-
Notifications
You must be signed in to change notification settings - Fork 56
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
文章中BERT模型的疑问 #4
Comments
非常感谢关注我们的工作,我们论文的方案是采用多任务的方式来解决ABSA问题。典型的ABSA任务是给定文本包含的Aspect,进而判断在该Aspect上的情感倾向。确实不可能每段评论文本都包含所有Aspect信息,在具体实现中我们定义了一个mask vector [p1, p2, ..., pN ] 来提示哪些aspect被包含,包含pi=1,不包含pi=0,计算loss的时候只会计算被包含aspect上的损失。详情请关注我们论文第四章节Methodology部分,我们发布的数据是有aspect未被包含这一类别信息的。 |
模型在验证集和测试集上做评估的时候,也是假定事先就知道了评论文本中包含了哪些aspect吧? |
你好,想问下你后面怎么解决的这个问题呢? |
想请问这个问题是怎么解决的呢?预测的时候,是假定已经知道文本中挺提及了哪些aspect么? |
是的,acsa本身的任务定义就是需要知道文本包含哪些aspect |
是的 |
是的。ACSA任务定义就是给定aspect,判断情感极性 |
感谢解答 |
本文的BERT模型应该是采用单阶段的方法来解决方面级的情感分类吧,一共有18个方面,每个方面有3类。
但是,不应该再多出一个类来表示评论中是否包含该方面的信息吗?针对每个方面进行情感分类的时候,用的是交叉熵损失函数吧。这样的话,不就默认评论包含所有的方面信息吗?
The text was updated successfully, but these errors were encountered: