DEEP Learning! What? Why? Where?

We will be discussing the three W’s to define Deep Learning, one of the trending technologies of today’s world.

Credit: Author

What is Deep Learning?

Deep learning (DL), as the name, suggests, makes the machine learn a given data in depth. Deep learning is a subset of machine learning in artificial intelligence (AI) that has networks capable of learning from the data by making a neural network, similar to the network present in our BRAIN.

Key features of Deep Learning:

• Deep learning is an AI function that mimics the workings of the human brain in processing data for use in detecting objects, recognizing speech, translating languages, and making decisions.
• Deep learning AI can learn without human supervision, drawing from data that is both unstructured and unlabeled.
• Deep learning, a form of machine learning, can be used to help detect fraud or money laundering, among other functions.

How Deep Learning Works?

Credit: Research Gate, edited by Author

The neurons that we use in deep learning aren’t actual biological neurons present in our brains. A neuron is a mathematical function that just models functioning similar to biological neurons. Arranging these neurons to form complex layers forms an Artificial Neural Network (ANN).

An artificial neural network is a computing system that is comprised of a collection of connected units called neurons that are organized into what we call layers. Working of ANN:

• A large amount of data is fed to the input layer of the ANN.
• The neural network automatically extracts important features from the data.
• ANN then processes the data according to the structure of the model made.
• The data propagates back and forth several times across ANN in each epoch(run). ANN tries to learn by updating the values in mathematical functions and improving accuracy.
• The final layer is the output layer.

How Deep Learning (DL)is different from Machine Learning(ML)?

Credit: Author

• The above image clearly states that ML performs better in less amount of data whereas DL beats ML with huge data (big data) with a huge difference in accuracy.
• Feature engineering is an exhaustive process, it requires to put the domain knowledge (based on a dataset) to know and extract important features. These features can best be identified only by the domain expert. Feature extraction is hand-coded in ML, and model accuracy mostly depends on it. Whereas in DL, the feature is extracted automatically by grouping similar data.
• Since DL mostly works on huge data, it forms a complex model and thus it requires high-end machinery including GPUs to train the model. On the other hand, ML models can be easily trained on personal computers.
• ML divides the problem into sub-problem and then solves each sub-problem step by step. DL in contrast advocates solving the problem end-to-end.
• DL involves a whole bunch of parameters making it slower in executing than ML models.

Conclusion: From the above-stated points we can easily conclude that DL is best suited for real-life complex problems whereas ML works best for simple problems. This gives us an answer to our second question Why.

Where is deep learning used?

Use cases today for deep learning include all types of big data analytics applications, especially those focused on natural language processing, language translation, medical diagnosis, stock market trading signals, network security, and image recognition.

Specific fields in which deep learning is currently being used include the following:

• Customer experience: Deep learning models are already being used for chatbots. And, as it continues to mature, deep learning is expected to be implemented in various businesses to improve customer experiences and increase customer satisfaction.
• Text generation: Machines are being taught the grammar and style of a piece of text and are then using this model to automatically create a completely new text matching the proper spelling, grammar, and style of the original text. Eg: Google Translate
• Aerospace and military: Deep learning is being used to detect objects from satellites that identify areas of interest, as well as safe or unsafe zones for troops.
• Adding colour: Color can be added to black-and-white photos and videos using deep learning models. In the past, this was an extremely time-consuming, manual process.
• Medical research: Cancer researchers have started implementing deep learning into their practice as a way to automatically detect cancer cells.
• Computer vision: Deep learning has greatly enhanced computer vision, providing computers with extreme accuracy for object detection and image classification, restoration, and segmentation.
• Robotics: Reinforcement Learning combined with Deep Learning forms a great new application. Eg: Self-driving cars

DL has a great future, it has been penetrating and taking over all business sectors.

Source: Google Trends from the past 5 years

As we can see in the above image, the trend of Deep Learning is always increasing in the past 5 years.

There is a lot more to research and improve our algorithms to use the data sufficiently. In my opinion, we are witnessing the popularity of using deep learning in many fields, and soon, it will be extended to all aspects of science, engineering, and so on.

In the next blog, we will be starting with a new story on DL algorithms. I will see ya in the next one!