State-of-the-art models for joint entity recognition and relation extraction strongly rely on external natural language processing (NLP) tools such as POS (part-of-speech) taggers and dependency parsers. Thus, the performance of such joint models depends on the quality of the features obtained from these NLP tools. However, these features are not always accurate for various languages and contexts. In this paper, we propose a joint neural model which performs entity recognition and relation extraction simultaneously, without the need of any manually extracted features or the use of any external tool. Specifically, we model the entity recognition task using a CRF (Conditional Random Fields) layer and the relation extraction task as a multi-head selection problem (i.e., potentially identify multiple relations for each entity). We present an extensive experimental setup, to demonstrate the effectiveness of our method using datasets from various contexts (i.e., news, biomedical, real estate) and languages (i.e., English, Dutch). Our model outperforms the previous neural models that use automatically extracted features, while it performs within a reasonable margin of feature-based neural models, or even beats them.
#doc 2050 0 Mrs. B-Peop ['N'] [0] 1 Rose I-Peop ['N'] [1] 2 hired O ['N'] [2] 3 Abebe B-Peop ['N'] [3] 4 Worke I-Peop ['Work_For', 'Live_In'] [22, 8] 5 , O ['N'] [5] 6 one O ['N'] [6] 7 of O ['N'] [7] 8 Ethiopia B-Loc ['N'] [8] 9 's O ['N'] [9] 10 most O ['N'] [10] 11 distinguished O ['N'] [11] 12 lawyers O ['N'] [12] 13 and O ['N'] [13] 14 a O ['N'] [14] 15 former O ['N'] [15] 16 member O ['N'] [16] 17 of O ['N'] [17] 18 the O ['N'] [18] 19 country O ['N'] [19] 20 's O ['N'] [20] 21 High B-Org ['N'] [21] 22 Court I-Org ['N'] [22] 23 , O ['N'] [23] 24 to O ['N'] [24] 25 investigate O ['N'] [25] 26 . O ['N'] [26]
说明：由于我们假设基于令牌的编码，我们只考虑实体的最后一个令牌作为另一个令牌的头部，从而消除了冗余关系。 例如，实体“John Smith”和“Disease Control Center”之间存在关联工作。 我们不是连接实体的所有标记，而是仅将“Smith”与“Center”连接起来。 此外，对于没有关系的情况，我们引入“N”标签，我们将令牌本身预测为头部。
Since we assume token-based encoding, we consider only the last token of the entity as head of another token, eliminating redundant relations. For instance, there is a Works for relation between entities “John Smith” and “Disease Control Center”. Instead of connecting all tokens of the entities, we connect only “Smith” with “Center”. Also, for the case of no relation, we introduce the “N” label and we predict the token itself as the head.
