Thinking in Many Modes: How Composite Reasoning Elevates Large Language Model Performance with Limited Data
This addresses the need for more robust and adaptive problem-solving in LLMs for applications like scientific and medical question-answering, though it appears incremental as it builds on existing reasoning methods.
The paper tackles the problem of large language models relying on singular reasoning paradigms by introducing Composite Reasoning, which dynamically combines multiple reasoning styles, resulting in outperforming existing baselines like Chain-of-Thought and DeepSeek-R1 on scientific and medical benchmarks with superior sample efficiency.
Large Language Models (LLMs), despite their remarkable capabilities, rely on singular, pre-dominant reasoning paradigms, hindering their performance on intricate problems that demand diverse cognitive strategies. To address this, we introduce Composite Reasoning (CR), a novel reasoning approach empowering LLMs to dynamically explore and combine multiple reasoning styles like deductive, inductive, and abductive for more nuanced problem-solving. Evaluated on scientific and medical question-answering benchmarks, our approach outperforms existing baselines like Chain-of-Thought (CoT) and also surpasses the accuracy of DeepSeek-R1 style reasoning (SR) capabilities, while demonstrating superior sample efficiency and adequate token usage. Notably, CR adaptively emphasizes domain-appropriate reasoning styles. It prioritizes abductive and deductive reasoning for medical question answering, but shifts to causal, deductive, and inductive methods for scientific reasoning. Our findings highlight that by cultivating internal reasoning style diversity, LLMs acquire more robust, adaptive, and efficient problem-solving abilities.