Mixture of Reasoning Experts (MoRE) technique is designed to improve the generalization of Large Language Models (LLMs) across different question types in question answering (QA). While LLMs have shown impressive performance, they often struggle when handling questions that require different reasoning skills—such as factual, multihop, mathematical, or commonsense reasoning. MoRE aims to address this challenge by using specialized language models, each trained for a specific reasoning type.
MoRE also introduces a novel approach to selective QA, where the system decides when to abstain from answering if the confidence in its prediction is low. This ensures the system answers accurately when possible and avoids incorrect answers.
MoRE leverages a pool of specialized experts, where each expert is optimized for a distinct reasoning type, such as:
MoRE uses an answer selector to choose the best response based on predictions from the specialized experts. If the system detects that none of the answers are reliable, it can abstain from answering. Another key feature of MoRE is its ability to abstain from answering when it's unsure, improving the system's reliability.
Each expert in MoRE is designed to handle different reasoning challenges:
By combining these experts, MoRE ensures higher generalizability compared to single-model approaches.
MoRE is especially useful in situations requiring robust, interpretable question answering, such as:
Here’s a simplified view of the process:
Mixture-of-Reasoning Experts technique offers a novel solution to the challenges of QA by combining specialized models and leveraging inter-expert agreement for both generalizability and selective answering. By specializing models for different reasoning tasks and abstaining when appropriate, MoRE delivers both higher accuracy and better interpretability.
Valeriia Kuka, Head of Content at Learn Prompting, is passionate about making AI and ML accessible. Valeriia previously grew a 60K+ follower AI-focused social media account, earning reposts from Stanford NLP, Amazon Research, Hugging Face, and AI researchers. She has also worked with AI/ML newsletters and global communities with 100K+ members and authored clear and concise explainers and historical articles.
Si, C., Shi, W., Zhao, C., Zettlemoyer, L., & Boyd-Graber, J. (2023). Getting MoRE out of Mixture of Language Model Reasoning Experts. https://arxiv.org/abs/2305.14628 ↩ ↩2 ↩3 ↩4