• Home
  • Search
  • AlphaILP: thinking visual scenes as differentiable logic programs
  • Cite Icon18
  • https://doi.org/10.1007/s10994-023-06320-1Copy DOI Icon

AlphaILP: thinking visual scenes as differentiable logic programs

Show More
  • Abstract
  • PDF
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Deep neural learning has shown remarkable performance at learning representations for visual object categorization. However, deep neural networks such as CNNs do not explicitly encode objects and relations among them. This limits their success on tasks that require a deep logical understanding of visual scenes, such as Kandinsky patterns and Bongard problems. To overcome these limitations, we introduce alpha {textit{ILP}}, a novel differentiable inductive logic programming framework that learns to represent scenes as logic programs—intuitively, logical atoms correspond to objects, attributes, and relations, and clauses encode high-level scene information. alphaILP has an end-to-end reasoning architecture from visual inputs. Using it, alphaILP performs differentiable inductive logic programming on complex visual scenes, i.e., the logical rules are learned by gradient descent. Our extensive experiments on Kandinsky patterns and CLEVR-Hans benchmarks demonstrate the accuracy and efficiency of alpha {textit{ILP}} in learning complex visual-logical concepts.

Loading PDF

Similar Papers
  • PDF
  • Research Article
  • Citations1

Neuro-symbolic Predicate Invention: Learning relational concepts from visual scenes

  • Mar 01, 2025
  • Neurosymbolic Artificial Intelligence
  • Jingyuan Sha +3
  • PDF
  • Research Article
  • Citations65

Representation of visual scenes by local neuronal populations in layer 2/3 of mouse visual cortex

  • Dec 12, 2011
  • Frontiers in Neural Circuits
  • Björn M Kampa +3
  • Research Article
  • Citations4

Artificial Intelligence and Deep Learning Applications: A Review

  • Oct 04, 2021
  • Journal of Environmental Impact and Management Policy
  • G. Ramachandran +1
  • Research Article
  • Citations67

Reconstruction of natural visual scenes from neural spikes with deep neural networks

  • Feb 08, 2020
  • Neural Networks
  • Yichen Zhang +7
  • Research Article
  • Citations23

Effects of hemoencephalographic biofeedback with virtual reality on selected aspects of attention in children with ADHD

  • Oct 13, 2021
  • International Journal of Psychophysiology
  • Sebastian Skalski +4
  • Research Article
  • Citations26

The role of view in human face detection

  • May 21, 2009
  • Vision Research
  • A Mike Burton +1
  • Research Article
  • Citations5

Zero-shot counting with a dual-stream neural network model.

  • Dec 01, 2024
  • Neuron
  • Jessica A F Thompson +5
  • Book Chapter
  • Citations1

Handwritten Digit Recognition Using Very Deep Convolutional Neural Network

  • Jan 01, 2022
  • M Dhilsath Fathima +2
  • PDF
  • Conference Article
  • Citations26

Deep Weighted MaxSAT for Aspect-based Opinion Extraction

  • Jan 01, 2020
  • Meixi Wu +2
  • Research Article
  • Citations13

Performance comparison of different momentum techniques on deep reinforcement learning

  • Feb 23, 2018
  • Journal of Information and Telecommunication
  • M Sarigül +1
  • Research Article
  • Citations9

ScieNet: Deep learning with spike-assisted contextual information extraction

  • May 01, 2021
  • Pattern Recognition
  • Xueyuan She +3
  • Conference Article
  • Citations6

FADEC: FPGA-based Acceleration of Video Depth Estimation by HW/SW Co-design

  • Dec 05, 2022
  • Nobuho Hashimoto +1
  • Supplementary Content

Nonparametric Perspective of Deep Learning

  • Dec 10, 2020
  • Figshare
  • Tianyang Hu
  • Research Article
  • Citations9

Information Bottleneck and its Applications in Deep Learning

  • Dec 29, 2018
  • حسن حافظ کلاهی +1
  • Research Article
  • Citations6

Break through the limits of learning by machines

  • Sep 20, 2016
  • Chinese Science Bulletin
  • Zhongzhi Shi
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.