ClarET: Pre-training a Correlation-Aware Context-To-Event Transformer for Event-Centric Generation and Classification
About
Generating new events given context with correlated ones plays a crucial role in many event-centric reasoning tasks. Existing works either limit their scope to specific scenarios or overlook event-level correlations. In this paper, we propose to pre-train a general Correlation-aware context-to-Event Transformer (ClarET) for event-centric reasoning. To achieve this, we propose three novel event-centric objectives, i.e., whole event recovering, contrastive event-correlation encoding and prompt-based event locating, which highlight event-level correlations with effective training. The proposed ClarET is applicable to a wide range of event-centric reasoning scenarios, considering its versatility of (i) event-correlation types (e.g., causal, temporal, contrast), (ii) application formulations (i.e., generation and classification), and (iii) reasoning types (e.g., abductive, counterfactual and ending reasoning). Empirical fine-tuning results, as well as zero- and few-shot learning, on 9 benchmarks (5 generation and 4 classification tasks covering 4 reasoning types with diverse event correlations), verify its effectiveness and generalization ability.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Story Ending Generation | ROCStories (test) | BLEU-157.47 | 43 | |
| Story Cloze Test | Story Cloze (test) | Accuracy91.18 | 38 | |
| Abductive Commonsense Reasoning | ANLI (test) | Accuracy82.77 | 23 | |
| Counterfactual Story Generation | Counterfactual Story (test) | BLEU-487.18 | 19 | |
| Script Reasoning | Script Reasoning | Accuracy64.61 | 18 | |
| Abductive Reasoning Generation | Abductive Commonsense Reasoning aNLG (test) | BLEU-417.67 | 14 | |
| Commonsense Story Generation | Commonsense (C.S.) Story Generation (test) | BLEU-157.47 | 14 | |
| Event Process Completion | Event Process Completion (test) | BLEU-1 Score58.88 | 10 | |
| Story Cloze Test | ROCStories Story Cloze Test | Accuracy0.9118 | 8 | |
| Narrative Incoherence Detection | ROCStories | Accuracy0.7488 | 7 |