Introduction to AI and Data science

       Greetings, This blog contains the required information to get started with Data Science and Artificial Intelligence. This blog focuses on 

1. What is Artificial Intelligence
2. What is Data Science 
3. Domains of Data Science 
4. The general logic of Artificial Intelligence
5. Skills required to learn Data Science and Artificial Intelligence

      "Artificial Intelligence is a kind of human-made intelligence that is embedded in the machine to perform a particular task". The end goal is to make an AI application. To create an AI application, the important factor is data. We feed data to the machine then that machine tries to create knowledge with respect to the available data. It is also important to process the data because if data is correctly processed AI machines give us highly accurate results. To accomplish this task, Data scientists are required. Data scientists are required for overall AI-based systems.  Data scientists are responsible to preprocess the data as well as create an AI model for the system. AI developers and AI engineers are only focused on creating AI models. Data scientists are also responsible to extract meaningful information from data.

        The main formula to remember what is data science is 

         Data science = Data preprocessing. + Artificial Intelligence

Domains in Data Science

 Fig.1 Domains in Data Science

        

1. Data Preprocessing : 

   

    When Data scientists are supposed to create an Artificially intelligent system at that time, the data which they get from the client-side is raw data or dirty data. Raw or dirty data means the dataset contains some missing values or there can be some text in between or any possible data like images or audio. AI system cannot be developed with that data. Hence dataset needs to be processed in a certain manner. Raw data needs to be converted in a proper format so that machine can understand. The accuracy totally depends on preprocessing of data done by data scientists.

2. Artificial Intelligence :

    

    Artificial Intelligence is the main goal. The application or system, which will be developed by a data scientist or AI developer has the capability to think like a human in a certain way. 

3. Machine Learning :

    

    Machine Learning is one of the sub-domain of Artificial Intelligence. Machine Learning is only related to predictive models. Predictive models are the models which will give predictions as an output. In AI systems, AI is known as the AI model.

4. Deep Learning : 

       Deep Learning is all about mimicking (copy) human thinking. The way humans think similarly our machines will think and take necessary decisions.  

General Logic of Artificial Intelligence

 Fig 2. School teaching VS AI training

    Let's take a use case to understand how data scientists create AI models. The process of giving intelligence is similar to the way we got intelligence. We got our basic understanding in school. In schools, there is a syllabus that is assigned to all students. The teacher teaches each and every concept to all students. After teaching, all students get some homework. Students solve the homework, they can get some correct answers and some wrong answers. The next day, the teacher clears the queries, and if possible teaches the same concepts again. Slowly, The amount of mistakes students does in their homework decreases. This process goes on until the syllabus is done. After the syllabus was done, the teacher conducts a test. If students pass the test they get some percentage and the students are promoted to the next batch. 

    This is the similar process how data scientist makes the machine to learn. Data scientists are the observer who observes the AI model. Teachers are kernels (compiler or interpreter of machines). Instead of the syllabus, AI models get datasets to learn. AI model is supposed to read the data multiple times. Dataset is divided into major and minor data i.e training dataset and validation dataset. Every time machine fetches training datasets to learn and to find the patterns. Instead of homework, it tests itself with the validation dataset. After testing it takes some measures to avoid errors i.e mistakes. Slowly, machines are also able to get accurate results. Finally, After the AI model is trained, Data scientists test that model with some unknown data also called testing dataset. If AI models give proper results and accuracy (percentage) then it is sent for production else that AI model is discarded. I hope this example gives a basic idea about ML workflow.

Skill Required for Artificial Intelligence

1. Anyone Programming language (Python, C++, Java, etc): Python preferable.
2. Concepts of Mathematics.
3. AI Algorithms
4. Data preprocessing techniques
5. A little bit of information about Data storage like Databases and Big data
6. A little bit of information about the deployment and Web development.

    All these skills are available in this blog channel If you know all these things. You can say that you are a Data Scientist. 

    

    Artificial Intelligence is our main goal, From the next blogs, We will be focusing on sub-domains of AI i.e machine learning and deep learning. 

Summary    

1. Artificial Intelligence is a kind of human-made intelligence that is embedded in the machine to perform a particular task
2. Data science is used to extract meaningful information from data.
3. Data Science = Data Preprocessing + Artificial Intelligence
4. Data preprocessing is for better accuracy and converting data so that machines can understand.
5. Artificial Intelligence is our end goal. 
6. Machine Learning is used for creating predictive models
7. Deep Learning is nothing but mimic (copy) of the human brain

    

                                                                                                                                         

- Santosh Saxena