Demystifying Key Buzzwords – Artificial intelligence and its emerging bundled technologies like machine learning, artificial neural networks, deep learning and many many more are simple and complex terms at the same time. These buzz words are agents of future analytics in one or the other way. Involvement of such emerging technologies are now part of our daily life and we use them knowingly or unknowingly. All of these buzzwords will work well not only in analytics business but also going to play a key and extremely important role in our daily and business life.
In this post, we will attempt to demystify some use cases of machine learning and deep learning at a high level to make them clear. To name a few industries where these technologies are making difference already are healthcare, finance, payments, e-commerce, manufacturing, engineering services etc.
Let’s Understand a Bit
Let’s make it simple to understand “how all-encompassing terms are actually correlated and speak to each other”. Artificial Intelligence, Machine Learning and BlockChain etc. are all the golden words of today’s time. Almost every technology (now even non-technology) company on this planet is claiming the share of extra revenue by putting these buzzwords on their product display.
It’s easy to get lost and not see the difference between hype and reality. Reality and real benefits are getting lost here in the race of these buzzwords swirling around everything. In AILabPage’s view, all these terms are representative of future analytics. Sometimes it is good to un-develop something existing to uncover the hidden gems underneath. Maybe its like un-develop to Innovate?
In the past, Alan Turing published “Turing Test” that speculates the possibility of creating machines that think. Where in order to pass the test, a computer must be able to carry on a conversation that was indistinctive from a conversation with a human being
Confusing Jargons – AI, ML and Many More
Defining confusing jargons like AI, ML, Data Science, DL and ANN with no scientific reasoning and meaning can create huge understanding gaps. Each one of these buzzwords has its own meaning and use cases. One needs to be careful about when to use what for what reasons.
AI definition also includes things like understanding language, planning, recognising objects & sounds learning and problem-solving.
It all about time and the right learning algorithms makes all the difference. Below examples clarify little more about the basic concept of artificial intelligence.
- A simple mathematical calculation or product of two given numbers; any computer can beat any human in terms of speed and accuracy.
- On the other hand to identify whether an image has a dog or a cat most of the humans (even a four-year-old kid) will easily outperform computers.
In scenarios like example 2 as above where artificial intelligence has to play a critical role to show its power. An animal picture is a dog or cat etc. AI doesn’t need to demystify anything more as its metaheuristic approach has almost overshadowed. Near natural intelligence technologies and computational power of cognitive systems techniques are real now.
Algorithm – The Set Of Instructions!
In an Algorithm set of rules gets followed to solve problems so in short, an algorithm is a set of rules or instructions. In machine learning; algorithms are key elements which take on the data and process all the rules to get some responses. This processing makes the algorithm complex or easy. One thing is clear here more the data stronger algorithm gets over a period of time.
Algorithms suppose to work much faster, accurate and self-resilient to demonstrate its capabilities; much beyond any human. No algorithm can be considered good or bad but surely can be data greedy or resource hungry. Algorithms need to be trained to learn how to classify and process information.
To simulate the mapping of inputs to outputs as it happens in a human brain which makes very difficult tasks for computers like image recognition, sarcasm detection, voice recognition, etc.
The efficiency and accuracy of the algorithm are dependent on how much data fed to it to train it. Using an algorithm to calculate something does not automatically mean machine learning or AI was being used.
Human intelligence exhibited by machines. AI as a branch of computer science deals with human mind simulation for machines but interesting natural intelligence plays a key role for same. In today’s time, AI is increasingly used as a wide term. It describes machines that can mimic human functions such as learning and problem-solving. In earlier times it was believed that human intelligence can be precisely described, and machines can simulate it with AI. Before the machine starts attempting simulation, it needs to do learning with lots of data.
We needed AI for real life not just PhD or scholar books material. In simple words, AI involves machines that behave and think like humans i.e Algorithmic Thinking in general. AI-powered computers have started simulating the human brain’s sensation, action, interaction, perception and cognition abilities.
The real fact is artificial intelligence is about the people, not the machines. Technology and non-technology companies are now investing and bringing out the real and materialistic values of Artificial Intelligence to the real world.
Simply an approach to achieve artificial intelligence. Machine Learning as a subset of AI. It was carrying neural networks as an important ingredient for some time. Only recently it has become the focus of AI and deep learning. Sadly it becoming more accessible to developers as their tool. What we need is simply an MLaaS (Machine Learning as a Service) for everyone.
Artificial intelligence and machine learning are used interchangeably often but they are not the same. Machine learning is one of the most active areas and a way to achieve AI. Why ML is so good today; for this, there are a couple of reasons like below but not limited to though.
