Artificial Intelligence — Inception
A series:
I am inspired by the term AI, trying to figure out past, progress, and propaganda… trying to jot down my daily learnings, thoughts over here as I progress and explore the world of AI… Stay tuned!!
Little Background:
It’s been more than 7 decades since the term Artificial Intelligence has been with us. Be it the decoding of German-ENIGMA by British or the belief of McCarthy that one-day computers can play chess without human intervention. So why it took so long before AI. could have been developed or utilized in the common world? …Well, somehow the progression of pragmatic implementation of AI could not follow the theories given between 1950–1970 due to technical failures and lack of funding. It’s utter truth that investors would like to pitch for the Technology which gives them ROI fast than waiting for years.
The sudden growth in programming languages and interest in Imperative/object-oriented Programing in the early 1980s brought significant improvement in AI. Be it google’s BERT (Bidirectional Encoder Representations from Transformers)or Youtube’s algorithms to find the next video that you may like, they are great examples of how far AI has come.
Fact about AI’s growth: This is an open secret that in order to gain knowledge one has to learn and to learn one needs data to turn that into comprehensive form — information and then derive facts out of that and to gather right facts one has to explore, then try various sources, apply different permutations & combinations, and during this experiment, one gains experience which is then identified as the relevant findings from this lengthy exercise and churn into knowledge. We read and gain knowledge often which is perceived by someone else and based on their experiences.
In the same manner, AI grows…
Problem Statement: Humans create & train AI Today.
We have been developing AI keeping our own intelligence in the back of our heads, for example; we know that humans are pre-programmed under Brain Stem, for survival, breathing, motor function, to react or defend ourselves in a dangerous situation. In the same way, we learn, adopt & develop specific thinking, beliefs over time by our experiences. Do you think that can such human brains can overcome these predominant thoughts developed over the years through their own learnings — Ever?
If you have children or can recall your childhood, the most interesting phase starts between 3 to 5 years as kids of such age are so eager to learn, explore, and ask questions about almost everything that is normal for adults or about their surroundings that they come across and irrespective of which language they speak their question starts and continues with a series of WHY?
This exact Why has been/is still missing in many AI programs, today the AI has become more intelligent and made great progress in Machine learning or Deep Learning, while they take input of terabytes of data, process it, generate new models or algorithms but, there is a barrier in the development of AI because not every program is designed to ask WHY and explore further…
The point is that we need to build & grant freedom to an AI program that is able to develop its own brain which can perceive, predict, presume, produce information Autonomously. Many researchers have found that humans are trained to presume answers to common questions on behalf of others based on their prior experiences and in most of the common cases these answers turn out to be accurate. This exact same thinking or experience, if applied in AI programs engenders a kind of ensnare which costs to the freedom and autonomy of AI.
Fear factor Hypothetically: Well, some learned people or politicians would definitely question or create barriers through policy governance and laws around this freedom factor for AI, because certain vital aspects like the value of human life must be encoded into all AI else the result could be preposterous. Many admire movies like Matrix or Ex Machina but if that becomes our reality one day then we would surely pay a huge price.
What AI is / what its not:
What AI is everyone talks today but I came across meeting people who could not explain what its not or sometimes misunderstood complex programs written to interact with Humans as AI. The applications written and pre-trained and fed with specific questions-answers and programmed to run quick searches to revert with the results that we have asked for. But, Alexa or SIRI from Amazon and Apple are put under AI category. Why???
They both depend upon natural language generation and processing, then the machine learning develops artificial intelligence around it, in order to effectively operate while interacting with humans and simultaneously learning how to perform better over the time.
Long ago, when OCR (Optical Character Recognition) was invented it was also considered as AI, because the programs were able to distinguish between similar-looking characters or numbers e.g Zero “0” vs Letter “O” or different shapes, handwritten texts, etc. Now, OCR is out of A.I list because the definition of A.I has evolved.
To be continued…
