Artificial Intelligence | Machine Learning | Deep Learning

What is the actual concept behind Artificial Intelligence, Machine Learning and Deep Learning

? Artificial Intelligence (AI)

AI is a technique which enables machines to mimic human behavior. AI is intelligent machines that think and act like human beings.

E.g. Apple Siri or Tesla/Google Self learning Car based on NLP (Natural Language Processing) and DL

AI types Reactive Machines, Limited memory, self-awareness, theory of mind.

ML is Systems learn things without being programmed to do so.

DL is Machine think like human brains using Artificial Neural Networks.

Data dependency as It depends on the data as data increases the deep learning understands when a smaller amount of data machine learning works well.

Hardware dependency as deep learning algo works on high end machine as machine learning algo works with low end machine within GPUs

Feature engineering as putting domain knowledge to reduce the complexity of the data and make patterns more visible to learn algo.

Execution time as DL takes a long time to train because more parameters involved make it longer than usual. On reverse basis while test the data DL take less time to learn data where Ml take more as data increase

? Machine Learning (ML)

ML is a subset of AI techniques which use statistical methods to enable machines to improve with experience. ML uses the algo to parse data, learn from that data, and make the informed decision based on what it has learned. ML the system is able to make predictions or take decisions based on past data.

ML types

·       Supervised learning. Systems are able to predict future outcomes based on past data. Requires both input and output to be given to model for it to be trained

·       Unsupervised learning; systems are able to identify hidden patterns from input data provided. By making data more readable and organised, the pattern, similarities become more evident

·       Reinforcement learning; systems are given no training. It learns on the basis of reward/punishment it received for last performed action.

ML features;

Better decision and prediction

Quick and accurate outcomes

More powerful processing capability

Analysing complex big data

Managing large amounts of data

Inexpensive

? Deep Learning (DL)

DL is a particular kind of ML that is inspired by the functionality of our brain cells called neurons which led to the concept of Artificial Neural Network. DL structures algo in layers to create an artificial “Neural Networks” that can learn and make intelligent decisions on its own. DL is a subfield of ML. while both fall under the board category of AI, DL is usually what is behind the most human-like AI.

DL takes large unlabelled data training the model using Artificial Neural Networks. Once it is trained then new data is given to the testing model and it will provide the required output. The information is converted into a bunch of different nodes and it goes into different layers and the outcome will come.

DL the systems think and learn like humans using artificial neural networks. Deep learning is machine learning and is the next evolution of machine learning.

DL features;

Performance improves with more data

Better scalability

Problem solved in an end to end method

Best features are selected by system

Is subset of ML

Lesser testing time

Comparison ML Vs DL

ML needs only a small amount of training data as DL needs a large amount of training data.

ML works well on low end systems as DL needs high end systems to work.

ML most features need to be identified in advance and manually coded as DL the machines learns the features from the data it is provided.

ML the problem is divided into parts and solved individually and then combines as DL the problem is solved in end to end manner.

ML testing takes longer as DL testing takes less time.

 - Apurva Bhandari