ZeroShotEvent

ZeroShotEvent copied to clipboard

I reproduced the code for your trigger classification. Experiment using 10 event types. The accuracy rate of the experiment was 62.32% on hit@5, which could not reach the 68.3% mentioned...

lisiming228

Question about python rsd2ltf.py

4

Hello, I encountered some trouble when running your code and would you please give me some suggestion? When I run **"java -jar removeXmlTag.jar [source_path] [source_without_tag_path]"**, I can only get 25...

carrie0307

关于论文实验的细节，触发词分类时候选事件类型是1194种吗

2

原文： _For our approach, we followed the experiment setting D in the previous section, using the same training and development data sets for the 10 seen types, but targeted all...

LarryLee-BD

额，能说下用的什么embedding吗。

R-craft

Hello, I am having troubles running rsd2ltf.py It seems from tokenizer import Tokenizer is the line where it fails. I have installed tokenizer and that seems to work fine, but...

sidhartha4

作者你好，我最近在读你的论文，非常受启发。但是我对一些细节仍然有所疑惑。 我想问一下作者在实现的过程中，整一个pipline训练了多少个不同的神经网络? 比如assign trigger和assign argument的是同一个神经网络吗?另外对于mention structure和type structure中对tuple表示的方式不同问题，为什么type structure中要使用三维的张量来表示关系呢? 希望作者能够解答我的疑惑，万分感谢!!!

cjlala