While Transformer-based pre-trained language models and their variants exhibit strong semantic representation capabilities, the question of comprehending the information gain derived from the additional components of PLMs remains an open question in this field. Motivated by recent efforts that prove Multilayer-Perceptrons (MLPs) modules achieving robust structural capture capabilities, even outperforming Graph Neural Networks (GNNs), this paper aims to quantify whether simple MLPs can further enhance the already potent ability of PLMs to capture linguistic information. Specifically, we design a simple yet effective probing framework containing MLPs components based on BERT structure and conduct extensive experiments encompassing 10 probing tasks spanning three distinct linguistic levels. The experimental results demonstrate that MLPs can indeed enhance the comprehension of linguistic structure by PLMs. Our research provides interpretable and valuable insights into crafting variations of PLMs utilizing MLPs for tasks that emphasize diverse linguistic structures.
Document-level event argument extraction poses new challenges of long input and cross-sentence inference compared to its sentence-level counterpart. However, most prior works focus on capturing the relations between candidate arguments and the event trigger in each event, ignoring two crucial points: a) non-argument contextual clue information; b) the relevance among argument roles. In this paper, we propose a SCPRG (Span-trigger-based Contextual Pooling and latent Role Guidance) model, which contains two novel and effective modules for the above problem. The Span-Trigger-based Contextual Pooling(STCP) adaptively selects and aggregates the information of non-argument clue words based on the context attention weights of specific argument-trigger pairs from pre-trained model. The Role-based Latent Information Guidance (RLIG) module constructs latent role representations, makes them interact through role-interactive encoding to capture semantic relevance, and merges them into candidate arguments. Both STCP and RLIG introduce no more than 1% new parameters compared with the base model and can be easily applied to other event extraction models, which are compact and transplantable. Experiments on two public datasets show that our SCPRG outperforms previous state-of-the-art methods, with 1.13 F1 and 2.64 F1 improvements on RAMS and WikiEvents respectively. Further analyses illustrate the interpretability of our model.
In the field of natural language understanding, the intersection of neural models and graph meaning representations (GMRs) remains a compelling area of research. Despite the growing interest, a critical gap persists in understanding the exact influence of GMRs, particularly concerning relation extraction tasks. Addressing this, we introduce DAGNN-plus, a simple and parameter-efficient neural architecture designed to decouple contextual representation learning from structural information propagation. Coupled with various sequence encoders and GMRs, this architecture provides a foundation for systematic experimentation on two English and two Chinese datasets. Our empirical analysis utilizes four different graph formalisms and nine parsers. The results yield a nuanced understanding of GMRs, showing improvements in three out of the four datasets, particularly favoring English over Chinese due to highly accurate parsers. Interestingly, GMRs appear less effective in literary-domain datasets compared to general-domain datasets. These findings lay the groundwork for better-informed design of GMRs and parsers to improve relation classification, which is expected to tangibly impact the future trajectory of natural language understanding research.
Document-level event argument extraction (EAE) is a crucial but challenging subtask in information extraction. Most existing approaches focus on the interaction between arguments and event triggers, ignoring two critical points: the information of contextual clues and the semantic correlations among argument roles. In this paper, we propose the CARLG model, which consists of two modules: Contextual Clues Aggregation (CCA) and Role-based Latent Information Guidance (RLIG), effectively leveraging contextual clues and role correlations for improving document-level EAE. The CCA module adaptively captures and integrates contextual clues by utilizing context attention weights from a pre-trained encoder. The RLIG module captures semantic correlations through role-interactive encoding and provides valuable information guidance with latent role representation. Notably, our CCA and RLIG modules are compact, transplantable and efficient, which introduce no more than 1% new parameters and can be easily equipped on other span-base methods with significant performance boost. Extensive experiments on the RAMS, WikiEvents, and MLEE datasets demonstrate the superiority of the proposed CARLG model. It outperforms previous state-of-the-art approaches by 1.26 F1, 1.22 F1, and 1.98 F1, respectively, while reducing the inference time by 31%. Furthermore, we provide detailed experimental analyses based on the performance gains and illustrate the interpretability of our model.
In the dataset of image captioning, each image is aligned with several captions. Despite the fact that the quality of these descriptions varies, existing captioning models treat them equally in the training process. In this paper, we propose a new control signal of sentence quality, which is taken as an additional input to the captioning model. By integrating the control signal information, captioning models are aware of the quality level of the target sentences and handle them differently. Moreover, we propose a novel reinforcement training method specially designed for the control signal of sentence quality: Quality-oriented Self-Annotated Training (Q-SAT). Equipped with R-Drop strategy, models controlled by the highest quality level surpass baseline models a lot on accuracy-based evaluation metrics, which validates the effectiveness of our proposed methods.
We consider the trade-off problem between exploration and exploitation under finite discounted Markov Decision Process, where the state transition matrix of the underlying environment stays unknown. We propose a double Thompson sampling reinforcement learning algorithm(DTS) to solve this kind of problem. This algorithm achieves a total regret bound of $\tilde{\mathcal{O}}(D\sqrt{SAT})$in time horizon $T$ with $S$ states, $A$ actions and diameter $D$. DTS consists of two parts, the first part is the traditional part where we apply the posterior sampling method on transition matrix based on prior distribution. In the second part, we employ a count-based posterior update method to balance between the local optimal action and the long-term optimal action in order to find the global optimal game value. We established a regret bound of $\tilde{\mathcal{O}}(\sqrt{T}/S^{2})$. Which is by far the best regret bound for finite discounted Markov Decision Process to our knowledge. Numerical results proves the efficiency and superiority of our approach.
Graph neural networks (GNNs) have drawn increasing attention in recent years and achieved remarkable performance in many graph-based tasks, especially in semi-supervised learning on graphs. However, most existing GNNs excessively rely on topological structures and aggregate multi-hop neighborhood information by simply stacking network layers, which may introduce superfluous noise information, limit the expressive power of GNNs and lead to the over-smoothing problem ultimately. In light of this, we propose a novel Dual-Perception Graph Neural Network (DPGNN) to address these issues. In DPGNN, we utilize node features to construct a feature graph, and perform node representations learning based on the original topology graph and the constructed feature graph simultaneously, which conduce to capture the structural neighborhood information and the feature-related information. Furthermore, we design a Multi-Hop Graph Generator (MHGG), which applies a node-to-hop attention mechanism to aggregate node-specific multi-hop neighborhood information adaptively. Finally, we apply self-ensembling to form a consistent prediction for unlabeled node representations. Experimental results on five datasets with different topological structures demonstrate that our proposed DPGNN outperforms all the latest state-of-the-art models on all datasets, which proves the superiority and versatility of our model. The source code of our model is available at https://github.com.