Kristina Gligoric

Tutorial Day kicks off tomorrow! Which sessions are you going to attend? 📚 Check the lineup and share your picks. #ic2s2 #tutorials

Excited to co-organize the Bridging Human and LLM Annotations for Social Science tutorial at #ic2s2 tomorrow!

Join us at 1:30pm CEST to learn how to combine Qualtrics+Prolific, MTurk, & LLMs for automated annotation and valid estimates.

Code&materials: github.com/kristinaglig...

Thank you, Patrick!

Thank you, Sarah!!

Thank you!

Thank you, Maria!

Thank you!!! Yaay east coast crew!

Thank you, Dallas!

Thank you!!

Thank you, Navita!

Thank you!!

Thank you, Marianne!!

Yaay!!!

Thank you! <3 <3

Thank you, Adam!

Sooo lucky to be colleagues! :))

I’ll be starting my lab and planning future projects. If you're interested in research at the intersection of NLP/CSS, please reach out! See my website for more details kristinagligoric.com

My amazing husband @tiziano.bsky.social is also joining Johns Hopkins CS as an Assistant Professor, working on social computing&HCI 💙. We’re beyond excited for this next chapter. I'm so grateful to my mentors and collaborators!

I'm excited to announce that I’ll be joining the Computer Science department at Johns Hopkins as an Assistant Professor this Fall! I’ll be working on large language models, computational social science, and AI & society—and will be recruiting PhD students. Apply to work with me!

Recently, I’ve been busy with the job market and exploring how LLMs can understand, simulate, and influence human behavior. If you’re working on anything related, I’d love to connect! 🤓

And, on Fri at 12pm at the AI & Scientific Discovery Workshop, I’ll present our work on how LLMs can support food scientists and chefs in designing more sustainable diets 🍽️🤖

📝 Paper: arxiv.org/pdf/2503.04734

I'm at #NAACL2025 & would love to chat!

On Thu, I'll be presenting our paper on LLM annotations for valid estimates in social science (oral talk @ 4pm, CSS session)

📝 Paper: aclanthology.org/2025.naacl-l...
*New* 🧪 tutorial (LLMs + Prolific data collection): github.com/kristinaglig...

Check out the full paper for more details!

URL: bit.ly/3ERJbUV

w/ @annatthomas.bsky.social, Adam Yee, Andrew Mayne, Maya Mathur & @jurafsky.bsky.social

For a menu design task, equipping LLMs with optimization tools decreases emissions of food choices by 79% in a hypothetical restaurant while maintaining participants’ satisfaction. Simpler baselines (vegetarian only, naive prompting) decrease emissions, too, but they also decrease satisfaction.

Highlights: Food science experts estimated that collaboration with an LLM reduces time spent on a sustainable protein design task by 45%, compared to 22% for collaboration with another expert human food scientist. LLMs generate accurate, specific, and complementary suggestions.

🤖🍲 What can LLMs do for sustainable food? 🤖🍲

We collaborated with domain experts (food scientists and chefs) to define a typology of food design and prediction tasks.

LLMs can assist in food and menu development, saving food scientists' time and reducing emissions!

URL: bit.ly/3ERJbUV

Highlights: Food science experts estimated that collaboration with an LLM reduces time spent on a sustainable protein design task by 45%, compared to 22% for collaboration with another expert human food scientist. LLMs generate accurate, specific, and complementary suggestions.

Our new piece in Nature Machine Intelligence: LLMs are replacing human participants, but can they simulate diverse respondents? Surveys use representative sampling for a reason, and our work shows how LLM training prevents accurate simulation of different human identities.

Can Unconfident LLM Annotations Be Used for Confident Conclusions? Large language models (LLMs) have shown high agreement with human raters across a variety of tasks, demonstrating potential to ease the challenges of human data collection. In computational social sci...

Thanks for sharing! We also have this NAACL work that shows how verbalized confidence is calibrated and useful in CSS tasks! arxiv.org/abs/2408.15204

whale breaching

spectrogram of whale song

Research in @science.org showing structures in whale song thought to be uniquely human! Why? We think cultural learning!

Led by @inbalarnon.bsky.social with @simonkirby.bsky.social @ellengarland.bsky.social @clairenea.bsky.social @emma-carroll.bsky.social and me!

www.science.org/doi/10.1126/...