One Self-Configurable Model to Solve Many Abstract Visual Reasoning Problems

Abstract

Abstract Visual Reasoning (AVR) comprises a wide selection of various problems similar to those used in human IQ tests. Recent years have brought dynamic progress in solving particular AVR tasks, however, in the contemporary literature AVR problems are largely dealt with in isolation, leading to highly specialized task-specific methods. With the aim of developing universal learning systems in the AVR domain, we propose the unified model for solving Single-Choice Abstract visual Reasoning tasks (SCAR), capable of solving various single-choice AVR tasks, without making any a priori assumptions about the task structure, in particular the number and location of panels. The proposed model relies on a novel Structure-Aware dynamic Layer (SAL), which adapts its weights to the structure of the considered AVR problem. Experiments conducted on Raven's Progressive Matrices, Visual Analogy Problems, and Odd One Out problems show that SCAR (SAL-based models, in general) effectively solves diverse AVR tasks, and its performance is on par with the state-of-the-art task-specific baselines. What is more, SCAR demonstrates effective knowledge reuse in multi-task and transfer learning settings. To our knowledge, this work is the first successful attempt to construct a general single-choice AVR solver relying on self-configurable architecture and unified solving method. With this work we aim to stimulate and foster progress on task-independent research paths in the AVR domain, with the long-term goal of development of a general AVR solver.

arXiv commentAccepted to The 38th Annual AAAI Conference on Artificial Intelligence (AAAI 2024)

References in corpus (18)

• Adam: A Method for Stochastic Optimization
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Gaussian Error Linear Units (GELUs)
• Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
• A simple neural network module for relational reasoning
• Dynamic Filter Networks
• Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
• Deformable ConvNets v2: More Deformable, Better Results
• Partial Transfer Learning with Selective Adversarial Networks
• LambdaNetworks: Modeling Long-Range Interactions Without Attention
• Learning to Make Analogies by Contrasting Abstract Relational Structure
• Multi-Label Contrastive Learning for Abstract Visual Reasoning
• Deformable Kernels: Adapting Effective Receptive Fields for Object Deformation
• RAVEN: A Dataset for Relational and Analogical Visual rEasoNing
• The Scattering Compositional Learner: Discovering Objects, Attributes, Relationships in Analogical Reasoning
• Bongard-LOGO: A New Benchmark for Human-Level Concept Learning and Reasoning
• A Review of Emerging Research Directions in Abstract Visual Reasoning
• Effective Abstract Reasoning with Dual-Contrast Network