- The explosion of big data
- Hunger for new business and revenue streams in this business shrinking times
- Advancements in machine learning algorithms
- Development of extremely powerful machine with high capacity & faster computing ability
- Storage capacity
Today’s machines are learning and performing tasks; that was only be done by humans in the past like making a better judgement, decisions, playing games etc. This is possible because machines can now analyse and read through patterns and remember learnings for future use. Today the major problem is to find resources that are skilled enough to demonstrate & differentiate their learning from university & PhD books in real business rather than just arguing on social media with others.
Machine learning should be treated as a culture in an organisation where business teams, managers and executives should have some basic knowledge of this technology. In order to achieve this as a culture, there have to be continuous programs and road shows for them. There are many courses which are designed for students, employees with little or no experience, managers, professionals and executive to give them a better understanding of how to harness this magnificent technology in their business.
MLaaS is needed for data scientists work, for architects, and data engineers who have domain expertise. It is important for everyone to have a better understanding of the possibilities of Machine Learning. What is all the fuss about machine learning anyways! Can machines be creative? Can machines empathise?
Now AI has started delivering values already. Using the contemporary view of computing exemplified by recent models and results from non-uniform complexity theory has proven the fact.
What machines can do and how creative it can be? I guess that we will see in another upcoming post “Machine Learning Evolution followed by Machine Learning Transformation”. Machine Learning is (mostly) a mathematics-specific AI technique for classification, regression and clustering.
What Machine Learning Offers !!
AILabPage defines Machine Learning as “A focal point where business, data and experience meets emerging technology and decides to work together“. Machine learning is responsible for assessing the impact of data. In machine learning algorithms are used for gaining knowledge from data sets. It completely focuses on algorithms.
Machine learning is where the traditional statistical modelling of data meets the algorithmic and computational field of data science. It focuses primarily on the development of several computer programs that can transform if and when exposed to newer sets of data.
Machine learning and data mining follow the relatively same process. Algorithms are built through which input is received and after statistical analysis output value is predicted. There are three general classes of machine learning —
- Supervised Machine Learning – Map inputs to output
- Unsupervised Machine Learning – Magnify hidden patterns and trends
- Reinforcement Learning – Reward for learning
Every machine learning algorithm has three components:
Artificial Neural Networks
Mimic human brain cells — Inspired from the biological neuronal structure. The effect is to raise or lower the electrical potential inside the body of the receiving cell. If this graded potential reaches a threshold, the neuron fires. It is this characteristic that the artificial neuron model attempt to reproduce.
In a biological neuron, the transmission of a signal from one neuron to another through synapses is a complex chemical process. In this specific transmitter, substances are released from the sending side of the junction. The biological neuron model is widely used in artificial neural networks with some minor modifications on it.
The artificial neural network we train for the prediction of image and stock data has an arbitrary number of hidden layers. Also has an arbitrary number of hidden nodes in each layer, both of which the user decides during run-time.
Neural networks find great application in data mining used in sectors. For example economics, forensics, etc and for pattern recognition. “ANN – Artificial Neural Systems, or Neural Networks, are physical cellular systems which can acquire, store, and utilize experiential knowledge” – Zurada (1992).
Neural Network Architecture
There are some specialized versions of neural networks also available. Such as convolutional neural networks and recurrent neural networks. These addresses special problem domains. Like image processing and text/speech processing that is based on methodologies like deep neural nets.
Neural Network can be of 1 hidden layer to 3 at max for all practical purposes. The example below for 3 hidden layers.
- Input layer – The activity of the input units represents the raw information that can feed into the network.
- Hidden layer – To determine the activity of each hidden unit. The activities of the input units and the weights on the connections between the input and the hidden units. There can be several hidden layers.
- Output layer – The behaviour of the output units depends on the activity of the hidden units and the weights between the hidden and output units.
In practice, Deep Learning methods, specifically Recurrent Neural Networks (RNN) models are used for complex predictive analytics. Like share price forecasting and it consists of several stages.
Deep Learning – Mandate for Humans, Not Just Machines
A technique for implementing extremely powerful and much better machine learning. Deep Learning’s capabilities and limits make it so powerful that it can stand its own separate domain. Some time deep learning appears as a supernatural power of machines.
The objective of the techniques is to achieve a goal or an artificial intelligence power that teaches computers to tasks and the ability to understand anything. This process is as similar as it comes naturally to humans i.e. learn by examples.
“Deep Learning is an algorithm which has no theoretical limitations of what it can learn; the more data you give and the more computational time you provide, the better it is” – Sir Geoffrey Hinton
Key Driver of Deep Learning
Deep learning’s main driver is artificial neural networks system or neural networks. Deep learning is based on multiple levels of features or representation in each layer with the layers forming a hierarchy of low-level to high-level features.
