Convolutional 2D Knowledge Graph Embeddings resources

Related tags

Deep LearningConvE
Overview

ConvE

Convolutional 2D Knowledge Graph Embeddings resources.

Paper: Convolutional 2D Knowledge Graph Embeddings

Used in the paper, but do not use these datasets for your research: FB15k and WN18. Please also note that the Kinship and Nations datasets have a high number of inverse relationships which makes them unsuitable for research. Nations has +95% inverse relationships and Kinship about 48%.

ConvE key facts

Predictive performance

Dataset MR MRR [email protected] [email protected] [email protected]
FB15k 64 0.75 0.87 0.80 0.67
WN18 504 0.94 0.96 0.95 0.94
FB15k-237 246 0.32 0.49 0.35 0.24
WN18RR 4766 0.43 0.51 0.44 0.39
YAGO3-10 2792 0.52 0.66 0.56 0.45
Nations 2 0.82 1.00 0.88 0.72
UMLS 1 0.94 0.99 0.97 0.92
Kinship 2 0.83 0.98 0.91 0.73

Run time performance

For an embedding size of 200 and batch size 128, a single batch takes on a GTX Titan X (Maxwell):

  • 64ms for 100,000 entities
  • 80ms for 1,000,000 entities

Parameter efficiency

Parameters ConvE/DistMult MRR ConvE/DistMult [email protected] ConvE/DistMult [email protected]
~5.0M 0.32 / 0.24 0.49 / 0.42 0.24 / 0.16
1.89M 0.32 / 0.23 0.49 / 0.41 0.23 / 0.15
0.95M 0.30 / 0.22 0.46 / 0.39 0.22 / 0.14
0.24M 0.26 / 0.16 0.39 / 0.31 0.19 / 0.09

ConvE with 8 times less parameters is still more powerful than DistMult. Relational Graph Convolutional Networks use roughly 32x more parameters to have the same performance as ConvE.

Installation

This repo supports Linux and Python installation via Anaconda.

  1. Install PyTorch using Anaconda.
  2. Install the requirements pip install -r requirements.txt
  3. Download the default English model used by spaCy, which is installed in the previous step python -m spacy download en
  4. Run the preprocessing script for WN18RR, FB15k-237, YAGO3-10, UMLS, Kinship, and Nations: sh preprocess.sh
  5. You can now run the model

Running a model

Parameters need to be specified by white-space tuples for example:

CUDA_VISIBLE_DEVICES=0 python main.py --model conve --data FB15k-237 \
                                      --input-drop 0.2 --hidden-drop 0.3 --feat-drop 0.2 \
                                      --lr 0.003 --preprocess

will run a ConvE model on FB15k-237.

To run a model, you first need to preprocess the data once. This can be done by specifying the --preprocess parameter:

CUDA_VISIBLE_DEVICES=0 python main.py --data DATASET_NAME --preprocess

After the dataset is preprocessed it will be saved to disk and this parameter can be omitted.

CUDA_VISIBLE_DEVICES=0 python main.py --data DATASET_NAME

The following parameters can be used for the --model parameter:

conve
distmult
complex

The following datasets can be used for the --data parameter:

FB15k-237
WN18RR
YAGO3-10
umls
kinship
nations

And here a complete list of parameters.

Link prediction for knowledge graphs

optional arguments:
  -h, --help            show this help message and exit
  --batch-size BATCH_SIZE
                        input batch size for training (default: 128)
  --test-batch-size TEST_BATCH_SIZE
                        input batch size for testing/validation (default: 128)
  --epochs EPOCHS       number of epochs to train (default: 1000)
  --lr LR               learning rate (default: 0.003)
  --seed S              random seed (default: 17)
  --log-interval LOG_INTERVAL
                        how many batches to wait before logging training
                        status
  --data DATA           Dataset to use: {FB15k-237, YAGO3-10, WN18RR, umls,
                        nations, kinship}, default: FB15k-237
  --l2 L2               Weight decay value to use in the optimizer. Default:
                        0.0
  --model MODEL         Choose from: {conve, distmult, complex}
  --embedding-dim EMBEDDING_DIM
                        The embedding dimension (1D). Default: 200
  --embedding-shape1 EMBEDDING_SHAPE1
                        The first dimension of the reshaped 2D embedding. The
                        second dimension is infered. Default: 20
  --hidden-drop HIDDEN_DROP
                        Dropout for the hidden layer. Default: 0.3.
  --input-drop INPUT_DROP
                        Dropout for the input embeddings. Default: 0.2.
  --feat-drop FEAT_DROP
                        Dropout for the convolutional features. Default: 0.2.
  --lr-decay LR_DECAY   Decay the learning rate by this factor every epoch.
                        Default: 0.995
  --loader-threads LOADER_THREADS
                        How many loader threads to use for the batch loaders.
                        Default: 4
  --preprocess          Preprocess the dataset. Needs to be executed only
                        once. Default: 4
  --resume              Resume a model.
  --use-bias            Use a bias in the convolutional layer. Default: True
  --label-smoothing LABEL_SMOOTHING
                        Label smoothing value to use. Default: 0.1
  --hidden-size HIDDEN_SIZE
                        The side of the hidden layer. The required size
                        changes with the size of the embeddings. Default: 9728
                        (embedding size 200).

