A very simple framework for state-of-the-art Natural Language Processing (NLP)

基于Pytorch和torchtext的知识图谱深度学习框架。

A neural network architecture for NLP tasks, using cython for fast performance. Currently, it can perform POS tagging, SRL and dependency parsing.

Deep Semantic Role Labeling with Self-Attention

Astock

Reimplementation of a BERT based model (Shi et al, 2019), currently the state-of-the-art for English SRL. This model implements also predicate disambiguation.

Collection of papers on Emotion Cause Analysis

[COLING'22] Code for "Semantic Role Labeling as Dependency Parsing: Exploring Latent Tree Structures Inside Arguments".

Natural Language Parsing and Feature Generation

VerbNet semantic parser and related utilities

The Mingled Structured Predictor

Scripts for preprocessing the CoNLL-2005 SRL dataset.

This is a meta-model distilled from LLMs for information extraction. This is an intermediate checkpoint that can be well-transferred to all kinds of downstream information extraction tasks.

SRL deep learning model is based on DB-LSTM which is described in this paper : [End-to-end learning of semantic role labeling using recurrent neural networks](http://www.aclweb.org/anthology/P15-1109)

*SEM 2018: Learning Distributed Event Representations with a Multi-Task Approach

A Structured Span Selector (NAACL 2022). A structured span selector with a WCFG for span selection tasks (coreference resolution, semantic role labelling, etc.)

Sequence Labeling in Tensorflow

GSRL is a seq2seq model for end-to-end dependency- and span-based SRL (IJCAI2021).

Unifying Cross-Lingual Semantic Role Labeling with Heterogeneous Linguistic Resources (NAACL-2021).

BERT-based nominal Semantic Role Labeling (SRL), both using the Nombank dataset and the Ontonotes dataset.