Learning Visual Abstract Reasoning through Dual-Stream Networks

Abstract

Visual abstract reasoning tasks present challenges for deep neural networks, exposing limitations in their capabilities. In this work, we present a neural network model that addresses the challenges posed by Raven's Progressive Matrices (RPM). Inspired by the two-stream hypothesis of visual processing, we introduce the Dual-stream Reasoning Network (DRNet), which utilizes two parallel branches to capture image features. On top of the two streams, a reasoning module first learns to merge the high-level features of the same image. Then, it employs a rule extractor to handle combinations involving the eight context images and each candidate image, extracting discrete abstract rules and utilizing an multilayer perceptron (MLP) to make predictions. Empirical results demonstrate that the proposed DRNet achieves state-of-the-art average performance across multiple RPM benchmarks. Furthermore, DRNet demonstrates robust generalization capabilities, even extending to various out-of-distribution scenarios. The dual streams within DRNet serve distinct functions by addressing local or spatial information. They are then integrated into the reasoning module, leveraging abstract rules to facilitate the execution of visual reasoning tasks. These findings indicate that the dual-stream architecture could play a crucial role in visual abstract reasoning.

10 pages, 6 figures

References in corpus (9)

• Two-Stream Convolutional Networks for Action Recognition in Videos
• A simple neural network module for relational reasoning
• Language Is Not All You Need: Aligning Perception with Language Models
• Two-Stream Network for Sign Language Recognition and Translation
• RAVEN: A Dataset for Relational and Analogical Visual rEasoNing
• The Scattering Compositional Learner: Discovering Objects, Attributes, Relationships in Analogical Reasoning
• Effective Abstract Reasoning with Dual-Contrast Network
• Dynamic Inference with Neural Interpreters
• Learning to reason over visual objects