop emotions and thought as we humans do? Or can humans quantify the same terms through mathematical equations and electronic devices?


To answer these questions we have to understand what thoughts and emotions are.


Thoughts


Thoughts come from nowhere and from everywhere! It contains an element of truth. Subjectively, our thoughts come from nowhere: they just pop into our heads or emerge in the form of words leaving our mouths. Objectively, we can say that thoughts emerge from neural processes and that neural processes come from everywhere. Thoughts are reflections of the experience over the span of life. What I meant by this is that the forms and dynamics of thought are influenced by everything that has a causal connection with you, your society, and your species.


Our brain likes to cluster likable information together with the input gathered from sensors in the body – like touch, vision, voice, taste, and smell. While clustering the information together it creates a pattern something like a "cloud of information" some sort of chain where all the information received is connected to each other. These clusters are not created randomly but quantitatively. The neuronal patterns that mediate and enable thought and behavior have proximal and distal causes.


The proximal causes are the stimuli and circumstances we experience. These experiences have causal impacts on our bodies and are also partly caused by our bodies. In the right circumstances, these nebulous patterns can condense into streams of thought.


The distal causes are our experiential history and our evolutionary pre-history. Our experiential history consists of the things we've learned, consciously and unconsciously, and the various events that have shaped our bodies and our neural connections in large and small ways.


To elaborate, our neural network in our brain works continuously 24/7 for all 365 days. Each of the neuron in the brain exchanges an electrical signal with each other anywhere from 20 to 1000 times a second and that it is too many electrical exchanges happening inside our brain with approximately 86 billion neurons. The objective of these neurons is to cluster likable information together for later use – such as thinking, storing, and recalling old memories, etc. For example, when we see words like bed, coffee, and exercise we can easily conclude that the words are related to a morning habit.


In a nutshell, thoughts are the result of the manifestations of the patterns created by the neural network.


Thinking


Thinking on the other hand is the process of sifting information such that it makes sense. For example, solving a mathematical equation, or planning a trip.


Daniel Kahneman in his book Thinking, Fast and Slow argues that there are two systems in the brain which he describes as system one and system two. Both of these systems determine our behavior.


System 1 is automatic and impulsive. It’s the system you use when someone sketchy enters the train and you instinctively turn towards the door and what makes you eat the entire bag of chips in front of the TV when you just wanted to have a small bowl.


System 2 is very conscious, aware, and considerate. It helps you exert self-control and deliberately focus your attention. This system is at work when you’re meeting a friend and trying to spot them in a huge crowd of people, as it helps you recall how they look and filter out all these other people.

Thinking is a tedious job hence most people refrain from doing it rather they would rely much on the intuitional responses that come from system 1. Hence they prone to errors and irresponsible responses. Only if we take time to think and engage in the process of sifting our stored information we could indeed right decisions and actions. Emotions are on the other hand are the responses that emerge from the same cloud of information.


The question arises "Can we control them in the moments necessary?"


Emotions


Lisa Feldman Barnett says:

"Emotions are guesses that the brain constructs at the moment where billions of brain cells are working together" and we might have control over them.

What we understood from the previous session is that the brain tries to find patterns by sifting the lifetime of experiences, reordering and making sense of it as much as it can by weighing the probabilities and trying to answer the question of what it could be...


When new information is received by the brain it stores it for future use by arranging it using some primal information. When a new piece of information is given to the brain it tries to find the match from the existing information or knowledge which is most probable or similar to the new piece of information received by the brain. Remember the example of the morning routine I explained earlier?


Predictions are the way your brain works.


Predictions are primal to any sentient beings. They are faster and reliable (not always). Using these past experiences the brain predicts and constructs our experience of the world. The brain builds these intuitions by the exposure it receives from the world around it. Or in other words, we become what we perceive. Our learning, our education, our society plays an important role in shaping our emotions and construct an experience of the world we live in.


System 1 does all the work that does not require any thinking, it is automatic and leverages no cognitive strain. System 2 on the other hand is lazy and most of the time it avoids getting any task. But system 2 plays an important role in build any sort of habits or emotions when they are repeated time and time again. Eventually transferring them to system 1 where it becomes automatic.


Let me explain with a scenario, our feeling might differ depending upon the environment we find ourself in. Imagine going to a Caffe' and smelling the aroma of freshly baked cookies, grounded coffee, and hearing a pleasing uniform noise of people chattering. The brain perceives all this information from the present and triggers a cue, possibly a certain churning in your stomach and thus in return will produce an emotional state so-called the feeling of being hungry. Now, imagine being in a hospital and seeing a patient with a deep cut wound being rushed to the emergency room, along with the smell of Accelerated Hydrogen Peroxide across the whole corridor. This time (as well) your brain perceives all this information from the present and triggers the same cue of churning in your stomach but this time your emotional state is not being hungry but rather being dreadful.


