Abstract:Recent years have witnessed the success of deep neural networks in many research areas. The fundamental idea behind the design of most neural networks is to learn similarity patterns from data for prediction and inference, which lacks the ability of cognitive reasoning. However, the concrete ability of reasoning is critical to many theoretical and practical problems. On the other hand, traditional symbolic reasoning methods do well in making logical inference, but they are mostly hard rule-based reasoning, which limits their generalization ability to different tasks since difference tasks may require different rules. Both reasoning and generalization ability are important for prediction tasks such as recommender systems, where reasoning provides strong connection between user history and target items for accurate prediction, and generalization helps the model to draw a robust user portrait over noisy inputs. In this paper, we propose Logic-Integrated Neural Network (LINN) to integrate the power of deep learning and logic reasoning. LINN is a dynamic neural architecture that builds the computational graph according to input logical expressions. It learns basic logical operations such as AND, OR, NOT as neural modules, and conducts propositional logical reasoning through the network for inference. Experiments on theoretical task show that LINN achieves significant performance on solving logical equations and variables. Furthermore, we test our approach on the practical task of recommendation by formulating the task into a logical inference problem. Experiments show that LINN significantly outperforms state-of-the-art recommendation models in Top-K recommendation, which verifies the potential of LINN in practice.

Realization of Neural Networks for Basic Operation on Multi-Valued Logic

Multi-thresholded neuron and it's application in multi-valued logic

Multi-Valued Logic Learning Control

A SYMMETRIC TERNARY LOGIC BASED MULTI-VALUED LEARNING NETWORK

Algebraic properties of a learning multiple-valued logic network

Learning Algorithms via Neural Logic Networks

Extracting Formulae in Many-Valued Logic from Deep Neural Networks

Binary Higher Order Neural Networks for Realizing Boolean Functions.

Interpretable neural networks based on continuous-valued logic and multicriteria decision operators

Looped ReLU MLPs May Be All You Need as Practical Programmable Computers

Learning Interpretable Differentiable Logic Networks

Reinforcement Learning with Markov Logic Networks

Logic Computing with Stateful Neural Networks of Resistive Switches

Perceptron Implementation of Triple-Valued Logic Operations

Local design for multi-layer perceptron

Neural Arithmetic Logic Units with Two Transition Matrix and Independent Gates

A Universal Logic Operator for Interpretable Deep Convolution Networks

Cellular automata, many-valued logic, and deep neural networks

Neural Logic Reasoning

Towards a Unifying Logical Framework for Neural Networks

A preferential interpretation of MultiLayer Perceptrons in a conditional logic with typicality