Search papers, labs, and topics across Lattice.
The paper introduces SciDC, a method to reduce LLM hallucination by integrating subject-specific scientific knowledge as strong constraints during decoding. SciDC uses LLMs to convert flexible knowledge into multi-layered, standardized rules, creating an extensible framework for domain-specific tasks. Experiments across industrial formulation design, clinical tumor diagnosis, and retrosynthesis planning show a 12% average accuracy improvement over vanilla generation, demonstrating the effectiveness of knowledge-driven constraints.
LLMs can be made significantly more reliable in scientific domains by automatically converting flexible knowledge into hard constraints during decoding, boosting accuracy by 12% on average.
Large language models (LLMs) have shown strong knowledge reserves and task-solving capabilities, but still face the challenge of severe hallucination, hindering their practical application. Though scientific theories and rules can efficiently direct the behaviors of human manipulators, LLMs still do not utilize these highly-condensed knowledge sufficiently through training or prompting. To address this issue, we propose \textbf{SciDC}, an LLM generation method that integrate subject-specific knowledge with strong constraints. By adopting strong LLMs to automatically convert flexible knowledge into multi-layered, standardized rules, we build an extensible framework to effectively constrain the model generation on domain tasks. Experiments on scientific tasks including industrial formulation design, clinical tumor diagnosis and retrosynthesis planning, consistently demonstrate the effectiveness of our method, achieving a 12\% accuracy improvement on average compared with vanilla generation. We further discuss the potential of LLMs in automatically inductively summarizing highly-condensed knowledge, looking ahead to practical solutions for accelerating the overall scientific research process. All the code of this paper can be obtained (https://github.com/Maotian-Ma/SciDC).
