The Human-In-the-Loop Drug Design Framework with Equivariant Rectified Flow
Bibliographic record
Abstract
Abstract Advancements in AI-based drug design often face obstacles due to incomplete datasets, hindering progress to clinical trials. Human experts bring invaluable expertise and nuanced contextual understanding to drug design. There are two main difficulties in integrating human knowledge into the drug development process. First, human annotations are costly, and traditional machine learning algorithms require a large number of samples to be effective. Second, human experts are unable to accurately describe their expertise using natural language, nor can they precisely specify what kind of molecule is needed before seeing the generated molecules. To address these problems, we propose a new platform, called HIL-DD, for experts to infuse their experience by selecting molecules generated by AI that meet their criteria, or discarding those that do not. The core generative technology utilizes an Equivariant Rectified Flow Model (ERFM), which offers faster generation speeds than conventional diffusion models, enabling efficient human-AI collaboration. More importantly, we provide a user-friendly interface to ensure smooth and effective collaboration between human experts and AI systems. Rigorous experiments demonstrate that our system can produce 3D molecules that align with expert expectations in minimal interactive sessions. These generated molecules maintain drug-like qualities comparable to those created by current state-of-the-art models.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".