Ruchira’s Homepage

🎓 Short Intro: Hello there! I’m a PhD Researcher in the Department of Computer Science at the University of Copenhagen. I work with Dr. Anders Søgaard and the amazing people at CoAStaL to develop cognitive and human-centric evaluation frameworks for AI systems like LLMs. My work focuses on creating explainable and interpretable evaluation methods that makes analyzing these systems more accessible for all stakeholders. Additionally, I am involved with the Centre of Philosophy of AI at the university, exploring the societal and ethical dimensions of AI evaluation including AI policy and governance.

🧠 Research Interests: My primary research interest lies in building evaluation frameworks that bridge the gap between technical AI capabilities and human-centric needs. This involves studying the cognitive abilities and humanlike behaviors of AI systems to make their capabilities interpretable and useful for diverse stakeholders, including developers, policymakers, and end-users. I integrate insights from AI, Machine Learning, Cognitive Science, Linguistics, and Philosophy to advance this goal.

👥 Volunteering: I’m also a community lead at Cohere for AI (C4AI), where I host talks, socials, and meetups to foster open-source research and collaborations in AI, emphasizing ethical and accessible approaches.

🚶🏽‍♀️‍➡️ Hobbies: When I’m not immersed in research, I enjoy playing badminton and taking long, contemplative walks to the ass end of nowhere.

Updates

Super happy to announce that our paper titled “Investigating the role of modality and training objective on representational alignment between transformers and the brain” was accepted to the Proceedings Track Unireps Workshop at NeurIPS 2024! Read the paper here. Coming out soon in PMLR!

My work with co-author Constanza Fierro, titled “Defining Knowledge: Bridging Epistemology and Large Language Models” was accepted to EMNLP Main Conference, 2024! Read the paper here

The first work of my PhD- “From Words to Worlds: Compositionality for Cognitive Architectures” - was accepted to the LLMs & Cognition workshop at International Conference on Machine Learning (ICML) 2024, Vienna. Read the paper here

Our abstract for an oral presentation titled “The Completeness Problem: Beyond Human Metrics in Assessing Abilities of Cognitive Systems” was accepted to the International Symposium on Artificial Intelligence and Neuroscience (ACAIN) 2024.