Demystifying Deep Learning – Back to Basics

Deep Learning – A very young and limitless field. It is a class of machine learning where theories of the subject aren’t strongly established and views quickly change almost daily. Deep Learning (DL) employs multiple layers of non-linear training units to facilitate feature extraction and transformation. The phenomenon of artificial intelligence has induced significant transformation in contemporary enterprises due to its demonstration of a degree of accuracy in specific endeavors that is akin to human capabilities.

“I think people need to understand that deep learning is making a lot of things, behind the scenes, much better” – Sir Geoffrey Hinton

This is part 1 of 2 parts story on DeepLearning & basic terms which revolve (may evolve around as well) around it.

Why Deep Learning?

• Human Brains have deep architecture.
• Humans organize their ideas hierarchically, through the composition of simpler ideas
• Insufficiently deep architectures can be exponentially inefficient.
• Deep architectures facilitate feature and sub-feature sharing.

What is Deep Learning?

“Deep learning is undeniably mind-blowing” and “deep learning can be used with too much ease to predict the unpredictable”. In my personal opinion, “We are all so busy creating artificial intelligence by using a combination of non-bio-neural networks and natural intelligence”. The domain of Deep Learning has surpassed standard machine learning techniques and their respective algorithms, which can be classified as either supervised or unsupervised strategies. 

Deep Learning, in short, is going much beyond machine learning and its algorithms that are either supervised or unsupervised. In DL, it uses many layers of nonlinear processing units for feature extraction and transformation. It has revolutionized today’s industries by demonstrating near-human-level accuracy in certain tasks. tasks like pattern recognition, image classification, voice and text decoding, and many more.

A variety of tasks, including but not limited to design acknowledgment, pattern recognition, picture classification, and voice or text interpreting, are frequently employed in the respective field under consideration. Self-driving cars are one of the best examples and biggest achievements so far. I hope and pray that there will be no time when we need to do the reverse and use artificial intelligence to create natural intelligence in the future.

That said, it is my impression that the representation-focused perspective of artificial neural networks is presently very popular. Deep learning is a learning scheme that approaches the learning problem by learning the underlying representations; too much learning I think that’s why it’s also called representational learning.

Because this is a young and limitless field as of date, theories and conclusions aren’t strongly established, and views quickly change almost daily. Some people also call it cutting-edge emerging technology and a breakthrough in the AI domain. This depicts what machines can do, and it’s just a start.

Deep Learning (DL) has revolutionized, image processing & classification and also speech recognition with high accuracy. Business leaders and developers’ communities absolutely need to understand what it is, what it can do, and how it works.

What is Neuron

On a high level, we have three types of neurons in the human body: sensory neurons, motor neurons, and relay neurons. Please note that we are not at all claiming or making any efforts to demonstrate any knowledge about biological science here; we neither have any idea nor do we claim so.

The information here is taken from the open internet to just give a small understanding of the artificial neuron. The typical structure of the human neuron is below.

Artificial neural networks were originally designed to approximate biological neurons. They are mathematical constructs. Each “neuron” is a relatively simple element—for example, summing its inputs and applying a threshold to the result to determine the output of that “neuron”.

Artificial Neural Networks (ANN)

As per Dr. Robert Hecht-Nielsen, the inventor of one of the first neurocomputers, a neural network, or artificial neural network (ANN), is “a computing system made up of several simple, highly interconnected processing elements, which process information by their dynamic state response to external inputs.”

Artificial Neural Networks (ANN) are inspired by the human brain. The structure of this neuron is now used for machine learning and artificial intelligence. With these artificial neuron networks, various problems, especially where layering is needed for refinement and getting more granular details are needed, are now being solved.

The neural networks are now part of research subjects like neuroinformatics (a research field concerned with the organization of neuroscience data by the application of computational models and analytical tools), and they require extensive training before being implemented in a real-time problem-solving environment. They can solve problems.

Working Method of Neural Network

Nural Network for Real can work as a visual guide for data evolution strategies and would also be a highly scalable alternative to deep reinforcement learning. Information flow in a neural network happens in two ways. The below-mentioned methods are common designs for a feedforward network.

• At the time of its learning or “being trained,”
• At the time of operating normally or “after being trained,”

Patterns of information are fed into the network via the input units, which trigger the layers of hidden units, and these, in turn, arrive at the output units. Though this subject is large and complex because there are many different types of artificial neural network models, DL’s most common model is backpropagation. This has become the foundation for most of the others, which are fully connected.

Please note : In the DL chain, not all units get triggered all the time. Each unit receives inputs from the units to its left, and the inputs are multiplied by the weights of the connections they travel along.

Characteristics of Deep Learning that make it different from Machine Learning

Though machine learning has laid the foundation for deep learning to grow and evolve, At the same time, deep learning has taken key features from the machine learning model. An interesting fact and truth here is that it takes it a step further by constantly teaching itself new abilities and adjusting existing ones.

• Deep learning is a cultured technology with the highest degree of accuracy for self-learning capabilities. This makes the results from this often accurate and faster to process.
• It learns high-level, non-linear features necessary for accurate classification.
• Deep Neural Networks are the first family of algorithms within machine learning that do not require manual feature engineering; rather, they learn on their own by processing and learning the high-level features from raw data.
• Deep neural networks are fed raw data and learn to identify the object on which they are trained, or “being trained”.

This unique type of algorithm (deep learning) has far surpassed any previous benchmarks for the classification of images, text, and voice.

In the next post, we will talk about below important aspect of this topic

• Deep learning Computational Models i.e Feed forward propagation and backward propagation
• Deep learning Algorithms – High-level of view
• Implementation of Deep Learning Models
• Notable Use Cases & Applications
• Deep learning limitations

Just for Fun 🙂

Let’s taste the real AI and see how it will amaze us soon. The below joke has been floating around social media. It paints an excellent picture and reality of the future (for real) impact of deep learning technology in our lives, businesses, and personal work. Note that this joke is taken from one of the social media platforms.

• Hello!  Somi’s pizza?
• No, sir, it is Naomi’s pizza.
• So is it the wrong number?
• No, sir, Naomi bought it.
• OK. Take my order, please.
• Well, sir, do you want the usual?
• The usual? You know me?
• According to your caller ID, in the last 12 times, you ordered pizza with cheese, sausage, and a thick crust.
• OK! This is it.
• May I suggest to you this time ricotta and arugula with dry tomatoes?
• What? I hate vegetables.
• But sir, your cholesterol is not good.
• How do you know? Through the subscribers guide, we have the results of your blood tests for the last seven years.
• Okay, but I do not want this pizza, please. I already take medicine.
• But sir, you have not taken medicine regularly; four months ago, you only purchased a box with 30 tablets at Tony’s drug store.
• I bought more from another drugstore.
• It is not showing on your credit card.
• I paid in cash.
• But sir, you did not withdraw that much cash, according to your bank statement.
• I have other sources of cash.
• This is not showing as per your last tax form unless you bought them from an undeclared income source.
• WHAT THE HELL?
• Enough! I am sick of Google, Facebook, Twitter, and Whatsapp. I am going to an island without the internet, where there is no cell phone and no one to spy on me.
• I understand, sir, but you need to renew your passport as it expired five weeks ago!

============================ About the Author =======================

Read about Author at : About Me

Thank you all, for spending your time reading this post. Please share your opinion / comments / critics / agreements or disagreement. Remark for more details about posts, subjects and relevance please read the disclaimer.

FacebookPage    ContactMe      Twitter         ====================================================================