Fixed Point Continuation

The essence of distantly supervised rela- tion extraction is that it is an incomplete multi-label classification problem with s- parse and noisy features. To tackle the s- parsity and noise …...