Neuro-symbolic approach boosts LLM storytelling coherence

Researchers have explored a neuro-symbolic architecture for interactive storytelling systems, aiming to improve coherence compared to purely LLM-based approaches. Their method uses LLMs to trigger pre-programmed world-state transformations, which helps maintain consistency while allowing for player creativity. An exploratory evaluation using Llama 3 70B and Gemini 1.5 Flash in English and Spanish showed that this hybrid approach can enhance player expression and address common incoherence issues. AI

IMPACT This hybrid approach could lead to more engaging and coherent AI-powered interactive narrative experiences.

RANK_REASON The cluster contains an academic paper detailing a new approach to interactive storytelling. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Gemini 1.5 Flash
• Llama 3 70B
• Santiago Góngora

• paper
• other

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