# Multi-Sourced Compositional Generalization in Visual Question Answering

arXiv: https://arxiv.org/abs/2505.23045 · HTML: https://mlsanity.com/paper/2505.23045/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 36 references that are in the map.
- First posted on arXiv 29 May 2025.

## Builds on (most cited first, 24 of 24)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [VQA: Visual Question Answering](https://mlsanity.com/paper/1505.00468/) (2015)
- [Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering](https://mlsanity.com/paper/1707.07998/) (2017)
- [Counterfactual VQA: A Cause-Effect Look at Language Bias](https://mlsanity.com/paper/2006.04315/) (2020)
- [Compositional Attention Networks for Machine Reasoning](https://mlsanity.com/paper/1803.03067/) (2018)
- [Learning Graph Embeddings for Compositional Zero-shot Learning](https://mlsanity.com/paper/2102.01987/) (2021)
- [Plug-and-Play VQA: Zero-shot VQA by Conjoining Large Pretrained Models with Zero Training](https://mlsanity.com/paper/2210.08773/) (2022)
- [The paradox of the compositionality of natural language: a neural machine translation case study](https://mlsanity.com/paper/2108.05885/) (2021)
- [Compositional Temporal Grounding with Structured Variational Cross-Graph Correspondence Learning](https://mlsanity.com/paper/2203.13049/) (2022)
- [Language-Conditioned Graph Networks for Relational Reasoning](https://mlsanity.com/paper/1905.04405/) (2019)
- [Meta Module Network for Compositional Visual Reasoning](https://mlsanity.com/paper/1910.03230/) (2019)
- [On Compositional Generalization of Neural Machine Translation](https://mlsanity.com/paper/2105.14802/) (2021)
- [Coarse-to-Fine Reasoning for Visual Question Answering](https://mlsanity.com/paper/2110.02526/) (2021)
- [Context-based and Diversity-driven Specificity in Compositional Zero-Shot Learning](https://mlsanity.com/paper/2402.17251/) (2024)
- [Variational Cross-Graph Reasoning and Adaptive Structured Semantics Learning for Compositional Temporal Grounding](https://mlsanity.com/paper/2301.09071/) (2023)
- [Compositional Generalization for Multi-label Text Classification: A Data-Augmentation Approach](https://mlsanity.com/paper/2312.11276/) (2023)
- [Transformer Module Networks for Systematic Generalization in Visual Question Answering](https://mlsanity.com/paper/2201.11316/) (2022)
- [Overcoming Language Priors for Visual Question Answering Based on Knowledge Distillation](https://mlsanity.com/paper/2501.05690/) (2025)

## Abstract (arXiv)

Compositional generalization is the ability of generalizing novel compositions from seen primitives, and has received much attention in vision-and-language (V&L) recently. Due to the multi-modal nature of V&L tasks, the primitives composing compositions source from different modalities, resulting in multi-sourced novel compositions. However, the generalization ability over multi-sourced novel compositions, textiti.e., multi-sourced compositional generalization (MSCG) remains unexplored. In this paper, we explore MSCG in the context of visual question answering (VQA), and propose a retrieval-augmented training framework to enhance the MSCG ability of VQA models by learning unified representations for primitives from different modalities. Specifically, semantically equivalent primitives are retrieved for each primitive in the training samples, and the retrieved features are aggregated with the original primitive to refine the model. This process helps the model learn consistent representations for the same semantic primitives across different modalities. To evaluate the MSCG ability of VQA models, we construct a new GQA-MSCG dataset based on the GQA dataset, in which samples include three types of novel compositions composed of primitives from different modalities. Experimental results demonstrate the effectiveness of the proposed framework. We release GQA-MSCG at https://github.com/NeverMoreLCH/MSCG.