{"slug": "mard-mirror-augmented-reasoning-distillation-for-mechanism-level-drug-drug", "title": "MARD: Mirror-Augmented Reasoning Distillation for Mechanism-Level Drug-Drug Interaction Prediction", "summary": "Researchers have developed MARD (Mirror-Augmented Reasoning Distillation), a 7-billion-parameter AI system that predicts how drugs interact at the mechanism level, identifying specific enzymes or pathways involved rather than just flagging potential interactions. In tests against 32 systems using the April-2026 DrugBank release, MARD-7B was the only model whose accuracy remained stable when encountering novel drug pairs, outperforming the best baseline by 13.9 percentage points and GPT-4o by 6.7 points at roughly 1% of the cost. The system's performance gains stem from structured pharmacological reasoning rather than memorization of frequently seen drugs, with accuracy actually improving on rarely encountered medications.", "body_md": "arXiv:2606.12578v1 Announce Type: new\nAbstract: Mechanism-level drug-drug interaction (DDI) prediction requires identifying which enzyme or pharmacodynamic axis is implicated, in which direction, and with which evidence -- not merely whether two drugs interact. We introduce a reproducible mechanism-level DDI labelling and evaluation protocol with a structured 7-family/147-subtype taxonomy, leakage-safe cold-split protocols, and auditable reasoning metrics for evaluating pharmacological prediction beyond flat interaction classification. We propose a pipeline that produces a 7B reasoning MARD (Mirror-Augmented Reasoning Distillation), combining three training innovations: a single-token KL divergence on direction tag that ties the model's prediction, per-loss PRM-weighted DPO with programmatic hard negatives, and a leakage-safe mechanism-aware retrieval channel. Process-reward step labels are automatically verifiable against DrugBank-structured fields, requiring no human or LLM judges. On the April-2026 DrugBank release, our MARD-7B is the only system in a 32-system comparison whose accuracy survives drug-pair novelty, beating the best baseline by +13.9 pp and GPT-4o by +6.7 pp at ~1% of frontier API cost. Further analysis reveals an anti-memorisation signature where accuracy improves on rarely seen drugs, suggesting that gain comes from structured pharmacological reasoning rather than drug-frequency memorisation. We release corpus, DDI-PRM, retrieval index, and training code.", "url": "https://wpnews.pro/news/mard-mirror-augmented-reasoning-distillation-for-mechanism-level-drug-drug", "canonical_source": "https://arxiv.org/abs/2606.12578", "published_at": "2026-06-12 04:00:00+00:00", "updated_at": "2026-06-12 04:54:45.073646+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "large-language-models", "natural-language-processing", "ai-research"], "entities": ["MARD", "DrugBank", "GPT-4o", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/mard-mirror-augmented-reasoning-distillation-for-mechanism-level-drug-drug", "markdown": "https://wpnews.pro/news/mard-mirror-augmented-reasoning-distillation-for-mechanism-level-drug-drug.md", "text": "https://wpnews.pro/news/mard-mirror-augmented-reasoning-distillation-for-mechanism-level-drug-drug.txt", "jsonld": "https://wpnews.pro/news/mard-mirror-augmented-reasoning-distillation-for-mechanism-level-drug-drug.jsonld"}}