Tianyu Liu
@meSuper8
Followers
226
Following
1K
Media
15
Statuses
250
BS @ #UIUC, PhD student @ #Yale
New Haven
Joined August 2021
🎵Glad to share our work (accepted by #NeurIPS2023) on using GNN for gene embedding analysis with @WangYuge4585 , Rex Ying and @HongyuZhao2. To efficiently integrate cross-modal relationships in multi-omics data, we propose a novel method to learn the representation of genes.
1
4
31
🚀 We’re hiring! Multiple postdocs & a program manager to push the frontiers of explainable AI in cutting-edge biomedical research—Alzheimer’s, aging, cancer & medical AI. Start immediately. Friends, please RT🙏 Learn more: https://t.co/pAR8O6EKNP
#postdocjobs #AI #Biomedicine
0
10
22
SuperGLUE facilitates an explainable training framework for multi-modal data analysis
cell.com
Liu et al. introduce SuperGLUE, a unified framework for integrating multi-modal single-cell data using probabilistic deep learning and graph-based explainability. It outperforms existing methods in...
0
4
2
🚀 Thrilled to share a preview of Biomni-R0 — we trained the first RL agent end-to-end for biomedical research. ➡️ nearly 2× stronger than its open-source base ➡️ >10% better than frontier closed-source models ➡️ Scalable path to hill climb to expert-level performance 🔗
18
67
318
I heard some NeurIPS ACs are rejecting papers with all positive reviews (5444) just to control acceptance rates, which is wrong. We had a similar AAAI experience years ago: all positive reviews 7776 rejected because of results reported as 0.84 instead of 84%. They didn't really
28
48
657
🔔 Two months ago, we released #IneqMath, which revealed the Soundness Gap: LLMs can guess answers to Olympiad-level inequalities problems, but still struggle to make rigorous proof steps. Since then, it's been downloaded 4K+ times on HuggingFace! ➡️ https://t.co/qnZLr3lKtn
Do LLMs truly understand math proofs, or just guess? 🤔Our new study on #IneqMath dives deep into Olympiad-level inequality proofs & reveals a critical gap: LLMs are often good at finding answers, but struggle with rigorous, sound proofs. ➡️ https://t.co/h5f8Qv8Xlv To tackle
0
7
25
Thank you to @MITEECS for highlighting @Schmidt_Center Director Caroline Uhler. Learn more about Caroline's work and why right now is a vital time to address problems across different levels of biological organization using ML and AI. https://t.co/4yAndKBSyl
@broadinstitute
0
4
20
What does it mean for an AI to reason? In our upcoming #WiAIR podcast (Aug 27), we chat with @HanSineng (PhD student at @Yale, intern at @Meta FAIR) about LLM reasoning, human reasoning chains, cognitive science insights, and what makes AI outputs trustworthy.
0
2
13
PepMLM is out in @NatureBiotech! 🎭Want a short binder for any target? Just input its sequence + desired binder length → PepMLM unmasks it. Simple? No way? The results speak for themselves. 👇 📜: https://t.co/RRZwQe82tu 💻: https://t.co/8l3nSNEtHo 🤗: https://t.co/9aSfxzZZTo
7
61
286
spVelo is now in Genome Biology!
0
2
4
I've written the full story of Attention Sinks — a technical deep-dive into how the mechanism was developed and how our research ended up being used in OpenAI's new OSS models. For those interested in the details: https://t.co/0EAi2KQMMx
39
275
2K
Biomni open source version 0.0.5 is now released! 🎉 Since our last release, we’ve seen amazing contributions from the community: 📊 New Data Sources - OpenFDA and Monarch databases added, panhumanpy for cell type annotation, DepMap data for CRISPR screens and cancer cell lines
0
7
32
✅ Published in @Nature today, the paper describing the initial whole-genome sequencing analysis of 500,000 UK Biobank participants. https://t.co/QmyovKXEy1
1
64
198
Deep-learning-based gene perturbation effect prediction does not yet outperform simple linear baselines https://t.co/jr8uIhzeDh 🧬🖥️🧪 https://t.co/QFYiTjcOPq
#Rstats
2
10
101
We have the totally new GLM-4.5 with very good performance on ARC (agent, reasoning, and coding). Simply one day, it already become no.1 at https://t.co/SGRr4AYLgH 's trending for coding. Welcome to try. A free provider will be available very soon!
2
10
55
James Collins Lab researchers have discovered broad-spectrum antiviral compounds through the use of a novel optogenetic screen, symbolized in this image by a beam of light piercing a virus. https://t.co/NBnbYum49L
@CellCellPress @MIT_IMES
0
6
11
Incredibly honored and grateful to receive the Overton Prize at #ISMBECCB2025♥️ Many thanks to ISCB and my amazing students, collaborators and mentors!🙏
27
9
251
🎉 Big week at ISMB/ECCB 2025! Congrats to #GENETICS Senior Editor Hongyu Zhao, named an ISCB Fellow for his contributions to computational biology! 👏 And don’t miss #G3Journal Editor in Chief Lauren McIntyre’s “Meet the Editor” session—OUP booth, tomorrow, July 22, 10–11am!
0
3
3
Excited to be at #ISMBECCB2025! I'll be presenting at RegSys on "Learning the Regulatory Genome by Destruction and Creation" (Thursday 3:15PM in Room 11BC). I'm looking forward to connecting with both new and old friends throughout the conference! 🧬 #CRISPR #Genomics #Diffusion
1
3
27
The @cziscience workshop on AI Modeling Benchmarking and Evaluation for Biology ( https://t.co/ZpmBoaSAuo) was well organized with representation from academic, industry, and independent groups. This is a useful summary even though it can't capture all the discussion and ideas.
virtualcellmodels.cziscience.com
Learn about insights from a recent CZI workshop on building robust benchmarking for AI in biology
Benchmarking and Evaluation of AI Models in Biology: Outcomes and Recommendations from the CZI Virtual Cells Workshop 1. A recent workshop convened by the Chan Zuckerberg Initiative (CZI) brought together experts from imaging, transcriptomics, proteomics, and genomics to address
1
14
50