Event representation targets to model the event-reasoning process as a machine-readable format. Previous studies on event representation mostly concentrate on a sole modeling perspective and have not well investigated the scenario-level knowledge, which can cause information loss. To cope with this dilemma, we propose a unified fine-tuning architecture-based approach ( UniFA-S ) that integrates all levels of trainings, including the scenario-level knowledge. However, another challenge for existing models is the ever-increasing computation overheads, restricting the deployment ability on limited resources devices. Hence, in this article, we aim to compress the cumbersome model UniFA-S into a lighter and easy-to-deploy one without much performance damage. To this end, we propose a sequence-aware knowledge distillation model (SaKD) that employs a dynamic self-distillation on the decouple-compress-couple framework for compressing UniFA-S , which cannot only realize the model compression, but also retain the integrity of individual components. We also design two fitting strategies to address the less-supervised issue at the distillation stage. Comprehensive experiments on representation-and-inference ability-based tasks validate the effectiveness of SaKD. Compared to UniFA-S , SaKD realizes a more portable event representation model at the cost of only 1.0% performance drop in terms of accuracy or Spearman’s correlation, which is far less than other knowledge distillation models.

Document classification has a broad application in the field of sentiment classification, document ranking and topic labeling, etc. Previous neural network-based work has mainly focused on investigating a so-called forward implication, i.e., the preceding text segments are taken as the context of the following text segments when generating the text representation. Such a scenario typically ignores the fact that the semantics of a document are a product of the mutual implication of all text segments in a document. Thus, in this paper, we introduce a concept of interaction and propose a text representation model with Self-interaction Attention Mechanism (TextSAM) for document classification. In particular, we design three aggregated strategies to integrate the interaction into a hierarchical architecture for document classification, i.e., averaging the interaction, maximizing the interaction and adding one more attention layer on the interaction, which leads to three models, i.e., TextSAM_AVE, TextSAM_MAX and TextSAM_ATT, respectively. Our comprehensive experimental results on two public datasets, i.e., Yelp 2016 and Amazon Reviews (Electronics), show that our proposals can significantly outperform the state-of-the-art neural-based baselines for document classification, presenting a general improvement in terms of accuracy ranging from 5.97% to 14.05% against the best baseline. Furthermore, we find that our proposals with a self-interaction attention mechanism can obviously alleviate the impact brought by the increase of sentence number as the relative improvement of our proposals against the baselines are enlarged when the sentence number increases.

Session-based recommendation (SBRS) aims to make recommendations for users merely based on the ongoing session. Existing GNN-based methods achieve satisfactory performance by exploiting the pair-wise item transition pattern; however, they ignore the temporal evolution of the session graphs over different time-steps. Moreover, the widely applied cross-entropy loss with softmax in SBRS faces the serious overfitting problem. To deal with the above issues, we propose dynamic graph learning for session-based recommendation (DGL-SR). Specifically, we design a dynamic graph neural network (DGNN) to simultaneously take the graph structural information and the temporal dynamics into consideration for learning the dynamic item representations. Moreover, we propose a corrective margin softmax (CMS) to prevent overfitting in the model optimization by correcting the gradient of the negative samples. Comprehensive experiments are conducted on two benchmark datasets, that is, Diginetica and Gowalla, and the experimental results show the superiority of DGL-SR over the state-of-the-art baselines in terms of Recall@20 and MRR@20, especially on hitting the target item in the recommendation list.

Knowledge-enhanced recommendation (KER) aims to integrate the knowledge graph (KG) into collaborative filtering (CF) for alleviating the sparsity and cold start problems. The state-of-the-art graph neural network (GNN)–based methods mainly focus on exploiting the connectivity between entities in the knowledge graph, while neglecting the interaction relation between items reflected in the user-item interactions. Moreover, the widely adopted BPR loss for model optimization fails to provide sufficient supervisions for learning discriminative representation of users and items. To address these issues, we propose the collaborative knowledge-enhanced recommendation (CKER) method. Specifically, CKER proposes a collaborative graph convolution network (CGCN) to learn the user and item representations from the connection between items in the constructed interaction graph and the connectivity between entities in the knowledge graph. Moreover, we introduce the self-supervised learning to maximize the mutual information between the interaction- and knowledge-aware user preferences by deriving additional supervision signals. We conduct comprehensive experiments on two benchmark datasets, namely Amazon-Book and Last-FM, and the experimental results show that CKER can outperform the state-of-the-art baselines in terms of recall and NDCG on knowledge-enhanced recommendation.

Fact verification aims to verify the authenticity of a given claim based on the retrieved evidence from Wikipedia articles. Existing works mainly focus on enhancing the semantic representation of evidence, e.g., introducing the graph structure to model the evidence relation. However, previous methods can’t well distinguish semantic-similar claims and evidences with distinct authenticity labels. In addition, the performances of graph-based models are limited by the over-smoothing problem of graph neural networks. To this end, we propose a graph-based contrastive learning method for fact verification abbreviated as CosG, which introduces a contrastive label-supervised task to help the encoder learn the discriminative representations for different-label claim-evidence pairs, as well as an unsupervised graph-contrast task, to alleviate the unique node features loss in the graph propagation. We conduct experiments on FEVER, a large benchmark dataset for fact verification. Experimental results show the superiority of our proposal against comparable baselines, especially for the claims that need multiple-evidences to verify. In addition, CosG presents better model robustness on the low-resource scenario.

Chinese Medical Named Entity Recognition (Chinese-MNER) aims to identify potential entities and their categories from the unstructured Chinese medical text. Existing methods for this task mainly incorporate the dictionary knowledge on the basis of traditional BiLSTM-CRF or BERT architecture. However, the construction of high-quality dictionaries is typically time consuming and labor-intensive, which may also damage the robustness of NER models. What is more, the limited amount of annotated Chinese-MNER data can easily lead to the over-fitting problem while training. With the aim of dealing with the above problems, we put forward a BERT-BiLSTM-CRF model by integrating the part-of-speech (POS) tagging features and a Regularization method (BBCPR) for Chinese-MNER. In BBCPR, we first leverage a POS fusion layer to incorporate external syntax knowledge. Next, we design a novel REgularization mothod with Adversarial training and Dropout (READ) to improve the model robustness. Specifically, READ focuses on reducing the difference between the predictions of two sub-models through minimizing the bidirectional KL divergence between the adversarial output and original output distributions for the same sample. Comprehensive evaluations on two public data sets, namely, cMedQANER and cEHRNER from the Chinese Biomedical Language Understanding Evaluation benchmark (ChineseBLUE), demonstrate the superiority of our proposal in Chinese-MNER. In addition, ablation study shows that READ can effectively improve the model performance. Our proposal does well in exploring the technical terms and identifying the word boundary.