Accelerating Drug Discovery with Deep Reinforcement Learning: Molecular Generation Using Deep Q-Network
Bibliographic record
Abstract
The drug discovery process is traditionally time-consuming and expensive, relying on manual trial and error methods through the Design-Make-Test-Analyze (DMTA) cycle. High-throughput screening has facilitated molecular selection but remains cost-prohibitive, limiting accessibility to many research institutions. In recent years, artificial intelligence (AI) has revolutionized computational drug discovery, enabling more efficient molecular generation and optimization. Among these AI-driven methods, deep reinforcement learning (DRL) has emerged as a powerful technique for novel molecular design, guiding the optimization process toward desired chemical properties. This study employs a Deep Q-Network (DQN) to generate novel drug-like molecules and evaluates their relevance using molecular features extracted from the ZINC dataset. Notably, increasing the number of actions in the DQN model enabled broader chemical space exploration, resulting in structurally diverse and meaningful compounds. The generated molecules exhibit favorable drug-like properties, such as optimized LogP values, molecular weight, and hydrogen bonding profiles. Both generated compounds achieved 100 % chemical validity, demonstrating that reinforcement learning (RL) can effectively navigate the chemical space and accelerate the identification of promising therapeutic candidates. Compared to prior works such as ReLeaSE, GCPN, and MolDQN, our model offers competitive results with simpler architecture and consistently high chemical validity. By streamlining the DMTA cycle, this approach offers a cost-effective and scalable path toward AI-guided drug discovery.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 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 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".