Straddling two disciplines is a challenge that researcher Dr Ioana (Jo) Ciucă took on when she joined The Australian National University (ANU) School of Computing as a Joint Jubilee Fellow. Now at the completion of her joint fellowship, she reflects on the experience and what’s next.
“It’s been a truly wild ride. There are so many paths you can take with this fellowship.”
Dr Ciucă works at an exciting intersection of science and technology. During her fellowship at ANU, she has enjoyed the opportunity to combine the fields of astrophysics and artificial intelligence (AI).
“As an academic, you have an idea of the box you fit into with your research. Before this, I saw myself primarily as an astronomer but with a great love for technology and computer science.
“This fellowship gave me the opportunity to be part of these two worlds. I learnt from my joint appointment how quickly your identity can expand and in ways I didn’t predict when I started.”
Exploring a galaxy of research pathways
The Joint Jubilee Fellowships were created at ANU to encourage interdisciplinary and multidisciplinary research – the kind of research that’s at the forefront of computing and addresses major challenges in science, health, and society.
For Dr Ciucă, the fellowship has delivered on what she hoped, creating opportunities that would not have been available if she had remained solely focused on one field.
“I knew I wanted to connect astronomy with computer science methodologies and AI, as well as bring people together from both worlds.”
In fact, with so many research pathways on offer, her initial proposal wildly differed to where she ended up.
Dr Ciucă originally planned to focus on astrophysics and dark matter. Her goal was to focus on the illusive search for the dark matter particle using machine learning methods capable of identifying the needle in a haystack.
However, with the boom of generative AI around 2022, conversations on ChatGPT gave rise to exciting new research questions.
“AI was exponentially growing. The tools were improving faster than we could rewrite policy or work through how we might utilise them.”
Dr Ciucă was inspired to explore ways to democratise science using artificial intelligence. It proved a fortuitous time to pivot; leaning into a passion for computing to make strides in AI for science and Large Language Models.
“One thing I didn’t realise then is the opportunity for growth. My experience made me realise that you are responsible for the vision behind your work and to accept that you’ll make mistakes as you learn how to get there. It brings a lot of freedom.”
Providing space to grow
Dr Ciucă credits her supervisor, Professor Amanda Barnard, for allowing space to explore and has enjoyed bringing people together for unlikely collaborations.
“I see this joint appointment as a bridge to discover what you don’t expect and to bring people together. I realised that is my favourite way to conduct research and enjoyed becoming a catalyst, one who opens doors for interdisciplinary research and brings together great teams of researchers.”
Transitioning from individual to group research was a rewarding step into leadership for Dr Ciucă, and an unexpected opportunity for personal growth.
“The fellowship made me realise that research is an ongoing process. Not just the concrete outcomes, but the process of improving and building yourself and your skills to lead projects.”
“There are many pathways you can take when doing research, and you are on that road without knowing what happens afterwards. It’s about trusting the process and trusting yourself as well.”
Dr Ciucă also marvels at the growth her students have demonstrated and is proud to see many go on to take on leadership roles.
“I am very proud of the collaborations and community I have built at ANU and with other institutions. I particularly enjoyed seeing our team share their ideas and our youngest undergraduate and honours students become so passionate about their projects.
“I find meaning in empowering my junior colleagues, and I am grateful that this fellowship gave me that opportunity.”
“Working at the interface between disciplines is difficult and takes courage. Jo is an inspiration to other computational scientists, demonstrating how the right motivation and hard work can pay off.”— Senior Professor Amanda Barnard, Assistant Director ANU School of Computing
A search engine with superpowers
Dr Ciucă co-founded UniverseTBD, an AI research group combining more than 30 astronomers, software engineers, computer scientists and artists from around the world. Their goal is to democratise science for everyone using tools such as AI to help others explore, synthesise, and create knowledge.
Among the team’s proudest achievements is Pathfinder, an AI-powered tool that helps make astronomy research more accessible for all. It is a collaboration led by her cofounder Dr Kartheik Iyer from Columbia University, working with the Evaluating Large Language Models for Astro team led by Dr John Wu from the Space Telescope Science Institute, and with Johns Hopkins University as part of the JSALT24 Frederick Jelinek Memorial Summer Workshop.
The tool allows users to ask a large language model questions like, ‘what can you tell me about the Sun?’’ or ‘how did the Milky Way form?’. It can even understand questions that require analysis of opinion, such as ‘what is the consensus of scientists on finding dark matter at the moment?’.
Anyone in the community can use Pathfinder to explain astronomy research and jargon in simplified terms, or even in other languages. It summarises vast amounts of data – more than 300,000 peer-reviewed papers – identifies patterns in information and finds a path to the answer.
While Pathfinder has been designed for speedy literature reviews of astronomy topics, Dr Ciucă and the team hope to expand its capabilities to other sciences.
“It’s kind of like a highly advanced search engine. Unlike other tools such as ChatGPT—which can hallucinate answers, especially in highly specialised fields—Pathfinder is focused on finding the correct answer.”
Other notable lab projects include HypoGen; AstroTalks and ‘AstroLLaVA’, one of the first specialised multimodal large language model for Astronomy.
Paving the path toward democratised science
Through her Jubilee Joint Fellowship with the ANU School of Computing, there’s no doubt Dr Ciucă has spearheaded innovative research at the intersection of astronomy and computer science.
However, an equally important objective has been to foster a collaborative community dedicated to making science more accessible.
Dr Ciucă’s project interests are all linked by a common thread: to tap into the vast potential of AI technology to revolutionise research and democratise scientific knowledge.
“Our mission has been to break down the barriers to knowledge,” she says. “Our two worlds were united by the mission to democratise science for everyone. With projects such as Pathfinder, we are making strides to get there.”
So, what’s next and what role could AI play in the future of academic research?
“I think there is the potential for AI to help generate novel and creative ideas. AI is an assistive tool that could play an active part in the co-creation process of scientific thought”, said Dr Ciucă.
“I deeply enjoy doing research and I want to do my part in empowering the next generation. Our hope at UniverseTBD is to be able to integrate this work on hypothesis generation with our scientific interests. Who knows, maybe AI can even help us identify dark matter.”
Watch Dr Ciucă’s STEMM guest lecture: Democratising science in the era of artificial intelligence | ANU College of Science