AIRI (Artificial Intelligence Research Institute)

We work on universal AI systems that solve real-world problems.

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

AmbiK was presented at #ACL2025.
Paper: https://t.co/02YQyVzxkq
GitHub: https://t.co/89wWASNuSU

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

We introduce AmbiK — a textual dataset of 1000 ambiguous vs. unambiguous task pairs for a kitchen robot. Tasks are labeled by ambiguity type (Human Preferences, Common Sense, Safety) and include environment descriptions, user intents, clarifying Q&A, and task plans.

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

Why AmbiK matters? It provides a standardized benchmark for testing how well LLMs handle ambiguity — a crucial step toward safer, smarter, more human-aligned agents.

[apolinario (poli) @multimodalart]

HairFastGan is here to prove GANs are not dead! 💇

Check how you'd look with any haircut 💇‍♀️ with the demo on @huggingface

Thanks for the super cool paper/pipeline @AIRI_inst ✂️

▶️ https://t.co/VYzUczXKqB

[Aibek Alanov @ai_alanov]

📢 Exciting News! Our paper on StyleDomain for One-shot and Few-shot Domain Adaptation, accepted to ICCV 2023, is out! 📝🔥

📄 Paper Link: https://t.co/XTvjw9jhEV
🔗 Source Code: https://t.co/M4die1h6DC

#ICCV2023 #GANs #StyleDomain #DomainAdaptation

1/N 🧵

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

All lectures will be posted on Youtube this week! Subscribe and stay tuned: https://t.co/qnXs961NHe

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

The 2nd AIRI Summer School for undergraduate and graduate students has started at Innopolis University 👾

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

The students will be able to apply the knowledge gained during the lectures at practical seminars and in project activities, following which they will present reports on the results of their work.

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

Weekend reading: a fresh blog post about how scientists made domain adaptation for generative adversarial networks five thousand times faster: https://t.co/Uw9KTFmra7 📃

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

Recently, research fellows from AIRI have been presenting their reports at 61st Annual Meeting of the Association for Computational Linguistics, Toronto, Canada! 👾

Here they are:

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

“Efficient Out-of-Domain Detection for Sequence to Sequence Models”. Here is a link to the article, poster and their presentation: https://t.co/IiWzhJeQjj

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

Meet you in 2024! #ICLR2023

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

List of AIRI publications at @iclr_conf 👀

[AIRI (Artificial Intelligence Research Institute) @AIRI_inst]

Neural Symbolic Integration research group took part in @NeSy_GeMs workshop:

– https://t.co/76aTSp7Ta4