feilin-must commited on
Commit
d15922a
Β·
verified Β·
1 Parent(s): 1bedc62

Update org card: rebrand DeepYoke -> HydroSophy

Browse files
Files changed (1) hide show
  1. README.md +16 -39
README.md CHANGED
@@ -7,55 +7,32 @@ sdk: static
7
  pinned: false
8
  ---
9
 
10
- # HydroSophy AI Lab
11
 
12
- ## πŸš€ Our Vision
13
 
14
- **HydroSophy** is a non-profit research collective dedicated to building bridges between **artificial intelligence** and the **natural sciences**. Inspired by the Greek roots *Hydro* (water) and *Sophia* (wisdom), we believe that knowledge β€” like water β€” should flow freely, nourishing discovery, dissolving disciplinary boundaries, and shaping the contours of tomorrow's science.
15
 
16
- Our mission is to democratize scientific discovery by making advanced AI tools, rigorous benchmarks and high-quality datasets accessible to researchers worldwide, accelerating the pace of innovation across multiple scientific domains.
17
-
18
- ## 🌟 Research Focus
19
-
20
- ### 🧬 AI for Science
21
-
22
- - **Molecular Modeling & Drug Design** β€” Accelerating pharmaceutical research through intelligent molecular design and structure-level editing
23
- - **Structural Biology** β€” Understanding protein and small-molecule structures with AI
24
- - **Cheminformatics** β€” Bridging chemistry and computational intelligence
25
- - **Scientific Computing Acceleration** β€” Optimizing computational workflows for scientific discovery
26
-
27
- ### πŸ€– Multimodal Intelligence
28
-
29
- - **Multimodal Large Language Models** β€” Integrating vision, language and scientific reasoning
30
- - **Vision–Language–Molecule Understanding** β€” Joint reasoning over SMILES, 2D depictions and natural language
31
- - **Intelligent Scientific Literature Analysis** β€” Extracting structured insight from vast research corpora
32
- - **Cross-modal Knowledge Fusion** β€” Synthesizing information across heterogeneous scientific modalities
33
-
34
- ### πŸ“Š Benchmark & Evaluation
35
-
36
- - **Scientific AI Benchmarking** β€” Establishing rigorous, reproducible evaluation standards for AI4Science
37
- - **Dataset Construction & Standardization** β€” Creating high-quality, expert-annotated, mechanism-aware datasets
38
- - **Evaluation Framework Development** β€” Building transparent, extensible assessment toolkits
39
- - **Open Science Ecosystem** β€” Fostering collaborative, reproducible research environments
40
-
41
- ## πŸ† Featured Projects
42
 
43
  ### πŸ§ͺ ToxiMol  Β·  *KDD 2026 Oral*
44
 
45
- A comprehensive benchmark for **molecular toxicity repair**, evaluating the capabilities of Multimodal Large Language Models in structure-level molecular detoxification.
46
 
47
- **πŸ“Š [Dataset](https://huggingface.co/datasets/HydroSophyTech/ToxiMol-benchmark)**  |  **πŸ“„ [Paper](https://arxiv.org/abs/2506.10912)**  |  **πŸ”¬ [Code](https://github.com/HydroSophy/ToxiMol)**
48
 
49
- *Breaking Bad Molecules: Are MLLMs Ready for Structure-Level Molecular Detoxification?*
50
 
51
- ### πŸ”¬ Upcoming Projects
52
 
53
- - Multimodal foundation models for molecular reasoning
54
- - Mechanism-aware editing across broader therapeutic modalities
55
- - Open evaluation infrastructure for AI4Science
56
 
57
- ---
 
 
58
 
59
- *"Where algorithms meet atoms, where intelligence illuminates discovery, where language bridges knowledge and practice β€” this is **HydroSophy**."*
60
 
61
- **Coding the Future of Science, Illuminating Tomorrow with Wisdom**
 
 
 
7
  pinned: false
8
  ---
9
 
10
+ # HydroSophy
11
 
12
+ ## About
13
 
14
+ We work on **AI for Science** β€” mostly molecular modeling, multimodal foundation models, and the benchmarks / datasets / evaluation tools that go with them.
15
 
16
+ ## Projects
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
 
18
  ### πŸ§ͺ ToxiMol  Β·  *KDD 2026 Oral*
19
 
20
+ **Breaking Bad Molecules: Are MLLMs Ready for Structure-Level Molecular Detoxification?**
21
 
22
+ A benchmark for molecular toxicity repair with MLLMs β€” 660 toxic molecules, 11 repair tasks, multimodal inputs (SMILES + 2D structures), plus the companion **ToxiEval** evaluation framework.
23
 
24
+ πŸ“Š [Dataset](https://huggingface.co/datasets/HydroSophyTech/ToxiMol-benchmark)  |  πŸ“„ [Paper](https://arxiv.org/abs/2506.10912)  |  πŸ”¬ [Code](https://github.com/HydroSophy/ToxiMol)
25
 
26
+ ### πŸ” ToxiEval-ZKP  Β·  *IFAC BIKA Best Paper Award*
27
 
28
+ **ToxiEval-ZKP: A Structure-Private Verification Framework for Molecular Toxicity Repair Tasks**
 
 
29
 
30
+ A structure-private, verifiable evaluation framework for molecular toxicity repair built on zero-knowledge proofs.
31
+
32
+ πŸ“„ [Paper](https://arxiv.org/abs/2508.12035)  |  πŸ”¬ [Code](https://github.com/HydroSophy/ToxiEval-ZKP) *(coming soon)*
33
 
34
+ ## Contact
35
 
36
+ - **GitHub:** [@HydroSophy](https://github.com/HydroSophy)
37
+ - **Email:** feilin@ieee.org
38
+ - **Location:** Macau, China