Liliang

Hello，I am just wondering about the setting of the data. As described in this paper "We conduct the experiments on the Amazon reviews dataset (Blitzer, Dredze, and Pereira 2007)". However,...

So, you mean the data is really from http://jmcauley.ucsd.edu/data/amazon/? I think that's not related the data (Blitzer, Dredze, and Pereira 2007). Aha, I also wanna your data transform codes like...

And by the way, Dear zheng. I surveyed some work on sentiment transfer and notice that the paper "Transfer Learning for Sequences via Learning to Collocate" was just accepted by...

Thanks for your generosity! I think it was chaos in the domain adaptation area of NLP. We are comparing different models for adaptation and find that even some mSDA based...

OK, I will try to reimplemented it.

Hi, i'm trying this great code. But as you say above, modeling the slot label dependency is benefit for the F1 score, but the code produces low intention classification accuracy...

I think you do more trick on the data preprocessing step or you just use the whole training set which not leave aside a validation data set(for I don't recognize...

Thanks for your reply, but it seems that some utterance may have more than one intent, (like flight+flight_fare). And according to the paper you referred(JOINT SEMANTIC UTTERANCE CLASSIFICATION AND SLOT...

因为close操作是同步的，这两个操作确实不需要放到bufferwriteProcessor类中，但是在刷新rowgroup的时候同样需要调用这两个操作，是异步完成的。`BufferWriteRecordWriter`调用`flushRowGroup`函数的时候需要异步的刷新，所以`bufferwriteCloseAction.act()`和`filenodeFlushAction.act()`两个操作有必要下放到bufferwriteProcessor类中完成。