By Admin 
Monsoon in Kerala reached 8 days before its forecast, while Delhi did not rain despite the Yellow alert being issued. At the same time, cloudburst in Kullu and Dharamshala of Himachal Pradesh caused havoc. When two people lost their lives in this accident, 9 people are missing and more than 2000 tourists are stranded there. In such a situation, if the accurate forecast of the weather was already found, there would probably not have been so much damage. & Nbsp;
Now it will also be possible, because scientists from IIT Delhi have prepared a new Artificial Intelligence (AI) based model, which can tell the arrival of the monsoon 18 days in advance. This technology is not only faster than the current weather forecast models, but also more accurate. It has used transformer neural network technology, which is also used in modern systems like Chatgpt. In this research, Professor Dr. Sandeep Sukumaran, Professor Hariprasad Kodamana and two PhD researchers KM Anirudh and Pankaj have worked together.
AI model is better?
Professor Sandeep Sukumaran said that this transformer model is extremely advanced and modern. During research, he found that the error growth in this model is much lower than the traditional models. It proved better when this model was compared to the weather forecast models of America, Britain and other countries. He said that to run the traditional model, big supercomputer and heavy resources are required, while the new AI model can also work with ordinary GPU, as used in gaming laptops. This will not only save time, but will also save a lot of government resources. This project has been funded by the Ministry of Earth Sciences, Government of India and the team is working closely with the Indian Meteorological Department (IMD).
5 -year data made from data
PhD scholar Anirudh said that he has compared the AI based transformer model by studying the weather data of the last 5 years to traditional models. Comparing with the existing models of the US and Britain, he explained through the graph that the AI model is not only more accurate, but it saves both time and money. Traditional models take a long time to estimate with the help of supercomputer, while the new AI model can give results in a few seconds.
40 also conducted on cyclone
PhD scholar Pankaj said that during his research, he investigated 40 cyclones during four years. Traditional models estimate the direction and position of the cyclone within a radius of about 500 km, but the AI model was successful in reducing this distance to a great extent. According to Pankaj, traditional models or machine learning models are able to give estimates up to 4 days ago and the track against it is not accurate. At the same time, accurate prediction of the track has become possible with the help of AI model. The new AI based model of IIT Delhi is being considered a revolutionary step towards weather forecast in the country. With this, people will get accurate information about the weather on time. At the same time, the resources of the government will also be saved. In future such models can prove to be very important in warning before disasters.