Computers are useless, all they can do is answer questions”― Picasso
Maybe this is true during the old times but a conventional PlayStation today is more compelling than the military supercomputer in 1996.
So it is evident that people want to know what Artificial Intelligence is and how it will help organisations evolve. Though there are tons and tons of articles, nothing seems to provide a modest definition. But in the present state AI will fall under the Soviet Harvard delusion (unscientific overestimation of the reach of its capability today)
We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run
But at the same time as we are in the midst of a technological upheaval that will transform the way business is proceeded. I sincerely consider that AI will take a pre-eminent role in this disruption. AI has advanced from the hypothetical into Real world business solutions. Let's start with a fundamental question
What is AI?
Artificial Intelligence is nothing but the intelligence exhibited by Machines. Any device that perceives its environment and takes actions that maximise its chance to succeed at a similar goal.
AI is like teenage sex: everyone talks about it, nobody knows how to do it, everyone thinks everyone else is doing it & so claims to do it
Artificial Intelligence Vs Machine Learning Vs Deep Learning
Machine Learning gives Computers the ability to learn without being explicitly programmed― Arthur Samuel 1959
Different types of Machine Learning
Applications in which the training data comprises examples of the input vectors along with their corresponding target vectors are known as supervised learning problems.
- Classification: Assign a label
- Regression: Predict a continuous numerical value
- Mapping Expenditure to Profits
- Capability of the Client to return the Loan amount
- Analyze transactions to find the potential cost cuts
In other pattern recognition problems, the training data consists of a set of input vectors x without any corresponding target values. The goal in such unsupervised learning problems may be to discover groups of similar examples within the data, where it is called clustering
- Dimensionality reduction
- Cluster the search results into grouping to give a small footprint
- Simplify complex dataset by reducing number of dimensions
Instead of feeding with the answer key, the system should learn from the experience by collecting training examples from the real scenarios and test itself by trial and error methods with the goal of maximizing the long-term reward
- Reward Maximization
Deep learning is the modern refinement of “machine learning” and uses a structure of “deep” neural network of more than three layers. These complex neural networks are more able to categorize intricate links between massive, high-dimensional data sets.
As data gets passed through one layer of nodes to another, each layer trains on a particular set of features based on the previous layer’s output. The deeper into the neural network it goes, the more complex data is processed. This process is commonly referred to as “feature hierarchy” and is best exemplified in tools like image recognition, etc.
Each input will cross multiple hidden functional layers to derive the output. And the number of hidden layers depends on the complexity of the task.
Self Driving Cars reply on the deep learning for visual tasks like understanding road signs, detecting lanes and recognizing obstacles
Foundations of AI
Achievements of AI
- Google's Alpha GO defeated one of the best human players at GO. It is pretty extraordinary given the full nuance and complexity of this ancient Chinese war strategy game with its 10 to the power of 170 possible positions (there are only 10 to the power of 80 atoms in the universe)
- Mars rover used AI to autonomously select inspection worthy soil and rock samples with high accuracy
- BenevolentAI is using artificial intelligence to mine and analyse biomedical information, from clinical trials data to academic papers
- Google DeepMind AI examined the problem of how it uses energy in Google’s data centres ( they use 70 billion kilowatt hours of electricity per year) and was able to slash it by 40%
- In 2014, IBM used its Watson cognitive computing system to help Africa solve business and social challenges
Humans are likely to retain some advantages over machines for the foreseeable future because of their ability to know more than what they can tell― Deduced from Polanyi’s Paradox
Concerns of AI
- With low interpretability, it is difficult to figure out how the systems reached their decisions
- Neural network systems deal with statistical truths rather than literal truths. This will question the certainty that the system will work in all cases
Major Companies in the AI Race
“By far the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.”― Eliezer Yudkowsky
Democratization of AI is under way, But human brain is still more advanced and complex than any artificial neural network. This doesn’t mean that AI won’t continue to move towards that direction over time. With every breakthrough, more opportunities will arise that go beyond our wildest imaginations.
On the other side, the newly enforced data regulations like GDPR (AI should explain how it came to those conclusions in every case) will pose a potential threat to the advancement of AI
Thank you for being here and If you have enjoyed the post leave your comments or share this article so that others can see it.