In the former scenario, it might be quite often that you might pass through a Caffe' on your way to work, and seeing that the environment is non-threatening your brain does not have to processes a lot of information and hence system 1 takes care of this situation. But in the latter scenario, the situation is different. If you are not a doctor then it is highly unlikely that you will ever go to the hospital. Thus when a normal person visits the hospital he/she feels dizzy that's because the emotions are complex reactions the body has to certain stimuli. Emotions are more or less controlled by system 1.


Lisa Feldman argues that all emotional responses are created similarly. Prior experiences and sensory inputs guide to action.

For instance, you might experience a spectrum of anger responses. Each comes with its own neural patterns as well as bodily changes and movements. The brain can generate any of these responses. It has selection mechanisms that determine which anger response fits the situation best.


But can these emotions be controlled?


Yes, if somehow we become attentive to our surroundings and engage our brain to think and control our physical responses we can control them all together. Some gurus ask us to take a deep breath to be aware of the situation we are in. Why does it work? Because deep breaths are controlled by us whereas normal isn't. When we do something where we are fully in control of ourselves we engage system 2, and it is wise enough to sift and collect appropriate responses.


Decision Making


Decision making is again related to the things that we have discussed previously. Right from thoughts to emotions and everything in between. Decision making again heavily relies on the two systems we discussed already. Most of the time people rely on their intuition i.e. system 1, where they don't think. Where they don't gather and process information. And most likely they are prone to make errors.


Errors in decision-making happen mostly due to the fact that we undermine the situation at hand especially if we have encountered any similar situation earlier. Again we can see that our prior information directly affects the decision that we make. The probability of a decision being right or wrong is 50/50, but the chances of a decision being right can always be higher we have the right amount of information and prior history of the event. A decision can only be examined thoroughly if:

  1. We have a sufficient amount of information and prior history.

  2. We make no early decision based upon our intuition.

  3. Consider each and every possibility that can occur.

  4. Map out the whole process to get a visual view.

  5. Delay the decision until the last moment.


Delaying the decision until the last moment is one thing that Kahneman forces us to do. He argues that if we have followed points 1,2,3 and 4 then our brain has gathered enough information to make decisions based on system 1 and system 2. In this manner, we have increased the probability of being right. By this, we are not entirely dependent upon our emotions but also taking time to think and analyze all the possibilities.


Machine Intelligence


So far we learn what thoughts, emotions, and decision making are from a neurological and psychological point of view. Now let's see how machines can offer their expertise to most of our everyday tasks.


To begin with, let's define what machine intelligence is.


A machine or a non-human system that can sift through a large amount of (relevant) data to come to a conclusion is called machine intelligence.


Recent advances in Artificial Intelligence or machine intelligence has shown how machines excel at repetitive tasks and not only that it can automate many processes, making them more efficient, sifting a large amount of data to analyze, conclude and predict events; helping humans to make vital decisions based upon the output they receive from the machines.


But, how do they achieve such a feat of accuracy and preciseness that we humans have started working together with them?


At the core of any machine-intelligence lie, a fundamental mathematical formula and a set of algorithms that when stacked together can break any complex problem into a simple fundamental solution, and from there humans can interpret the cause of the problem, build a predictive model or even generate similar complex output to study the effects by tuning the parameter of the algorithm.


The algorithm is designed by studying the working of the human brain, and it turns out that while designing an artificial brain we were actually being able to discover how our human brain functions. Machine learning offers hypotheses sophisticated enough to push forward our expanding knowledge of the brain; and insights from neuroscience guide and inspires AI development.


Another kind of phenomenon is called ‘replay’. The ‘replay’ is a key point of contact between the two fields because like the brain, AI uses experience to learn and improve. And each piece of experience offers much more potential for learning than can be absorbed in real-time–so continued offline learning is crucial for both brains and artificial neural nets.


Artificial Neural Network


An Artificial Neural Network or ANN is a computing system designed to simulate the working of the human brain which includes analyzes and processing information.


The ANN works exactly like the brain but it is more of a mathematical representation of the biological neural nets. As we discussed in the previous section that the machine intelligence at the core has a simple mathematical formula. And that formula is written as:


f(x) = y,

where x is the input and y is the output.


Essentially we use a linear equation to expand that function:


w.x + b = y

where w and b are the parameters that manipulate the input x in a way that it reaches the final output.


Although this formula does not seem much if this formula is stacked together in a particular order then it has the capacity to break the complex informational structure into simple ones and then we can derive a lot of processes that suit our needs.


The beauty of ANN lies in the fundamental concept of tailoring it according to our needs, or even according to the complexity of the data that we are dealing with. Once we give it enough data to learn from and good computing power it can achieve the unfathomable.


