metadata
license: cc-by-nc-4.0
language:
- en
base_model:
- anferico/bert-for-patents
PatentMap V0 Models
This repository serves as the main entry point for all models released as part of the PatentMapV0 project.
It does not contain model weights itself. Instead, it provides links and documentation for each individual model hosted on Hugging Face.
🤖 Available Models
🔹 Augmentation Method Variants (Abstract-Only)
Models trained on abstract text with different data augmentation strategies:
- PatentMap-V0-Dropout - Baseline with dropout augmentation
- PatentMap-V0-IPC8 - IPC8 classification-based positive pairs
- PatentMap-V0-Paraphrase - Paraphrasing augmentation
- PatentMap-V0-RandCrop - Random text cropping
- PatentMap-V0-SentShuffle - Sentence shuffling
- PatentMap-V0-CropShuffle - Combined cropping and shuffling
🔹 Section-Pair Variants
Models trained with section-pair augmentation (abstract + additional sections):
Single Section Additions:
- PatentMap-V0-SecPair-Claim - Abstract + Claims
- PatentMap-V0-SecPair-Summary - Abstract + Summary
- PatentMap-V0-SecPair-Background - Abstract + Background
- PatentMap-V0-SecPair-Drawing - Abstract + Drawing descriptions
- PatentMap-V0-SecPair-Description - Abstract + Detailed description
Two Section Additions:
- PatentMap-V0-SecPair-ClaimSummary - Abstract + Claims + Summary
- PatentMap-V0-SecPair-ClaimBackground - Abstract + Claims + Background
- PatentMap-V0-SecPair-ClaimDrawing - Abstract + Claims + Drawing
- PatentMap-V0-SecPair-ClaimDescription - Abstract + Claims + Description
- PatentMap-V0-SecPair-SummaryBackground - Abstract + Summary + Background
- PatentMap-V0-SecPair-SummaryDrawing - Abstract + Summary + Drawing
- PatentMap-V0-SecPair-SummaryDescription - Abstract + Summary + Description
- PatentMap-V0-SecPair-BackgroundDrawing - Abstract + Background + Drawing
- PatentMap-V0-SecPair-BackgroundDescription - Abstract + Background + Description
- PatentMap-V0-SecPair-DrawingDescription - Abstract + Drawing + Description
🧪 Usage
Each model repository contains:
- Model weights
- Configuration
- Usage examples
- Training details
Please refer to the individual model pages for loading instructions.
📜 License
All models are released under CC BY-NC 4.0 licenses.
📖 Citation
If you use any of these models, please cite the following paper:
@article{zuo2025patent,
title={Patent Representation Learning via Self-supervision},
author={Zuo, You and Gerdes, Kim and de La Clergerie, Eric Villemonte and Sagot, Beno{\^\i}t},
journal={arXiv preprint arXiv:2511.10657},
year={2025}
}