2021/06/22 by Ruotian Ma, Ma, Ruotian, Tao Gui +9
Computer Science · Engineering · #Artificial intelligence #Computation and Language (cs.CL) #Computer science #Construct (python library) #Data mining #Engineering #FOS: Computer and information sciences #Machine learning #Natural Language Processing Techniques #Natural language processing #Noise (video) #Noisy data #Pattern recognition (psychology) #Process (computing) #Relation (database) #Relationship extraction #Sentence #Sequence labeling #Speech recognition #Text and Document Classification Technologies #Topic Modeling #Training set #cs.CL
paper · pdf · doi:10.48550/arxiv.2106.11566
published in arXiv (Cornell University) (Cornell University) · Accepted by ACL 2021
arxiv created 2021/06/22 · openalex publication_date 2021/06/22 · arxiv updated 2021/06/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
Distant supervision for relation extraction provides uniform bag labels for each sentence inside the bag, while accurate sentence labels are important for downstream applications that need the exact relation type. Directly using bag labels for sentence-level training will introduce much noise, thus severely degrading performance. In this work, we propose the use of negative training (NT), in which a model is trained using complementary labels regarding that ``the instance does not belong to these complementary labels". Since the probability of selecting a true label as a complementary label is low, NT provides less noisy information. Furthermore, the model trained with NT is able to separate the noisy data from the training data. Based on NT, we propose a sentence-level framework, SENT, for distant relation extraction. SENT not only filters the noisy data to construct a cleaner dataset, but also performs a re-labeling process to transform the noisy data into useful training data, thus further benefiting the model's performance. Experimental results show the significant improvement of the proposed method over previous methods on sentence-level evaluation and de-noise effect.