ANN not only decipher complex data structure into smaller ones but it also lays off humans to spend time on such tedious tasks which would take hours, days, or even months.


Bridging Human and Machine Intelligence


Although there is a widespread misconception that AI systems will gradually replace humans in most of the everyday work, that’s not true entirely. There may be certain jobs such as automation but the true power of AI lies in the collaboration with humans. For instance, in India where the population is approximately 1,380,004,385 and most of them being in the rural parts, the medical services and doctors are very few in numbers. And if there is a medical program conducted by any organization or institution then it would become very hard to cover such a large population. In cases like this, we can leverage the AI to recommend medical prescriptions and even analyze scans for a particular disease and give feedback to the radiologist or doctors so that they can necessary actions for the treatment of the patient.


In essence, machines are doing what they do best -- performing repetitive tasks, analyzing huge data, and handling routine cases. And humans are doing what they are best at doing -- resolving ambiguous information, exercising judgments in difficult cases, and dealing with patients or dissatisfied customers.


Not only is medicine but in other areas of life, AI is proving to be quite innovative. In arts, for example, artists can use AI to empower themselves with generative art. All they need is to feed some images into algorithms and the machine itself will generate a new image altogether. Architects, designers, musicians can all use the same technique to enhance their creativity, imagination, and skills.


Machine Intelligence is not an adversary but rather a tool that can improve the state of our life if used properly. Yes, it is smarter than us, faster than us, and does not get tired but it lacks the ability to understand things like we humans do. A machine needs a huge amount of data to learn patterns and come up with a conclusion. But a human does not require that much information maybe a couple of examples is enough for humans to understand the basic nature of the task, and this is what makes humans creative, and compassionate to other human beings and life forms. The slowness within us carries us through a journey of finite number and types of emotions, it helps us to imagine and dream things and situations of who we are and what we can become. This slowness and stillness help us to thrive in situations where there no hope. The same slowness helps us to share information and experiences with other companions and friends. An amazing thing about biological life forms such as ourselves is that we evolve and help others to evolve as well. The slowness of human beings helps them to reflect and it is in this reflection that new innovations and creativities are hidden and this is what makes us special and dominant over everything.


Machines are linear i.e. they do what they are asked to do but on the other hand, humans can possibly do anything.


Human-machine collaboration can open new doors of opportunities in all the known professions to humans. An architect can use it to generate a quick model for a client, an artist when finds no inspiration can find a new image which not only challenges but improves the skills as well, a musical can find a new chord or a new musical signature for herself. Doctors can predict a new virus before it being discovered. And likewise, find the biology footprint and develop vaccines.


Possibilities are endless only we are willing to accept AI as a friend and not an adversary.


Conclusion


At the core of everything, we should that human beings are still in control of most of the things that we do. Machine intelligence or AI can only be harmful or helpful when we want it to be. Ultimately the training process of the machines depends upon the information we feed into it, and the purpose for which we will be using it.


We haven’t yet created the algorithm which can bring emotions in machines exactly the way humans display it but work and the research is still going. What are sure of before developing emotions for the machines is that we want to create an intelligent system that can do all the task a normal human can do. This by the way is known as Artificial General Intelligence.


But before such time comes we will be developing machines that can do a particular task. And thus, we can rest assured that machines are friendly neighbors that can help us prosper and also help us evolve as a human being.


References

Books

  1. Thinking Fast and Slow – Daniel Kahneman

  2. How Emotions are made – Lisa Feldman Barrett

  3. The Organised Mind – Daniel Levitin

  4. Why We Sleep – Mathew Walker

  5. Human + Machines -- Paul R. Daugherty

  6. Stillness Is The Key -- Ryan Holiday


Podcast and Tedtalks

  1. Daniel Kahneman: Thinking Fast and Slow, Deep Learning, and AI | Lex Fridman Podcast #65

  2. Lisa Feldman Barrett: How the Brain Creates Emotions | MIT Artificial General Intelligence (AGI)

  3. Lisa Feldman Barrett: Counterintuitive Ideas About How the Brain Works | Lex Fridman Podcast #129

  4. You aren't at the mercy of your emotions -- your brain creates them | Lisa Feldman Barrett

  5. What is a Thought? How the Brain Creates New Ideas | Henning Beck | TEDxHHL

  6. Thinking Straight in an Age of Information Overload | Daniel Levitin | Talks at Google


Articles and Blogs

  1. The Strange Similarity of Neuron and Galaxy Networks

  2. Are Our Brains Wired for Categorization?

  3. On the complexity and the information content of cosmic structures

  4. Feeling Our Emotions

  5. Emotions and Types of Emotional Responses

  6. What Is a Thought?

  7. Replay by Deepmind