Search papers, labs, and topics across Lattice.
The paper introduces Mozi, a dual-layer architecture for tool-augmented LLM agents in drug discovery, designed to address issues of unconstrained tool use and poor long-horizon reliability. Mozi employs a Control Plane (Layer A) for governed tool use and reflection-based replanning, and a Workflow Plane (Layer B) for structured execution of drug discovery stages with data contracts and human-in-the-loop checkpoints. Evaluations on PharmaBench and end-to-end case studies demonstrate Mozi's superior orchestration accuracy and ability to generate competitive in silico candidates while mitigating error accumulation.
LLMs can navigate massive chemical spaces and enforce toxicity filters in drug discovery, but only if you constrain them with a dual-layer architecture that combines free-form reasoning with structured execution.
Tool-augmented large language model (LLM) agents promise to unify scientific reasoning with computation, yet their deployment in high-stakes domains like drug discovery is bottlenecked by two critical barriers: unconstrained tool-use governance and poor long-horizon reliability. In dependency-heavy pharmaceutical pipelines, autonomous agents often drift into irreproducible trajectories, where early-stage hallucinations multiplicatively compound into downstream failures. To overcome this, we present Mozi, a dual-layer architecture that bridges the flexibility of generative AI with the deterministic rigor of computational biology. Layer A (Control Plane) establishes a governed supervisor--worker hierarchy that enforces role-based tool isolation, limits execution to constrained action spaces, and drives reflection-based replanning. Layer B (Workflow Plane) operationalizes canonical drug discovery stages -- from Target Identification to Lead Optimization -- as stateful, composable skill graphs. This layer integrates strict data contracts and strategic human-in-the-loop (HITL) checkpoints to safeguard scientific validity at high-uncertainty decision boundaries. Operating on the design principle of ``free-form reasoning for safe tasks, structured execution for long-horizon pipelines,''Mozi provides built-in robustness mechanisms and trace-level audibility to completely mitigate error accumulation. We evaluate Mozi on PharmaBench, a curated benchmark for biomedical agents, demonstrating superior orchestration accuracy over existing baselines. Furthermore, through end-to-end therapeutic case studies, we demonstrate Mozi's ability to navigate massive chemical spaces, enforce stringent toxicity filters, and generate highly competitive in silico candidates, effectively transforming the LLM from a fragile conversationalist into a reliable, governed co-scientist.
