Let’s discuss the Theory of Mind and Artificial Intelligence
Can machines think? Can they do what a human brain can do? For ages, scientists have been trying to develop such machines that can replicate the human brain and can perform tasks as equivalent to humans or better than them.
Unfortunately, such machines haven’t been developed because they lack the two fundamental properties of the human brain, one is cognition and the other is consciousness.
Computational Theory of Mind and Artificial Intelligence
This brings us to the main theme of the article “The Computational Theory of Mind (CTM) and Artificial Intelligence (AI)”.
According to CTM, a human brain is an information processing system. It receives the input and makes the decision. But how does all this happens??. The human brain is a container that consists of neurons, connected to each other, which fire up at certain points, and thus the human brain is capable of making a decision.
Apart from these the two important factors to consider here are consciousness and emotions. Humans have different emotions at different times. These are the missing blocks in a machine. No matter what advancements can be made in a machine they can not have emotions like humans. They can just make decisions but they can not sense it.
According to CTM minds possesses
a representation of data. It may be visual or in
the form of language.
The human brain encapsulates the incoming data and
then draw results by understanding the meaning of symbols. When talking about
machines, they interpret the data by encoding and decoding the symbols so it
becomes important for a machine that how to encode the characters and how to extract
special meanings from them .
Talking about the fiction we can see in movies that human
defying machines have been developed.
Consider the example of ‘ex-Machina’ or ‘The Terminator’ which presents with a model which can defy humans and develop senses but still we are very far away from such systems.
Coming towards AI, it tends to produce human thinking capable machines by certain algorithms whose basis lies in mathematics. An algo is just a step by step procedure that how to tackle a problem. According to Warren McCulloch and Walter Pitts (1943), the operations performed by neurons are computational. They also considered that neural activities can explain cognition. Until the earlier 20th century mathematicians have been relying on notions of algos but the paper from Alan Turing “On Computable Numbers, With an Application to the Entscheidungsproblem”, provided such an approach which was most impactful.
The Turing machine, proposed as a model of computation devised an abstract machine that concludes special meanings from different notions. Alan Turing is considered the father of Artificial Intelligence. The Turing machines work as
- The CPU
can have many machine states.
- The four basic operations performed by the Turing machine include: to write a symbol at the specific memory location, to remove it from there, get the memory location of the next element and also get the previous element in the memory.
- Which operation to be performed by the processor hangs on two facts: the symbol present at the current memory location; and the scanner’s own current machine state.
Artificial Intelligence tends to make models that can think like humans. It does this by the famous words like a machine and deep learning. Deep Learning algos are based on the activity of the human brain. How the neurons work and how they are connected to each other. The models tend to replicate a human brain.
Advancements in AI
Over the years we have seen different deep learning architectures like CNNs, the RNNs, or the GANs. The different architectures have different purposes like they can be used for the textual data as well as visual data, still, we are far away from intelligent robots. Over the past few years, there have been exceptional advancements in AI. In March 2016, Alpha Go beat the nine times chess champion Lee Sadol. But one main hindrance followed by AI is that it requires a very large amount of data as the algos are data-hungry. The data collection for these models is a very big problem. In the real world with different properties, data gathering is a very challenging task.
The AI lead machines lack the property of relations among representations. AI models work on the principles of probability. They deduce the results from the Bayesian theory of probability. Consider the example of seeing dark clouds so we can mentally represent that it will rain so there is a mental representation between rain and clouds. But not only cognition there can be conventions like seeing the red light we have to stop, it is not a rule but just a convention.
According to Thomas Hobbes, considered as one of the earliest advocates of the computational theory quoted, “computation means reasoning. And computation means adding all the things which were added simultaneously. So, the reasoning is the same as adding or removing.
From the above discussion we can conclude comparing the computation power of the mind with AI still has a long way to go. Researchers are hopeful to produce intelligent robots but the main question which has been bothering the Researchers is how to produce the senses in a robot as that of the human brain.
- https://www.google.com/search? q=Computational+theory+of+mind+and+Artificial+Intelligences&source=lnms&tbm=isch&sa
You may also know: How can we use Artificial Intelligence to fight against coronavirus