The analogy to deep learning is that the rocket engine is the deep learning models and the fuel is the huge amounts of data we can feed to these algorithms. — Sir Andrew Ng
Decision tree learning, inductive logic programming, clustering, reinforcement learning, and Bayesian networks, among others, are key components here. A key advantage of deep learning networks is that they often continue to improve as the size of your data increases.
Deep learning Computational Models
The human brain is a deep and complex recurrent neural network. Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. In very simple words and not to confuse anything/anyone here, we can define both models as below.
- Feedforward propagation – Type of Neural Network architecture where the connections are “fed forward” only i.e. input to hidden to output The values are “fed forward”.
- Backpropagation (supervised learning algorithm) is a training algorithm with 2 steps:
- Feedforward the values
- Calculate the error and propagate it back to the layer before.
In short, forward-propagation is part of the backpropagation algorithm but comes before back-propagating.
Data science succinctly communicate the value of analysis & benefits in real life to stakeholders. Data science is the smartest BI or most intelligent BI i.e IBI to do laborious task of ‘solving intelligence’ with a range of tools and speciality areas. Fundamentally solving intelligence means the artificial seeking and production of knowledge that answers questions whether they were asked or not, known or unknown.
Successful data science requires more than just programming and coding. However, in data mining algorithms are only combined that too as the part of a process and entirely focuses on.
Artificial Learning vs. Machine Learning vs. Deep Learning
Artificial Intelligence or so-called artificial learning and machine learning are often used interchangeably, especially in this era of big data. By now we are sure you can make out clearly these terms aren’t the same, and it is important to understand how these can be applied.
Machine learning is a subset of AI and focuses on the ability of machines to receive a set of data and learn for themselves, changing algorithms as they learn more about the information they are processing.
Mixing Up All
Why DeepLearning – If supremacy is the basis for popularity then surely Deep Learning is almost there (At least for supervised learning tasks). Deep learning attains the highest rank in terms of accuracy when trained with a huge amount of data.
Data Science does not necessarily involve big data, but the fact that data is scaling up makes big data an important aspect of data science. On the other hand, i.e. thinking with machines or intelligence augmentation is serious evolutionary epistemology, and semiotic. It has extreme potentials to take business intelligence to competitive intelligence that can infer competitive measures using augmented site-centric data.
One of the best strength of DeepLearning does not require predefined features to find peculiar patterns that humans will always struggle or probably would never be able to define beforehand. Robots or AI enabled AI taking our job is still far not in the next 50 years or so.
We have excellent young professionals working in today’s time with the dream of changing the world. Their dreams go on to bring good values to society, business and people they work with.
During my time (around 19 years back) we had no such option to learn or to choose popular professions. It was mainly RnD for anything unusual. We were forced to use regression techniques if anyone wishing to be a data scientist or simply an engineer in computer software.
Robots can fail; as reason is very simple and known which is machine learning models are not sufficiently accurate or can’t be accurate without lots of data and lots of training. Artificial Intelligence with cloud computing adds advancements using new use cases to improvise the systems developed so far.
As investment in artificial intelligence is growing fast. Tech giants like Google, Microsoft, Apple and Baidu are known for their dominance in digital technologies globally are spending couple of billion united state dollars on AI. 90% of this is going to RnD & deployment, and 10% on AI acquisitions.
Are we ready to relinquish control to autonomous cars, software bots, or trust AI-based recommendation engines?
Points to Note:
All credits if any remains on the original contributor only. We have covered Artificial Intelligence, Machine Learning, Neural Network and Deep Learning to give some high-level understanding and difference. In the next upcoming post will talk about Reinforcement machine learning.
Feedback & Further Question
Do you have any questions about Deep Learning or Machine Learning? Leave a comment or ask your question via email. Will try my best to answer it.
Books + Other readings Referred
- Open Internet – Research Papers and ebooks
- Personal hand on work on data & experience of @AILabPage members
- Book “Artificial Intelligence: A Modern Approach”
Conclusion – Artificial intelligence is a broad and active area of research, but it’s no longer the sole province of academics; increasingly, companies are incorporating AI into their products. AI is controlled by humans and I wish in long term it should remain the same. AI should never get to a level where it gets into driver seat to control humans. Baidu’s speech-to-text services are outperforming humans in similar tasks. Amazon is also applying deep learning for best-in-class product recommendations.
======================= About the Author ==========================
Read about Author at : About Me
Thank you all, for spending your time reading this post. Please share your feedback / comments / critics / agreements or disagreement. Remark for more details about posts, subjects and relevance please read the disclaimer.