Research map / Vision-language and multimodal / Visual question answering

Visual Question Answering for Peruvian Cuisine in Regional Spanish

AAAI 2025 · Vision-language and multimodal › Visual question answering · approach: Universal VQA models · 0 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a planned VQA system for Peruvian cuisine in regional Spanish, combining a 4,000-image dish dataset with LLaVA-1.5 and GPT-2 Large, evaluated by metrics and culinary experts.

Problem

Existing VQA systems are mostly English-centric and do not capture regional Spanish naming variations of Peruvian dishes or Peruvian culinary culture.

Approach

Build a culturally grounded, regionally diverse Spanish VQA dataset for Peruvian gastronomy (images from PeruFoodNet, social media, Google Maps, culinary books, annotated by students from coast, highlands and jungle) and adapt LLaVA-1.5 and GPT-2 Large to it, with expert human evaluation.

Key results

• No experimental results are reported; the paper is a project proposal.
• Cites prior work reporting LLaVA-1.5 achieving 32.5% accuracy on the food topic in a Spanish VQA dataset (Romero et al. 2024).

Limitations

• No experiments or results are reported; the work is a proposal of planned research.
• Evaluation design is vague: zero-shot evaluation is described while also calling the models fine-tuned.
• Dataset construction details (annotation protocol, licensing for scraped content) are not specified.

Experimental setup

Datasets: PeruFoodNet

Metrics: accuracy perplexity F1 score

Baselines: LLaVA-1.5 GPT-2 Large

Builds on: LLaVA-1.5 GPT-2 Large MarIA

Closest papers by meaning

• CVQA: Culturally-diverse Multilingual Visual Question Answering Benchmark (NeurIPS 2024 · similarity 0.86)
• When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 2025 · similarity 0.83)
• Answer Them All! Toward Universal Visual Question Answering Models (CVPR 2019 · similarity 0.83)
• KVQA: Knowledge-Aware Visual Question Answering (AAAI 2019 · similarity 0.83)
• BOK-VQA: Bilingual outside Knowledge-Based Visual Question Answering via Graph Representation Pretraining (AAAI 2024 · similarity 0.82)
• World in a Frame: Understanding Culture Mixing as a New Challenge for Vision-Language Models (CVPR 2026 · similarity 0.82)