AgentBuild: A Framework for Constructing Scientific Agents Through Scientist-Authored Contracts
Researchers introduced AgentBuild, a framework that constructs LLM-based scientific agents from scientist-authored contracts rather than relying solely on fine-tuning or prompt engineering. The system was demonstrated on Rietveld refinement of X-ray diffraction data, using a rubric-driven approach with difficulty-graded curricula and curated knowledge bases. This approach aims to preserve scientific judgment in agent development while enabling systematic evaluation and adaptation as underlying models evolve.
AgentBuild proposes a new development paradigm for scientific agents that shifts control from automated optimization methods back to scientists through explicit contracts. Rather than relying on fine-tuning, reinforcement learning, or prompt-and-go approaches, the framework treats agent construction as a deliberate workflow stage where scientists author a version-controlled rubric, difficulty-graded curriculum, and curated external knowledge base. A meta-optimizer coding agent then edits the scientific agent within declared boundaries, with a rubric-driven judge ensuring quality. The authors instantiated this for Rietveld refinement of X-ray diffraction data using GSAS-II, testing on lithium lanthanum zirconium oxide (LLZO) samples across a signal-to-noise ladder. The framework identifies workflow-scope limits as contract failures rather than pattern-fitting failures, and allows re-tuning when base models evolve without requiring complete rebuilds.
What's missing
The paper does not discuss computational costs, runtime performance comparisons with alternative approaches, or quantitative benchmarks against other scientific agent construction methods. The study's own limitations regarding generalization to other scientific domains beyond Rietveld refinement are not explicitly addressed in the abstract.
What different sources said
- arXiv cs.AICenter
Fantastic Scientific Agents and How to Build Them: AgentBuild for Rietveld Refinement
Related
Topology-Aware Thermodynamics Improves DNA Probe Specificity Design
Researchers developed a new framework for designing DNA probes that accounts for the spatial organization of matched sequences, not just overall thermodynamic stability. Traditional methods rely on scalar measures like melting temperature and free energy, which miss how mismatches are distributed along the probe. The approach could improve diagnostic accuracy in applications like HPV detection and gene expression profiling.
Study Identifies Optimal Thermal Dose for Combining Focused Ultrasound with Immunotherapy in Tumors
Researchers used multimodal PET imaging to identify an optimal thermal dose range for focused ultrasound ablation that destroys tumor tissue while preserving conditions for immunotherapy delivery. The study found that excessive heating collapses blood vessels needed for antibody access, while insufficient heating fails to adequately reduce tumor burden. The findings could guide clinical design of combination treatments pairing thermal ablation with immunotherapies.
Plant MSH1 Protein Functions as Mismatch-Directed Nuclease for Organelle Genome Maintenance
Researchers have identified the precise mechanism by which the AtMSH1 protein in Arabidopsis plants recognizes and cleaves DNA mismatches and lesions, preventing mutations in organellar genomes. The protein combines a DNA mismatch recognition module with a nuclease domain that makes staggered cuts at specific positions relative to DNA damage. This discovery explains how plants maintain unusually low mutation rates in their mitochondrial and chloroplast DNA compared to other eukaryotes.