To reproduce most of the results in the ConvE paper, you can use the default parameters and execute the command below:

CUDA_VISIBLE_DEVICES=0 python main.py --data DATASET_NAME

For the reverse model, you can run the provided file with the name of the dataset name and a threshold probability:

python inverse_model.py WN18RR 0.9

Changing the embedding size for ConvE

If you want to change the embedding size you can do that via the ``--embedding-dim parameter. However, for ConvE, since the embedding is reshaped as a 2D embedding one also needs to pass the first dimension of the reshaped embedding (--embedding-shape1`) while the second dimension is infered. When once changes the embedding size, the hidden layer size `--hidden-size` also needs to be different but it is difficult to determine before run time. The easiest way to determine the hidden size is to run the model, let it run on an error due to wrong shape, and then reshape according to the dimension in the error message.

Example: Change embedding size to be 100. We want 10x10 2D embeddings. We run python main.py --embedding-dim 100 --embedding-shape1 10 and we run on an error due to wrong hidden dimension:

   ret = torch.addmm(bias, input, weight.t())
RuntimeError: size mismatch, m1: [128 x 4608], m2: [9728 x 100] at /opt/conda/conda-bld/pytorch_1565272271120/work/aten/src/THC/generic/THCTensorMathBlas.cu:273

Now we change the hidden dimension to 4608 accordingly: python main.py --embedding-dim 100 --embedding-shape1 10 --hidden-size 4608. Now the model runs with an embedding size of 100 and 10x10 2D embeddings.

Adding new datasets

To run it on a new datasets, copy your dataset folder into the data folder and make sure your dataset split files have the name train.txt, valid.txt, and test.txt which contain tab separated triples of a knowledge graph. Then execute python wrangle_KG.py FOLDER_NAME, afterwards, you can use the folder name of your dataset in the dataset parameter.

Adding your own model

You can easily write your own knowledge graph model by extending the barebone model MyModel that can be found in the model.py file.

Quirks

There are some quirks of this framework.

  1. The model currently ignores data that does not fit into the specified batch size, for example if your batch size is 100 and your test data is 220, then 20 samples will be ignored. This is designed in that way to improve performance on small datasets. To test on the full test-data you can save the model checkpoint, load the model (with the --resume True variable) and then evaluate with a batch size that fits the test data (for 220 you could use a batch size of 110). Another solution is to just use a fitting batch size from the start, that is, you could train with a batch size of 110.

Issues

It has been noted that #6 WN18RR does contain 212 entities in the test set that do not appear in the training set. About 6.7% of the test set is affected. This means that most models will find it impossible to make any reasonable predictions for these entities. This will make WN18RR appear more difficult than it really is, but it should not affect the usefulness of the dataset. If all researchers compared to the same datasets the scores will still be comparable.

Logs

Some log files of the original research are included in the repo (logs.tar.gz). These log files are mostly unstructured in names and might be created from checkpoints so that it is difficult to comprehend them. Nevertheless, it might help to replicate the results or study the behavior of the training under certain conditions and thus I included them here.

Citation

If you found this codebase or our work useful please cite us:

@inproceedings{dettmers2018conve,
	Author = {Dettmers, Tim and Pasquale, Minervini and Pontus, Stenetorp and Riedel, Sebastian},
	Booktitle = {Proceedings of the 32th AAAI Conference on Artificial Intelligence},
	Title = {Convolutional 2D Knowledge Graph Embeddings},
	Url = {https://arxiv.org/abs/1707.01476},
	Year = {2018},
        pages  = {1811--1818},
  	Month = {February}
}



Owner
Tim Dettmers
Tim Dettmers
Unsupervised Video Interpolation using Cycle Consistency

Unsupervised Video Interpolation using Cycle Consistency Project | Paper | YouTube Unsupervised Video Interpolation using Cycle Consistency Fitsum A.

NVIDIA Corporation 100 Nov 30, 2022
This project provides the code and datasets for 'CapSal: Leveraging Captioning to Boost Semantics for Salient Object Detection', CVPR 2019.

