Abstract:Patent classification aims to assign multiple International Patent Classification (IPC) codes to a given patent. Existing methods for automated patent classification primarily focus on analyzing the text descriptions of patents. However, apart from the textual information, each patent is also associated with some assignees, and the knowledge of their previously applied patents can often be valuable for accurate classification. Furthermore, the hierarchical taxonomy defined by the IPC system provides crucial contextual information and enables models to leverage the correlations between IPC codes for improved classification accuracy. However, existing methods fail to incorporate the above aspects and lead to reduced performance. To address these limitations, we propose an integrated framework that comprehensively considers patent-related information for patent classification. To be specific, we first present an IPC codes correlations learning module to capture both horizontal and vertical information within the IPC codes. This module effectively captures the correlations by adaptively exchanging and aggregating messages among IPC codes at the same level (horizontal information) and from both parent and children codes (vertical information), which allows for a comprehensive integration of knowledge and relationships within the IPC hierarchical taxonomy. Additionally, we design a historical application patterns learning component to incorporate previous patents of the corresponding assignee by aggregating high-order temporal information via a dual-channel graph neural network. Finally, our approach combines the contextual information from patent texts, which encompasses the semantics of IPC codes, with assignees’ sequential preferences to make predictions. Experimental evaluations on real-world datasets demonstrate the superiority of our proposed approach over existing methods. Moreover, we present the model’s ability to capture the temporal patterns of assignees and the semantic dependencies among IPC codes.

Support Vector Based Method for Acquiring Domain Specific Patents

Exploiting Semantic Knowledge Base for Patent Retrieval

Domain Lexicon-Based Query Expansion for Patent Retrieval

A Semantic Query Expansion-Based Patent Retrieval Approach

An Ontology-Based Automatic Semantic Annotation Approach for Patent Document Retrieval in Product Innovation Design

An automatic classification method for patents

DeepPatent: patent classification with convolutional neural networks and word embedding

A Patent Keyword Extraction Method Based on Corpus Classification

A patent retrieval method based on automatic query expansion

Automatic Abstraction of Long Chinese Patent Texts Based on P-Bertsum Model

Method of Discovering Similar Patents Based on Vector Space Model and Characteristics of Patent Documents

Enhancing patent text classification with Bi-LSTM technique and alpine skiing optimization for improved diagnostic accuracy

Adaptive Taxonomy Learning and Historical Patterns Modelling for Patent Classification

Towards Accurate Word Segmentation for Chinese Patents

An Automatic Generation Method of Patent Specification Abstract Based on "Extraction- Abstraction "Model

Research on the Classification and Identification Method of Fintech Patents

Information and Strategies Supplied Patent Management Platform Based on Network

An Improved P-SVM Model for Patent Application Filings Prediction

Multi label classification of Artificial Intelligence related patents using Modified D2SBERT and Sentence Attention mechanism

Reliable Multi-View Deep Patent Classification