Explicit Knowledge Learning

Explicit knowledge learning focuses on developing machine learning models that not only achieve high performance but also explicitly represent and utilize learned knowledge in a transparent and interpretable manner. Current research emphasizes enhancing existing models, such as collaborative filtering and neural networks, through techniques like embedding extraction, vector field representations, and knowledge graph construction, to improve both accuracy and explainability. This approach is significant because it addresses the "black box" nature of many deep learning models, paving the way for more trustworthy and understandable AI systems across diverse applications, including recommendation systems, 3D reconstruction, and mathematical reasoning.

Papers

May 31, 2024

ImplicitSLIM and How it Improves Embedding-based Collaborative Filtering
Ilya Shenbin, Sergey Nikolenko
High Dimensional Collaborative Filtering Implicit Model Traditional Recommender Embedding Based Explicit Knowledge Learning

March 8, 2023

Neural Vector Fields: Implicit Representation by Explicit Learning
Xianghui Yang, Guosheng Lin, Zhenghao Chen, Luping Zhou
Implicit Representation Vector Field 3D Shape Representation 3D Reconstruction Task Dual Domain Reconstruction Explicit Knowledge Learning

February 11, 2023

Learning by Applying: A General Framework for Mathematical Reasoning via Enhancing Explicit Knowledge Learning
Jiayu Liu, Zhenya Huang, Chengxiang Zhai, Qi Liu
LeArning Abstract Mathematical Reasoning General Framework Knowledge Representation Reasoning Module Explicit Knowledge Learning

October 25, 2022

Learning Explicit Object-Centric Representations with Vision Transformers
Oscar Vikström, Alexander Ilin
Vision Transformer Object Segmentation Object Centric Representation Based Method Explicit Knowledge Learning

January 17, 2022

Differentiable Rule Induction with Learned Relational Features
Remy Kusters, Yusik Kim, Marine Collery, Christian de Sainte Marie, Shubham Gupta
First Order Interpretable Rule Rule Induction Explicit Knowledge Learning

Explicit Knowledge Learning

Papers

ImplicitSLIM and How it Improves Embedding-based Collaborative Filtering

Neural Vector Fields: Implicit Representation by Explicit Learning

Learning by Applying: A General Framework for Mathematical Reasoning via Enhancing Explicit Knowledge Learning

Learning Explicit Object-Centric Representations with Vision Transformers

Differentiable Rule Induction with Learned Relational Features