Code-and-Dataset-for-CapSal This project provides the code and datasets for 'CapSal: Leveraging Captioning to Boost Semantics for Salient Object Detec

lu zhang 48 Aug 19, 2022
OpenMMLab Computer Vision Foundation

English | 简体中文 Introduction MMCV is a foundational library for computer vision research and supports many research projects as below: MMCV: OpenMMLab

OpenMMLab 4.6k Jan 09, 2023
Efficient Two-Step Networks for Temporal Action Segmentation (Neurocomputing 2021)

Efficient Two-Step Networks for Temporal Action Segmentation This repository provides a PyTorch implementation of the paper Efficient Two-Step Network

8 Apr 16, 2022
this is a lite easy to use virtual keyboard project for anyone to use

virtual_Keyboard this is a lite easy to use virtual keyboard project for anyone to use motivation I made this for this year's recruitment for RobEn AA

Mohamed Emad 3 Oct 23, 2021
Official Implementation (PyTorch) of "Point Cloud Augmentation with Weighted Local Transformations", ICCV 2021

PointWOLF: Point Cloud Augmentation with Weighted Local Transformations This repository is the implementation of PointWOLF(To appear). Sihyeon Kim1*,

MLV Lab (Machine Learning and Vision Lab at Korea University) 16 Nov 03, 2022
Visualizer using audio and semantic analysis to explore BigGAN (Brock et al., 2018) latent space.

BigGAN Audio Visualizer Description This visualizer explores BigGAN (Brock et al., 2018) latent space by using pitch/tempo of an audio file to generat

Rush Kapoor 2 Nov 21, 2022
PyTorch Implementation for Fracture Detection in Wrist Bone X-ray Images

wrist-d PyTorch Implementation for Fracture Detection in Wrist Bone X-ray Images note: Paper: Under Review at MPDI Diagnostics Submission Date: Novemb

Fatih UYSAL 5 Oct 12, 2022
A Keras implementation of YOLOv3 (Tensorflow backend)

keras-yolo3 Introduction A Keras implementation of YOLOv3 (Tensorflow backend) inspired by allanzelener/YAD2K. Quick Start Download YOLOv3 weights fro

7.1k Jan 03, 2023
Systematic generalisation with group invariant predictions

Requirements are Python 3, TensorFlow v1.14, Numpy, Scipy, Scikit-Learn, Matplotlib, Pillow, Scikit-Image, h5py, tqdm. Experiments were run on V100 GPUs (16 and 32GB).

Faruk Ahmed 30 Dec 01, 2022
Code for Robust Contrastive Learning against Noisy Views

Robust Contrastive Learning against Noisy Views This repository provides a PyTorch implementation of the Robust InfoNCE loss proposed in paper Robust

Ching-Yao Chuang 53 Jan 08, 2023
This repository is the official implementation of Unleashing the Power of Contrastive Self-Supervised Visual Models via Contrast-Regularized Fine-Tuning (NeurIPS21).

Core-tuning This repository is the official implementation of ``Unleashing the Power of Contrastive Self-Supervised Visual Models via Contrast-Regular

vanint 18 Dec 17, 2022
Evolution Strategies in PyTorch

Evolution Strategies This is a PyTorch implementation of Evolution Strategies. Requirements Python 3.5, PyTorch = 0.2.0, numpy, gym, universe, cv2 Wh

Andrew Gambardella 333 Nov 14, 2022
deep learning model with only python and numpy with test accuracy 99 % on mnist dataset and different optimization choices

deep_nn_model_with_only_python_100%_test_accuracy deep learning model with only python and numpy with test accuracy 99 % on mnist dataset and differen

0 Aug 28, 2022
Official implementation of "DSP: Dual Soft-Paste for Unsupervised Domain Adaptive Semantic Segmentation"

DSP Official implementation of "DSP: Dual Soft-Paste for Unsupervised Domain Adaptive Semantic Segmentation". Accepted by ACM Multimedia 2021. Authors

20 Oct 24, 2022
[SIGGRAPH Asia 2019] Artistic Glyph Image Synthesis via One-Stage Few-Shot Learning

AGIS-Net Introduction This is the official PyTorch implementation of the Artistic Glyph Image Synthesis via One-Stage Few-Shot Learning. paper | suppl

Yue Gao 102 Jan 02, 2023
ATOMIC 2020: On Symbolic and Neural Commonsense Knowledge Graphs

(Comet-) ATOMIC 2020: On Symbolic and Neural Commonsense Knowledge Graphs Paper Jena D. Hwang, Chandra Bhagavatula, Ronan Le Bras, Jeff Da, Keisuke Sa

AI2 152 Dec 27, 2022
[CVPR2021] Invertible Image Signal Processing

Invertible Image Signal Processing This repository includes official codes for "Invertible Image Signal Processing (CVPR2021)". Figure: Our framework

Yazhou XING 281 Dec 31, 2022
Direct design of biquad filter cascades with deep learning by sampling random polynomials.

IIRNet Direct design of biquad filter cascades with deep learning by sampling random polynomials. Usage git clone https://github.com/csteinmetz1/IIRNe

Christian J. Steinmetz 55 Nov 02, 2022
ROSITA: Enhancing Vision-and-Language Semantic Alignments via Cross- and Intra-modal Knowledge Integration

ROSITA News & Updates (24/08/2021) Release the demo to perform fine-grained semantic alignments using the pretrained ROSITA model. (15/08/2021) Releas

Vision and Language Group@ MIL 48 Dec 23, 2022