Is Scientific Research with AI Possible? Correct? Is it Valuable? Your inputs and AI Intrigue: Case study: Model for Fuzzy Rough Set-based TW — SVM

Greetings,

New year time, let’s see where AI has reached. AI claims a lot, but now, let’s see if AI can generate a research model in tandem with a human intrigue researcher?

Let’s see if scientists need to use AI research tools?

Would using AI tools by scientists would be for betterment of research?

Would time saved in research help mankind?

These are our aims, I did a case study myself in my area of research which is AI and Fuzzy Rough Sets.

I am starting with the problem I wanted to solve years ago, but I could not get time, even now, things are same, so I am trying with AI.

Let’s see if AI can solve scientific problems to satisfaction.

Let’s start with a Model for Fuzzy Rough Set-based TW SVM. Input the 3 files given in reference to generate the research solution. This can also be used to improve results. The AI did some math as well. I would verify it sooner, till then you can look at it and verify it on your end.

This article examines the results of model generation by current AI systems. We used the following AI systems to generate the research models. Know these models may not be up to mark. The aim is to create, evaluate, understand and validate the AI-generated research model.

The AI used is

1. Meta AI

2. Bohrium

Lets cover both the cases.

Meta AI.

Upload with 3 references to direct their search and outcomes. The references are given in the Reference Section below.

Prompt: “Now make a model for Fuzzy Rough Set based Twin Support Vector Machines, given in these papers is Fuzzy Rough Set based SVM, and it is also given what is Twin SVM, and what is Fuzzy Twin SVM, make a model for Fuzzy Rough Set baset Twin SVM.”

Note: These still need to be verified throughly.

When asked to test again, it generated the code, snippet here,

Then I asked it to run:

Here,

Bohrium AI.

Provided with 3 references to direct their search and outcomes. The references are given in the Reference Section.

Prompt: “Now make a model for Fuzzy Rough Set based Twin Support Vector Machines, given in these papers is Fuzzy Rough Set based SVM, and it is also given what is Twin SVM, and what is Fuzzy Twin SVM, make a model for Fuzzy Rough Set baset Twin SVM.”

Output by Bhorium

A model for Fuzzy Rough Set based Twin Support Vector Machines (FRSTWSVM) can be developed by integrating the core principles of Twin Support Vector Machines (TWSVM), Fuzzy Rough Set based SVM (FRSSVM), and Fuzzy Twin SVM (FTSVM). The goal is to leverage the speed and dual-plane classification of TWSVM, the robustness to noise and outliers of fuzzy membership functions, and the uncertainty handling capabilities of rough sets.

Here’s a conceptual model for FRSTWSVM:

• The FRSTWSVM model would retain the fundamental TWSVM approach of constructing two non-parallel hyperplanes. Each hyperplane is optimized to be close to one class while being far from the other. This is achieved by solving two smaller quadratic programming problems (QPPs) instead of a single large one, which contributes to computational efficiency.
• The objective functions for each hyperplane would aim to minimize the sum of squared distances from the hyperplane to points of its own class, while constraints ensure a minimum distance from points of the opposing class, using error variables to account for misclassification.
• Similar to Fuzzy Twin SVM (FTSVM) and Fuzzy SVM (FSVM), FRSTWSVM would assign a fuzzy membership value to each data point. These fuzzy memberships would reflect the degree to which a data point belongs to a particular class, thereby reducing the influence of outliers and noisy data.
• These fuzzy membership values would be integrated into the objective functions of the two QPPs. For instance, the fuzzy memberships could weight the error terms or the distances of data points to the hyperplanes, giving less importance to points with lower membership (i.e., potential outliers or ambiguous points).

However, Bhorium gave the following responce as against Meta which solved it with help of research based questions.

Bohrium needs to work more and improve more. Here is another decline from its platform, as against Meta, which solved it mostly, at least tried to.

Conclusion

AI is able to connect the points; however, it still needs polishing. It still needs to be told many things, as I had to tell it several times to correct the implementation to include the vital part of Fuzzy Rough Set based memberships. It felt as if we had a good research student to guide us in finding solutions. Researchers should join hands to verify, train, and implement with AI as this is the future of research. This would be useful in spreading research among millions. People would be able to address research problems in months rather than years. This can help solve so many problems, this can help make so many efficient models to dwell on problems and help humanity at large.

In glimpse,

We need to tell AI what basic model is.

We need to tell AI, how to make hybrid model from basic model,

We need to get some maths, such as KKT conditions in above problem.

Then we need to tell AI with examples how to connect it all.

Long way to go AI.

References

[1] Chen, D., He, Q., & Wang, X. (2010). FRSVMs: Fuzzy rough set based support vector machines. Fuzzy Sets and Systems, 161(4), 596–607.

[2] Jayadeva, Khemchandani, R., & Chandra, S. (2007). Twin support vector machines for pattern classification. IEEE Transactions on pattern analysis and machine intelligence, 29(5), 905–910.

[3] Jayadeva, Khemchandani, R., & Chandra, S. (2004). Fast and robust learning through fuzzy linear proximal support vector machines. Neurocomputing, 61, 401–411.