Note: This does not mean thinking can be automated. Research thinking can’t be automated.

So, how to do research with current AI,

Lets take chatGPT, you can take Anthropic or likes, I prefer chatGPT always, I remember how much of a research time goes in coding and coding and research again. This too and fro can be automated. Let’s see how, however its correctness is still a question.

But what remains is thinking..the main thing..It may reduce research time, but thinking is a gift of mind which cant be replaced..

So, steps to implement a research idea/problem in Maths and Physics:

1. Introduce the chatGPT the research topic, for example, neo twin support vector machine.
2. Then, tell it to write a code for it in python on itself or with SageMaths like toolkits. See if it can generate Matlab to integrate with your research language.
3. Download the code written in Python.
4. Run the code, for inputs.
5. Verify the outputs manually.
6. Verify the code generated by ChatGPT or similar tools. And give feedback to chatGPT.
7. This would be perfect in sometime when CODE AUTOMATION would be done.

Advantageous:

1. What is considered as a 5 year PhD can be completed in 2–3 years then. A lot of money and energies would be saved
2. Research can reach the hands of software developers too.
3. Common people who can’t qualify for top institutions can follow research done in top institutions.

Example: A small research problem which a researcher may like to accomplish before creating say a new model for Twin Support Vector Machines. Here, are some advanced topics in research.

Using fitness function as classification of Twin Support Vector machines with Genetic Algorithms to optimise a quadratic function. Use 5 fold verification.

Write prompt above in your chatGPT to get the automation.

Another one, here,

“Neo-Twin Support Vector Machines for Pattern Classification Code Implementation”

Here is what chatGPT did the output as:

Another latest research topic is “Artificial Electric Field Algorithm”

Here is what chatGPT generated for it:

Lets verify what chatGPT generated.

And let’s see if the steps given at the start of the article can all be joined together in one prompt.

Let’s verify everyone. This can increase time efficiency, but we need to verify manually.Lets verify everyone, it can increase time efficiency, but we need to verify manually,

Software developers and data scientists are needed for this work.

To be continued…

Thanks for reading..