DEGREE: A Data-Efficient Generation-Based Event Extraction Model

Abstract

Event extraction requires high-quality expert human annotations, which are usually expensive. Therefore, learning a data-efficient event extraction model that can be trained with only a few labeled examples has become a crucial challenge. In this paper, we focus on low-resource end-to-end event extraction and propose DE-GREE, a data-efficient model that formulates event extraction as a conditional generation problem. Given a passage and a manually designed prompt, DEGREE learns to summarize the events mentioned in the passage into a natural sentence that follows a predefined pattern. The final event predictions are then extracted from the generated sentence with a deterministic algorithm. DEGREE has three advantages to learn well with less training data. First, our designed prompts provide semantic guidance for DEGREE to leverage label semantics and thus better capture the event arguments. Moreover, DEGREE is capable of using additional weaklysupervised information, such as the description of events encoded in the prompts. Finally, DE-GREE learns triggers and arguments jointly in an end-to-end manner, which encourages the model to better utilize the shared knowledge and dependencies among them. Our experimental results demonstrate the strong performance of DEGREE for low-resource event extraction.

Dynamic Prefix-Tuning for Generative Template-based Event Extraction

“…The man returned to Los Angeles from Mexico following his capture Tuesday by bounty hunters. the previous work Hsu et al, 2021), extracting event records one type by one type, using the pretrained encoder-decoder language model BART (Lewis et al, 2020) for conditional generation. For each event type, we first initialize a typespecific prefix consisting of a sequence of tunable vectors as transformer history values (Li and Liang, 2021).…”

“…There is a rising trend of casting the task of event extraction as a sequence generation problem by applying special decoding strategies (Paolini et al, 2021;Lu et al, 2021) or steering pretrained language models to output conditional generation sequences with discrete prompts Hsu et al, 2021). Compared with classification-based methods, this line of work is more data-efficient and flexible, which requires less annotated data to achieve acceptable model performances, being easier to extend to new event types by slightly modifying the designed prompts and decoding strategies.…”

“…However, these generation-based methods have two significant challenges, which impede achieving competitive results with the classification-based methods. (1) suboptimal prompts: First, they manually design prompts for each event type Hsu et al, 2021), which are suboptimal without tuning and largely affect the model performances.…”

KC-GEE: Knowledge-based Conditioning for Generative Event Extraction

“…Inspired by the success of pretrained language models and the corresponding natural language generation-based paradigm for various NLP tasks [4,[21][22][23] tackle event extraction as controlled event generation. [6] is an end-to-end conditional generation method with manually designed discrete prompts for each event type, which needs more human effort to find the Fig. 2 The event extraction task.…”

“…Natural language generation techniques have been successfully applied to a number of NLP tasks [4][5][6]. They have inspired the use of controlled event generation to tackle event extraction.…”