An implementation for `Text2Event: Controllable Sequence-to-Structure Generation for End-to-end Event Extraction`

Overview

Text2Event

Update

  • [2021-08-03] Update pre-trained models

Quick links

Requirements

General

  • Python (verified on 3.8)
  • CUDA (verified on 11.1)

Python Packages

  • see requirements.txt
conda create -n text2event python=3.8
conda activate text2event
pip install -r requirements.txt

Quick Start

Data Format

Data folder contains four files:

data/text2tree/one_ie_ace2005_subtype
├── event.schema
├── test.json
├── train.json
└── val.json

train/val/test.json are data files, and each line is a JSON instance. Each JSON instance contains text and event fields, in which text is plain text, and event is event linearized form. If you want to use other key names, it is easy to change the input format in run_seq2seq.py.

{"text": "He also owns a television and a radio station and a newspaper .", "event": "<extra_id_0>  <extra_id_1>"}
{"text": "' ' For us the United Natgions is the key authority '' in resolving the Iraq crisis , Fischer told reporters opn arrival at the EU meeting .", "event": "<extra_id_0> <extra_id_0> Meet meeting <extra_id_0> Entity EU <extra_id_1> <extra_id_1> <extra_id_1>"}

Note:

  • Use the extra character of T5 as the structure indicators, such as <extra_id_0>, <extra_id_1>, etc.

  • event.schema is the event schema file for building the trie of constrained decoding. It contains three lines: the first line is event type name list, the second line is event role name list, the third line is type-to-role dictionary.

    ["Declare-Bankruptcy", "Convict", ...]
    ["Plaintiff", "Target", ...]
    {"End-Position": ["Place", "Person", "Entity"], ...}
    

Model Training

Training scripts as follows:

  • run_seq2seq.py: Python code entry, modified from the transformers/examples/seq2seq/run_seq2seq.py
  • run_seq2seq.bash: Model training script logging to the log file.
  • run_seq2seq_verbose.bash: Same model training script as run_seq2seq.bash but output to the screen directly.
  • run_seq2seq_with_pretrain.bash: Model training script for curriculum learning, which contains substructure learning and full structure learning.

The command for the training is as follows (see bash scripts and Python files for the corresponding command-line arguments):

bash run_seq2seq_verbose.bash -d 0 -f tree -m t5-base --label_smoothing 0 -l 1e-4 --lr_scheduler linear --warmup_steps 2000 -b 16
  • -d refers to the GPU device id.
  • -m t5-base refers to using T5-base.
  • Currently, constrained decoding algorithms do not support use_fast_tokenizer=True and beam search yet.

Trained models are saved in the models/ folder.

Model Evaluation

Offset-level Evaluation

python evaluation.py -g <data-folder-path> -r <offset-folder-path> -p <model-folder-path> -f <data-format>
  • This evaluation script converts the eval_preds_seq2seq.txt and test_preds_seq2seq.txt in the model folder <model-folder-path> into the corresponding offset prediction results for model evaluation.
  • -f <data-format> refers to dyiepp or oneie

Record-level Evaluation (approximate, used in training)

bash run_eval.bash -d 0 -m <model-folder-path> -i <data-folder-path> -c -b 8
  • -d refers to the GPU device id.
  • -c represents the use of constrained decoding, otherwise not apply
  • -b 8 represents batch_size=8

How to expand to other tasks

  1. prepare the corresponding data format
  2. Writ the code for reading corresponding data format: elif data_args.task.startswith("event") in seq2seq.py
  3. Writ the code for evaluating the corresponding task result: def compute_metrics(eval_preds) in seq2seq.py

Completing the above process can finish the simple Seq2Seq training and inference process.

If you need to use constrained decoding, you need to write the corresponding decoding mode (decoding_format), refer to extraction.extract_constraint.get_constraint_decoder

Pre-trained Model

You can find the pre-trained models as following google drive links or download models using command gdown (pip install gdown).

dyiepp_ace2005_en_t5_base.zip

gdown --id 1_fOmnSatNfceL9DZPxpof5AT9Oo7vTrC && unzip dyiepp_ace2005_en_t5_base.zip

dyiepp_ace2005_en_t5_large.zip

gdown --id 10iY1obkbgJtTKwfoOFevqL5AwG-hLvhU && unzip dyiepp_ace2005_en_t5_large.zip

oneie_ace2005_en_t5_large.zip

gdown --id 1zwnptRbdZntPT4ucqSANeaJ3vvwKliUe && unzip oneie_ace2005_en_t5_large.zip

oneie_ere_en_t5_large.zip

gdown --id 1WG7-pTZ3K49VMbQIONaDq_0pUXAcoXrZ && unzip oneie_ere_en_t5_large.zip

Event Datasets Preprocessing

We first refer to the following code and environments [dygiepp] and [oneie v0.4.7] for data preprocessing. Thanks to them!

After data preprocessing and we get the following data files:

 $ tree data/raw_data/
data/raw_data/
├── ace05-EN
│   ├── dev.oneie.json
│   ├── test.oneie.json
│   └── train.oneie.json
├── dyiepp_ace2005
│   ├── dev.json
│   ├── test.json
│   └── train.json
└── ERE-EN
    ├── dev.oneie.json
    ├── test.oneie.json
    └── train.oneie.json

We then convert the above data files to tree format. The following scripts generate the corresponding data folder in data/text2tree. The conversion will automatically generate train/dev/test JSON files and event.schema file.

bash scripts/processing_data.bash
data/text2tree
├── dyiepp_ace2005_subtype
│   ├── event.schema
│   ├── test.json
│   ├── train.json
│   └── val.json
├── dyiepp_ace2005_subtype_span
│   ├── event.schema
│   ├── test.json
│   ├── train.json
│   └── val.json
├── one_ie_ace2005_subtype
│   ├── event.schema
│   ├── test.json
│   ├── train.json
│   └── val.json
├── one_ie_ace2005_subtype_span
│   ├── event.schema
│   ├── test.json
│   ├── train.json
│   └── val.json
├── one_ie_ere_en_subtype
│   ├── event.schema
│   ├── test.json
│   ├── train.json
│   └── val.json
└── one_ie_ere_en_subtype_span
    ├── event.schema
    ├── test.json
    ├── train.json
    └── val.json
  • dyiepp_ace2005_subtype for Full Structure Learning and dyiepp_ace2005_subtype_span for Substructure Learning.

Citation

If this repository helps you, please cite this paper:

Yaojie Lu, Hongyu Lin, Jin Xu, Xianpei Han, Jialong Tang, Annan Li, Le Sun, Meng Liao, Shaoyi Chen. Text2Event: Controllable Sequence-to-Structure Generation for End-to-end Event Extraction. The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021).

@inproceedings{lu-etal-2021-text2event,
    title = "{T}ext2{E}vent: Controllable Sequence-to-Structure Generation for End-to-end Event Extraction",
    author = "Lu, Yaojie  and
      Lin, Hongyu  and
      Xu, Jin  and
      Han, Xianpei  and
      Tang, Jialong  and
      Li, Annan  and
      Sun, Le  and
      Liao, Meng  and
      Chen, Shaoyi",
    booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.acl-long.217",
    pages = "2795--2806",
    abstract = "Event extraction is challenging due to the complex structure of event records and the semantic gap between text and event. Traditional methods usually extract event records by decomposing the complex structure prediction task into multiple subtasks. In this paper, we propose Text2Event, a sequence-to-structure generation paradigm that can directly extract events from the text in an end-to-end manner. Specifically, we design a sequence-to-structure network for unified event extraction, a constrained decoding algorithm for event knowledge injection during inference, and a curriculum learning algorithm for efficient model learning. Experimental results show that, by uniformly modeling all tasks in a single model and universally predicting different labels, our method can achieve competitive performance using only record-level annotations in both supervised learning and transfer learning settings.",
}
Owner
Roger
Roger
A Broad Study on the Transferability of Visual Representations with Contrastive Learning

A Broad Study on the Transferability of Visual Representations with Contrastive Learning This repository contains code for the paper: A Broad Study on

Ashraful Islam 29 Nov 09, 2022
FAMIE is a comprehensive and efficient active learning (AL) toolkit for multilingual information extraction (IE)

FAMIE: A Fast Active Learning Framework for Multilingual Information Extraction

18 Sep 01, 2022
Code implementation of Data Efficient Stagewise Knowledge Distillation paper.

Data Efficient Stagewise Knowledge Distillation Table of Contents Data Efficient Stagewise Knowledge Distillation Table of Contents Requirements Image

IvLabs 112 Dec 02, 2022
TOOD: Task-aligned One-stage Object Detection, ICCV2021 Oral

One-stage object detection is commonly implemented by optimizing two sub-tasks: object classification and localization, using heads with two parallel branches, which might lead to a certain level of

264 Jan 09, 2023
An LSTM for time-series classification

Update 10-April-2017 And now it works with Python3 and Tensorflow 1.1.0 Update 02-Jan-2017 I updated this repo. Now it works with Tensorflow 0.12. In

Rob Romijnders 391 Dec 27, 2022
This initial strategy was developed specifically for larger pools and is based on taking a moving average and deriving Bollinger Bands to create a projected active liquidity range.

Gamma's Strategy One This initial strategy was developed specifically for larger pools and is based on taking a moving average and deriving Bollinger

Gamma Strategies 46 Dec 02, 2022
[ICML 2022] The official implementation of Graph Stochastic Attention (GSAT).

Graph Stochastic Attention (GSAT) The official implementation of GSAT for our paper: Interpretable and Generalizable Graph Learning via Stochastic Att

85 Nov 27, 2022
No-Reference Image Quality Assessment via Transformers, Relative Ranking, and Self-Consistency

This repository contains the implementation for the paper: No-Reference Image Quality Assessment via Transformers, Relative Ranking, and Self-Consiste

Alireza Golestaneh 75 Dec 30, 2022
Code for: https://berkeleyautomation.github.io/bags/

DeformableRavens Code for the paper Learning to Rearrange Deformable Cables, Fabrics, and Bags with Goal-Conditioned Transporter Networks. Here is the

Daniel Seita 121 Dec 30, 2022
🔎 Monitor deep learning model training and hardware usage from your mobile phone 📱

Monitor deep learning model training and hardware usage from mobile. 🔥 Features Monitor running experiments from mobile phone (or laptop) Monitor har

labml.ai 1.2k Dec 25, 2022
Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20. model in Tensorflow Lite.

TFLite-msg_chn_wacv20-depth-completion Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20. model

Ibai Gorordo 2 Oct 04, 2021
Geometry-Free View Synthesis: Transformers and no 3D Priors

Geometry-Free View Synthesis: Transformers and no 3D Priors Geometry-Free View Synthesis: Transformers and no 3D Priors Robin Rombach*, Patrick Esser*

CompVis Heidelberg 293 Dec 22, 2022
Open source annotation tool for machine learning practitioners.

doccano doccano is an open source text annotation tool for humans. It provides annotation features for text classification, sequence labeling and sequ

7.1k Jan 01, 2023
Deep Learning for Human Part Discovery in Images - Chainer implementation

Deep Learning for Human Part Discovery in Images - Chainer implementation NOTE: This is not official implementation. Original paper is Deep Learning f

Shintaro Shiba 63 Sep 25, 2022
A Python implementation of global optimization with gaussian processes.

Bayesian Optimization Pure Python implementation of bayesian global optimization with gaussian processes. PyPI (pip): $ pip install bayesian-optimizat

fernando 6.5k Jan 02, 2023
Source code for models described in the paper "AudioCLIP: Extending CLIP to Image, Text and Audio" (https://arxiv.org/abs/2106.13043)

AudioCLIP Extending CLIP to Image, Text and Audio This repository contains implementation of the models described in the paper arXiv:2106.13043. This

458 Jan 02, 2023
SOTR: Segmenting Objects with Transformers [ICCV 2021]

SOTR: Segmenting Objects with Transformers [ICCV 2021] By Ruohao Guo, Dantong Niu, Liao Qu, Zhenbo Li Introduction This is the official implementation

186 Dec 20, 2022
Official implementation of the ICLR 2021 paper

You Only Need Adversarial Supervision for Semantic Image Synthesis Official PyTorch implementation of the ICLR 2021 paper "You Only Need Adversarial S

Bosch Research 272 Dec 28, 2022
A high-performance distributed deep learning system targeting large-scale and automated distributed training.

HETU Documentation | Examples Hetu is a high-performance distributed deep learning system targeting trillions of parameters DL model training, develop

DAIR Lab 150 Dec 21, 2022
AgML is a comprehensive library for agricultural machine learning

AgML is a comprehensive library for agricultural machine learning. Currently, AgML provides access to a wealth of public agricultural datasets for common agricultural deep learning tasks.

Plant AI and Biophysics Lab 1 Jul 07, 2022