Objective: There have been immense developments in AI over recent years, culminating in the development of the Large Language and Multimodal models. These have made AI accessible to everyone, majorly changing the way in which people can use AI in a whole range of tasks, from seeking information to carrying out routine tasks to quickly developing sophisticated codes. These changes are greatly affecting the way in which astronomical research and projects are carried out. It is absolutely necessary to familiarise the larger astronomical community, particularly the younger generation, with these developments and provide them new directions in which their research and development could move with the help of modern AI.
We are proposing a one-day workshop on Recent Developments in AI With Applications to Astronomy at the 45th meeting of the Astronomical Society of India (ASI) at BHU. In this workshop, we will aim to
- Acquaint participants with important developments in AI which have taken place over the last few years
- Have presentations of case studies on applications of these topics and
- Have hands-on sessions for demonstrating the use of some of the techniques.
The lectures will be given by experts, from India and possibly abroad, in AI and the use of AI in astronomy. The presentations and demo sessions will be mainly by younger astronomers who have demonstrated skills in the areas being covered.
The tentative plan of the workshop is as follows.
The session before lunch will have lectures and case studies covering
- Introduction to Deep Learning including Convolutional Neural Networks and Autoencoders
- Attention and Transformers
- Large Language Models
- Foundational Models for Astronomy
The post lecture session may have a remaining lecture and case studies, which will be followed by hands-on sessions. For these sessions, Python jupyter-notebooks with detailed documentation will be shared with the participants. These jupyter-notebooks will be run on Google’s Colab environment that provides GPU processing capabilities. Google Colab will minimize the time required for software installation on individual machines of the participants.
A detailed time table will be provided well before the workshop date to facilitate arrangements. A list of main speakers and other resource persons will also be provided.
Expected number of Participants: We expect about 100 participants and around 10 resource persons.
Organizers: Dr. Alkendra Pratap Singh (Banaras Hindu University), Dr. Akash Garg (IUCAA), and Prof. Ajit Kembhavi (IUCAA)