Learning To Navigate The Synthetically Accessible Chemical Space Using\n Reinforcement Learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Over the last decade, there has been significant progress in the field of\nmachine learning for de novo drug design, particularly in deep generative\nmodels. However, current generative approaches exhibit a significant challenge\nas they do not ensure that the proposed molecular structures can be feasibly\nsynthesized nor do they provide the synthesis routes of the proposed small\nmolecules, thereby seriously limiting their practical applicability. In this\nwork, we propose a novel forward synthesis framework powered by reinforcement\nlearning (RL) for de novo drug design, Policy Gradient for Forward Synthesis\n(PGFS), that addresses this challenge by embedding the concept of synthetic\naccessibility directly into the de novo drug design system. In this setup, the\nagent learns to navigate through the immense synthetically accessible chemical\nspace by subjecting commercially available small molecule building blocks to\nvalid chemical reactions at every time step of the iterative virtual multi-step\nsynthesis process. The proposed environment for drug discovery provides a\nhighly challenging test-bed for RL algorithms owing to the large state space\nand high-dimensional continuous action space with hierarchical actions. PGFS\nachieves state-of-the-art performance in generating structures with high QED\nand penalized clogP. Moreover, we validate PGFS in an in-silico\nproof-of-concept associated with three HIV targets. Finally, we describe how\nthe end-to-end training conceptualized in this study represents an important\nparadigm in radically expanding the synthesizable chemical space and automating\nthe drug discovery process.\n
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it