Everyday Physics in Korean Contexts: A Culturally Grounded Physical Reasoning Benchmark

2025-09-22

AI Summary

Existing physical commonsense reasoning benchmarks predominantly reflect Western cultural contexts, overlooking how cultural differences influence physical problem-solving. To address this gap, we introduce EPiK—the first Korean-culture-specific physical commonsense reasoning benchmark—comprising 181 binary-choice questions spanning nine reasoning categories (e.g., kimchi fermentation) and 84 culturally grounded scenarios. EPiK is constructed via a two-stage, culture-context-driven generation pipeline followed by rigorous expert validation, ensuring both physical accuracy and cultural authenticity. Experimental results demonstrate that culturally adapted models significantly outperform general-purpose foundation models, exposing critical limitations of current large language models in culture-specific physical reasoning. EPiK thus fills a key void in non-Western physical commonsense evaluation and empirically validates the essential role of culturally aware benchmarks in enhancing language models’ real-world situational understanding.

Technology Category

• Knowledge Representation and Reasoning: Common-Sense Reasoning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Philosophy and Ethics of AI: AI & Epistemology

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Existing physical commonsense reasoning benchmarks predominantly focus on Western contexts, overlooking cultural variations in physical problem-solving. To address this gap, we introduce EPiK (Everyday Physics in Korean Contexts), a novel benchmark comprising 181 binary-choice problems that test physical reasoning within Korean cultural contexts, ranging from kimchi (Korean food) to traditional fermentation. EPiK is constructed using a two-stage generation and verification pipeline to create culturally-authentic problems across 9 reasoning subtasks and 84 scenarios. Unlike approaches based on simple translation, our method generates problems organically from Korean contexts while upholding rigorous physical reasoning standards. Our evaluations show that Korean-specialized models consistently outperform general-purpose models of comparable size. This performance gap highlights the limitations of culturally-agnostic models and demonstrates the critical need for culturally-aware benchmarks to truly measure language understanding. Our EPiK is publicly available at https://huggingface.co/datasets/jjae/EPiK.

Problem

• Addressing cultural bias in physical reasoning benchmarks favoring Western contexts
• Creating culturally-authentic Korean physical reasoning problems across diverse scenarios
• Evaluating limitations of culturally-agnostic models in physical commonsense reasoning

Innovation

• Culturally-grounded Korean physical reasoning benchmark
• Two-stage generation and verification pipeline method
• Organically generated problems from Korean contexts