GLM-4-Voice uses RL to achieve state-of-the-art spoken math reasoning

Researchers have applied reinforcement learning to the GLM-4-Voice speech model to improve its mathematical reasoning capabilities. After supervised fine-tuning on spoken question-answering data, the model showed improved accuracy on the GSM8K benchmark, surpassing previous speech model performance without additional reasoning tokens. Further enhancements were achieved by integrating streaming reasoning techniques, leading to a new state-of-the-art accuracy of 74.8% for speech-native models in mathematical tasks. AI

IMPACT Establishes new state-of-the-art for speech-native models in mathematical reasoning, potentially improving human-machine interaction for complex tasks.

RANK_REASON The cluster describes a research paper published on arXiv detailing a new method for improving a speech model's mathematical reasoning capabilities. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• DagsHub
• GLM-4-Voice
• GSM8K
• Hugging Face
• Timothée Weisselberger
• Zeng et al.

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →