Code for "Parallel Instance Query Network for Named Entity Recognition", accepted at ACL 2022.

Related tags

Text Data & NLP piqn
Overview

README

Code for Two-stage Identifier: "Parallel Instance Query Network for Named Entity Recognition", accepted at ACL 2022. For details of the model and experiments, please see our paper.

Setup

Requirements

conda create --name acl python=3.8
conda activate acl
pip install -r requirements.txt

Datasets

Nested NER:

Flat NER:

Data format:

{
    "tokens": ["Others", ",", "though", ",", "are", "novices", "."], 
    "entities": [{"type": "PER", "start": 0, "end": 1}, {"type": "PER", "start": 5, "end": 6}], "relations": [], "org_id": "CNN_IP_20030328.1600.07", 
    "ltokens": ["WOODRUFF", "We", "know", "that", "some", "of", "the", "American", "troops", "now", "fighting", "in", "Iraq", "are", "longtime", "veterans", "of", "warfare", ",", "probably", "not", "most", ",", "but", "some", ".", "Their", "military", "service", "goes", "back", "to", "the", "Vietnam", "era", "."], 
    "rtokens": ["So", "what", "is", "it", "like", "for", "them", "to", "face", "combat", "far", "from", "home", "?", "For", "an", "idea", ",", "here", "is", "CNN", "'s", "Candy", "Crowley", "with", "some", "war", "stories", "."]
}

The ltokens contains the tokens from the previous sentence. And The rtokens contains the tokens from the next sentence.

Due to the license, we cannot directly release our preprocessed datasets of ACE04, ACE05, KBP17, NNE and OntoNotes. We only release the preprocessed GENIA, FewNERD, MSRA and CoNLL03 datasets. Download them from here.

If you need other datasets, please contact me ([email protected]) by email. Note that you need to state your identity and prove that you have obtained the license.

Example

Train

python piqn.py train --config configs/nested.conf

Note: You should edit this line in config_reader.py according to the actual number of GPUs.

Evaluation

You can download our checkpoints on ACE04 and ACE05, or train your own model and then evaluate the model. Because of the limited space of Google Cloud Drive, we share the other models in Baidu Cloud Drive, please download at this link (code: js9z).

python identifier.py eval --config configs/batch_eval.conf

If you use the checkpoints (ACE05 and ACE04) we provided, you will get the following results:

  • ACE05:
2022-03-30 12:56:52,447 [MainThread  ] [INFO ]  --- NER ---
2022-03-30 12:56:52,447 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                  type    precision       recall     f1-score      support
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   PER        88.07        92.92        90.43         1724
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   LOC        63.93        73.58        68.42           53
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   WEA        86.27        88.00        87.13           50
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   GPE        87.22        87.65        87.44          405
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   ORG        85.74        81.64        83.64          523
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   VEH        83.87        77.23        80.41          101
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                   FAC        75.54        77.21        76.36          136
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                 micro        86.38        88.57        87.46         2992
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]                 macro        81.52        82.61        81.98         2992
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]  --- NER on Localization ---
2022-03-30 12:56:52,475 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]                  type    precision       recall     f1-score      support
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]                Entity        90.58        92.91        91.73         2991
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]                 micro        90.58        92.91        91.73         2991
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]                 macro        90.58        92.91        91.73         2991
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]  --- NER on Classification ---
2022-03-30 12:56:52,496 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                  type    precision       recall     f1-score      support
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   PER        97.09        92.92        94.96         1724
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   LOC        76.47        73.58        75.00           53
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   WEA        95.65        88.00        91.67           50
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   GPE        92.93        87.65        90.22          405
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   ORG        93.85        81.64        87.32          523
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   VEH       100.00        77.23        87.15          101
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                   FAC        89.74        77.21        83.00          136
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]  
2022-03-30 12:56:52,516 [MainThread  ] [INFO ]                 micro        95.36        88.57        91.84         2992
2022-03-30 12:56:52,517 [MainThread  ] [INFO ]                 macro        92.25        82.61        87.05         2992
  • ACE04
2021-11-15 22:06:50,896 [MainThread  ] [INFO ]  --- NER ---
2021-11-15 22:06:50,896 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                  type    precision       recall     f1-score      support
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                   VEH        88.89        94.12        91.43           17
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                   WEA        74.07        62.50        67.80           32
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                   GPE        89.11        87.62        88.36          719
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                   ORG        85.06        84.60        84.83          552
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                   FAC        83.15        66.07        73.63          112
2021-11-15 22:06:50,932 [MainThread  ] [INFO ]                   PER        91.09        92.12        91.60         1498
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]                   LOC        72.90        74.29        73.58          105
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]                 micro        88.48        87.81        88.14         3035
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]                 macro        83.47        80.19        81.61         3035
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]  --- NER on Localization ---
2021-11-15 22:06:50,933 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]                  type    precision       recall     f1-score      support
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]                Entity        92.56        91.89        92.23         3034
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]                 micro        92.56        91.89        92.23         3034
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]                 macro        92.56        91.89        92.23         3034
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,954 [MainThread  ] [INFO ]  --- NER on Classification ---
2021-11-15 22:06:50,955 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                  type    precision       recall     f1-score      support
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   VEH        94.12        94.12        94.12           17
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   WEA        95.24        62.50        75.47           32
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   GPE        95.60        87.62        91.44          719
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   ORG        93.59        84.60        88.87          552
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   FAC        93.67        66.07        77.49          112
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   PER        97.11        92.12        94.55         1498
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                   LOC        84.78        74.29        79.19          105
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]  
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                 micro        95.59        87.81        91.53         3035
2021-11-15 22:06:50,976 [MainThread  ] [INFO ]                 macro        93.44        80.19        85.87         3035

Citation

If you have any questions related to the code or the paper, feel free to email [email protected].

@inproceedings{shen-etal-2022-piqn,
    title = "Parallel Instance Query Network for Named Entity Recognition",
    author = "Shen, Yongliang  and
      Wang, Xiaobin  and
      Tan, Zeqi  and
      Xu, Guangwei  and
      Xie, Pengjun  and
      Huang, Fei and
      Lu, Weiming and
      Zhuang, Yueting",
    booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics",
    year = "2022",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/2203.10545",
}
You might also like...
Code for our ACL 2021 (Findings) Paper - Fingerprinting Fine-tuned Language Models in the wild .

🌳 Fingerprinting Fine-tuned Language Models in the wild This is the code and dataset for our ACL 2021 (Findings) Paper - Fingerprinting Fine-tuned La

LCS2-IIITDelhi 5 Sep 13, 2022
Code for our paper "Transfer Learning for Sequence Generation: from Single-source to Multi-source" in ACL 2021.

TRICE: a task-agnostic transferring framework for multi-source sequence generation This is the source code of our work Transfer Learning for Sequence

THUNLP-MT 9 Jun 27, 2022
Code for papers "Generation-Augmented Retrieval for Open-Domain Question Answering" and "Reader-Guided Passage Reranking for Open-Domain Question Answering", ACL 2021

This repo provides the code of the following papers: (GAR) "Generation-Augmented Retrieval for Open-domain Question Answering", ACL 2021 (RIDER) "Read

morning 49 Dec 26, 2022
This repository contains the code, data, and models of the paper titled "XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Languages" published in Findings of the Association for Computational Linguistics: ACL 2021.

XL-Sum This repository contains the code, data, and models of the paper titled "XL-Sum: Large-Scale Multilingual Abstractive Summarization for 44 Lang

null 189 Jan 2, 2023
🍊 PAUSE (Positive and Annealed Unlabeled Sentence Embedding), accepted by EMNLP'2021 🌴

PAUSE: Positive and Annealed Unlabeled Sentence Embedding Sentence embedding refers to a set of effective and versatile techniques for converting raw

EQT 21 Dec 15, 2022
A simple tool to update bib entries with their official information (e.g., DBLP or the ACL anthology).

Rebiber: A tool for normalizing bibtex with official info. We often cite papers using their arXiv versions without noting that they are already PUBLIS

(Bill) Yuchen Lin 2k Jan 1, 2023
Findings of ACL 2021

Assessing Dialogue Systems with Distribution Distances [arXiv][code] We propose to measure the performance of a dialogue system by computing the distr

Yahui Liu 16 Feb 24, 2022
A PyTorch implementation of paper
A PyTorch implementation of paper "Learning Shared Semantic Space for Speech-to-Text Translation", ACL (Findings) 2021

Chimera: Learning Shared Semantic Space for Speech-to-Text Translation This is a Pytorch implementation for the "Chimera" paper Learning Shared Semant

Chi Han 43 Dec 28, 2022
LV-BERT: Exploiting Layer Variety for BERT (Findings of ACL 2021)

LV-BERT Introduction In this repo, we introduce LV-BERT by exploiting layer variety for BERT. For detailed description and experimental results, pleas

Weihao Yu 14 Aug 24, 2022
Comments
  • 【matrix contains invalid numeric entries】匈牙利算法优化过程

    【matrix contains invalid numeric entries】匈牙利算法优化过程

    作者你好，非常喜欢你们在locate and label和piqn的工作，在复现过程中遇到了以下问题。如果你们遇到过类似的问题，可以告知一下解决方案吗？ 由于硬件限制，我们使用了最新版的torch和transformer。希望不会带来影响。 在复现的过程中，我们遇到了以下错误。在github，CSDN个Stack Overflow查询后发现是匈牙利算法的优化问题。 错误发生在非常后面的epoch里，这给我们的debug带来了很大的困难。

    Process SpawnProcess-1:
Traceback (most recent call last):
  File "/home/kk/anaconda3/envs/piqn/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
    self.run()
  File "/home/kk/anaconda3/envs/piqn/lib/python3.8/multiprocessing/process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
  File "/home/kk/piqn/piqn.py", line 15, in __train
    trainer.train(train_path=run_args.train_path[:-5]+ '_' + run_args.index + '.json',
  File "/home/kk/piqn/piqn/piqn_trainer.py", line 188, in train
    self._train_epoch(model, compute_loss, optimizer, train_dataset, updates_epoch, epoch)
  File "/home/kk/piqn/piqn/piqn_trainer.py", line 285, in _train_epoch
    batch_loss = compute_loss.compute(entity_logits, p_left, p_right, output, gt_types=batch['gt_types'], gt_spans = batch['gt_spans'], entity_masks=batch['entity_masks'], epoch = epoch,  deeply_weight = args.deeply_weight, seq_logits = masked_seq_logits, gt_seq_labels=batch['gt_seq_labels'], batch = batch)
  File "/home/kk/piqn/piqn/loss.py", line 59, in compute
    loss_dict = self.criterion(outputs, targets, epoch)
  File "/home/kk/anaconda3/envs/piqn/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/kk/piqn/piqn/loss.py", line 261, in forward
    indices = self.matcher(outputs, targets)
  File "/home/kk/anaconda3/envs/piqn/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/kk/anaconda3/envs/piqn/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context
    return func(*args, **kwargs)
  File "/home/kk/piqn/piqn/matcher.py", line 76, in forward
    indices = [linear_sum_assignment(c[i]) for i, c in enumerate(C.split(sizes, -1))]
  File "/home/kk/piqn/piqn/matcher.py", line 76, in <listcomp>
    indices = [linear_sum_assignment(c[i]) for i, c in enumerate(C.split(sizes, -1))]
  File "/home/kk/anaconda3/envs/piqn/lib/python3.8/site-packages/scipy/optimize/_lsap.py", line 93, in linear_sum_assignment
    raise ValueError("matrix contains invalid numeric entries")
ValueError: matrix contains invalid numeric entries

    参考链接： stack overflow scipy发现了这个问题，并且似乎改进了它，但依然出错 scipy issue

    opened by kk19990709 4
  • 【multiprocessing】

    【multiprocessing】 

    Traceback (most recent call last):
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
    self.run()
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/multiprocessing/process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
  File "/data/kk/GEEK/piqn/piqn.py", line 14, in __train
    trainer.train(train_path=run_args.train_path, valid_path=run_args.valid_path,
  File "/data/kk/GEEK/piqn/piqn/piqn_trainer.py", line 164, in train
    if args.local_rank != -1:
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 612, in to
    return self._apply(convert)
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 359, in _apply
    module._apply(fn)
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 359, in _apply
    module._apply(fn)
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 359, in _apply
    module._apply(fn)
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 381, in _apply
    param_applied = fn(param)
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 610, in convert
    return t.to(device, dtype if t.is_floating_point() else None, non_blocking)
  File "/home/deeplearning/anaconda3/envs/acl/lib/python3.8/site-packages/torch/cuda/__init__.py", line 163, in _lazy_init
    raise RuntimeError(
RuntimeError: Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method
    opened by kk19990709 2
  • OntoNotes 5.0 Version

    OntoNotes 5.0 Version

    Dear authors, thanks a lot for the nice work and the git repository.

    I have the access of the original/raw OnotoNotes 5.0 data, yet I am not clear how the preprocessing has been done on your side.

    1. Your repository offers this link for preprocessing; it is mentioned that there are two versions: v4 and v12. May I know which one are you using?
    2. According to the paper, it seems that you neither use v4 nor v12. Instead, you seem to use the flat version of OntoNotes 5.0. If so, may I know the reason for that? Also, I don't know how to access the flat version of the OntoNotes 5.0 dataset, even on their official repository.
    3. In general, do you know which OntoNotes version we should use? v4, v12 or flat?

    Thanks a lot!

    opened by addiu 2
  • 关于entity point阶段的一些问题

    关于entity point阶段的一些问题

    首先感谢如此出色的工作，有一些问题需要向您请教。 文中指出，每一个query能够指出一个实体，也就是指出一个span的左右两个边界，请问模型是如何保证每一个query一定会选取左右两个边界的呢？我理解的做法是，每一个query对每一个token做询问，得到一个p，最后选取p值最大的两个token作为query所对应选取的边界，请问这样理解对吗？

    opened by WangSheng21s 1
Owner
Yongliang Shen
Knowledge is power.
Yongliang Shen
GitHub
Hostapd-mac-tod-acl - Setup a hostapd AP with MAC ToD ACL

A brief explanation This script provides a quick way to setup a Time-of-day (Tod

null 2 Feb 3, 2022
(ACL 2022) The source code for the paper "Towards Abstractive Grounded Summarization of Podcast Transcripts"

Towards Abstractive Grounded Summarization of Podcast Transcripts We provide the source code for the paper "Towards Abstractive Grounded Summarization

null 10 Jul 1, 2022
Code for Findings of ACL 2022 Paper "Sentiment Word Aware Multimodal Refinement for Multimodal Sentiment Analysis with ASR Errors"

SWRM Code for Findings of ACL 2022 Paper "Sentiment Word Aware Multimodal Refinement for Multimodal Sentiment Analysis with ASR Errors" Clone Clone th

null 14 Jan 3, 2023
Code for ACL 2022 main conference paper "STEMM: Self-learning with Speech-text Manifold Mixup for Speech Translation".

STEMM: Self-learning with Speech-Text Manifold Mixup for Speech Translation This is a PyTorch implementation for the ACL 2022 main conference paper ST

ICTNLP 29 Oct 16, 2022
Implementaion of our ACL 2022 paper Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation

Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation This is the implementaion of our paper: Bridging the

hezw.tkcw 20 Dec 12, 2022
PyTorch Implementation of "Bridging Pre-trained Language Models and Hand-crafted Features for Unsupervised POS Tagging" (Findings of ACL 2022)

Feature_CRF_AE Feature_CRF_AE provides a implementation of Bridging Pre-trained Language Models and Hand-crafted Features for Unsupervised POS Tagging

Jacob Zhou 6 Apr 29, 2022
Entity Disambiguation as text extraction (ACL 2022)

ExtEnD: Extractive Entity Disambiguation This repository contains the code of ExtEnD: Extractive Entity Disambiguation, a novel approach to Entity Dis

Sapienza NLP group 121 Jan 3, 2023
Code for ACL 2021 main conference paper "Conversations are not Flat: Modeling the Intrinsic Information Flow between Dialogue Utterances".

Conversations are not Flat: Modeling the Intrinsic Information Flow between Dialogue Utterances This repository contains the code and pre-trained mode

ICTNLP 90 Dec 27, 2022
Code for our paper "Mask-Align: Self-Supervised Neural Word Alignment" in ACL 2021

Mask-Align: Self-Supervised Neural Word Alignment This is the implementation of our work Mask-Align: Self-Supervised Neural Word Alignment. @inproceed

THUNLP-MT 46 Dec 15, 2022
Code for our ACL 2021 paper - ConSERT: A Contrastive Framework for Self-Supervised Sentence Representation Transfer

ConSERT Code for our ACL 2021 paper - ConSERT: A Contrastive Framework for Self-Supervised Sentence Representation Transfer Requirements torch==1.6.0

Yan Yuanmeng 478 Dec 25, 2022