Bryan Hooi
@BryanHooi1
Followers
234
Following
20
Media
0
Statuses
14
Assistant Prof at NUS
Joined March 2022
๐ขCFP: Join us for the "AI for Scientific Research" workshop at AAAI 2026 in Singapore (Jan 26, 2026)! https://t.co/62YMMB9KJw ๐
We invite submissions of papers by Oct 22, 2025: https://t.co/yVO5zBon3r
#AAAI2026 #AI4Research
2
3
6
๐ Excited to share our work โConfTuner: Training Large Language Models to Express Their Confidence Verballyโ accepted at #NeurIPS2025! Arxiv: https://t.co/xAQywt82hb Code and checkpoints: https://t.co/oxJSW5Jzjx
5
4
10
๐คHow well can AI agents conduct open-ended machine learning research? ๐Excited to share our latest #AI4Research benchmark, MLR-Bench, for evaluating AI agents on open-ended machine learning research!!๐ https://t.co/qTtVY4YPlh 1/
1
6
11
๐จ Models like GPT-5-mini and Gemini 2.5 Flash hallucinate more with longer thinking! ๐คLonger thinking encourages the model to attempt more questions. It also induces confirmation bias, resulting in overconfident hallucinations. Check out https://t.co/nsVArMooaZ
#LLM
2
2
5
๐ Exciting News! Workshop on Reasoning and Planning for Large Language Models @ ICLR 2025 is coming ๐ Please visit our official website: ๐ https://t.co/KNp4kzyQvG With the release of o1 Pro and the growing interest in research on reasoning and planning capabilities of LLMs,
8
16
105
Congrats, very well deserved!! @he_xiaoxin is indeed an outstanding researcher and has done pioneering work in Graphs+LLMs: https://t.co/2nKKOt0hPg It's my privilege to advise her and other excellent students, and very glad to see her success being recognized!
Thrilled to announce that I've been awarded the ๐๐ผ๐ผ๐ด๐น๐ฒ ๐ฃ๐ต๐ ๐๐ฒ๐น๐น๐ผ๐๐๐ต๐ถ๐ฝ! Huge thanks to my advisors, @BryanHooi1 and @xbresson, for their incredible support and guidance! Grateful to @GoogleAI for this valuable honor! ๐๐ https://t.co/ry77GmFf6R
0
0
13
๐งฌ๐ค Introducing RNA-FrameFlow โโ an unconditional generative model for 3D RNA backbone design! ๐: https://t.co/JR9AzSArty ๐งฐ: https://t.co/Z7RwcdpR7f Our method generates โฅ 40% self-consistent *all-atom* RNA backbones that are globally and locally realistic ๐ช๐ป 1/9
3
29
134
Delighted to announce that our work has been accepted for presentation at the ICLR 2024 @LLMAgents workshop in Vienna. Received exceptionally positive feedback from the reviewers and looking forward to sharing our insights and contributions with the community.
How do LLMs conduct reasoning and planning given partial information with uncertainty? Whether they can proactively ask questions to improve decision-making? In joint work with UW, NTU, Yale and UCL, we introduce the UoT method, which boosts the information-seeking and
0
2
10
๐ช๐๐๐ ๐๐๐๐ ๐๐๐๐ ๐ฎ๐๐๐๐ ๐ฃ๏ธ G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering ArXiv: https://t.co/NdcIKy2urF GitHub: https://t.co/J9yEdE5y14 BlogPost: https://t.co/JvddZq9xxo
4
36
194
One of my favourite letters, insightful & humane, from Richard Feynman to a former student who was having a rough time I was reflecting on it as an argument in favour of scope insensitivity, or even in favour of smaller problems:
50
585
3K
๐ Excited to introduce MultiModal Graph Learning (MMGL) - exploring beyond the typical one-to-one data modality pairs๐ to uncover more complex and multifaceted relationships among data modalities! ๐ ๐ Paper: https://t.co/nP6JtEbdOA ๐งต Thread continues...
4
57
217
Thrilled to share paper on LLMs' calibration capability! "Can LLMs Express Their Uncertainty?" We explore confidence elicitation in LLM,revealing mitigation of LLMs' overconfidence. However,improvements are needed. Arxiv: https://t.co/yxCImfH9Zt
#LLM #Calibration #Uncertainty
0
14
66
New paper w/ @he_xiaoxin, T. Laurent, @BryanHooi1 "Explanations as Features: LLM-Based Features for Text-Attributed Graphs" ArXiv: https://t.co/TerY447HYJ GitHub: https://t.co/eQNnP8BTso
2
31
105
๐Awesome-Graph-LLM Check out this GitHub repo for a collection of AWESOME things about Graph-Related Large Language Models (LLMs):
github.com
A collection of AWESOME things about Graph-Related LLMs. - XiaoxinHe/Awesome-Graph-LLM
7
59
326