A Prior Information Enhanced Extraction Framework for Document-level Financial Event Extraction

Document-level financial event extraction (DFEE) is the task of detecting event and extracting the corresponding event arguments in financial documents, which plays an important role in information extraction in the financial domain. This task is challenging as the financial documents are generally long text and event arguments of one event may be scattered in different sentences. To address this issue, we propose a novel Prior Information Enhanced Extraction framework (PIEE) for DFEE, leveraging prior information from both event types and pre-trained language models. Specifically, PIEE consists of three components: event detection, event argument extraction, and event table filling. In event detection, we identify the event type. Then, the event type is explicitly used for event argument extraction. Meanwhile, the implicit information within language models also provides considerable cues for event arguments localization. Finally, all the event arguments are filled in an event table by a set of predefined heuristic rules. To demonstrate the effectiveness of our proposed framework, we participate the share task of CCKS2020 Task5-2: Document-level Event Arguments Extraction. On both Leaderboard A and Leaderboard B, PIEE takes the first place and significantly outperforms the other systems.

Revisiting Olga, the beautiful dancer: An intersective A-analysis

The paper presents a novel semantic account of the so-called "intersective/non-intersective" ambiguity of structures such as beautiful dancer. The proposal contrasts with Larson's (1998) famous N-analysis in taking the adjective as the ambiguity trigger and in unmasking the bracketing paradox perception of the non-intersective reading as a grammatical illusion. The adjective has no compositional access to the verbal root's event argument but is always linked to the referential argument of the noun. -er nominals are analyzed as a special kind of role noun (such as king, guest, judge). They introduce a social role r that manifests itself via the verbal root's e-argument. (However, neither r nor e are compositionally active.) An evaluative adjective such as beautiful introduces an underspecified trope variable, which calls for a pragmatic specification of the adjectival predicate's ultimate target. A general pragmatic parsimony condition ensures that referents introduced by linguistic material are chosen as best target candidates whenever possible. The -er nominal's social role r is an ideal choice in this respect. The linking of the adjective to the verbal root's e-argument is mediated via r and thus a secondary pragmatic effect. The proposal provides a unified analysis for modified -er nominals (beautiful dancer) and other instances of role- and event-related interpretations for adnominal modification such as, e.g., just king.