Extractive approaches have been the mainstream paradigm for identifying overlapping entity–relation extraction. However, limited by their inherently methodological flaws, which hardly deal with three issues: hierarchical dependent entity–relations, implicit entity–relations, and entity normalization. Recent advances have proposed an effective solution based on generative language models, which cast entity–relation extraction as a sequence-to-sequence text generation task. Inspired by the observation that humans learn by getting to the bottom of things, we propose a novel framework, namely GenRE, Generative multi-turn question answering with contrastive learning for entity–relation extraction. Specifically, a template-based question prompt generation first is designed to answer in different turns. We then formulate entity–relation extraction as a generative question answering task based on the general language model instead of span-based machine reading comprehension. Meanwhile, the contrastive learning strategy in fine-tuning is introduced to add negative samples to mitigate the exposure bias inherent in generative models. Our extensive experiments demonstrate that GenRE performs competitively on two public datasets and a custom dataset, highlighting its superiority in entity normalization and implicit entity–relation extraction. (The code is available at https://github.com/lovelyllwang/GenRE).
Recently, the focus of entity linking research has centered on the zero-shot scenario, where the entity purposed to be labeled at the time of testing was never observed during the training phase, or it may belong to a different domain than the source domain. Current studies have used BERT as the base encoder, as it effectively establishes distributional links between source and target domains. The currently available negative sampling methods all use an extractive approach, which makes it difficult for the models to learn diverse and more challenging negative samples. To address this problem, we propose a generative negative sampling method, adaptive_mixup_hard, which generates more difficult negative entities by fusing the features of both positive and negative samples on top of hard negative sampling and introduces a transformable adaptive parameter, W, to increase the diversity of negative samples. Next, we fuse our method with the Biencoder architecture and evaluate its performance under three different score functions. Ultimately, experimental results on the standard benchmark dataset, Zeshel, demonstrate the effectiveness of our method.
Recommender systems (RecSys) have become essential in modern society, driving user engagement and satisfaction across diverse online platforms. Most RecSys focuses on designing a powerful encoder to embed users and items into high-dimensional vector representation space, with loss functions optimizing their representation distributions. Recent studies reveal that directly optimizing key properties of the representation distribution, such as alignment and uniformity, can outperform complex encoder designs. However, existing methods for optimizing critical attributes overlook the impact of dataset sparsity on the model: limited user-item interactions lead to sparse alignment, while excessive interactions result in uneven uniformity, both of which degrade performance. In this paper, we identify the sparse alignment and uneven uniformity issues, and further propose Regularized Alignment and Uniformity (RAU) to cope with these two issues accordingly. RAU consists of two novel regularization methods for alignment and uniformity to learn better user/item representation. 1) Center-strengthened alignment further aligns the average in-batch user/item representation to provide an enhanced alignment signal and further minimize the disparity between user and item representation. 2) Low-variance-guided uniformity minimizes the variance of pairwise distances along with uniformity, which provides extra guidance to a more stabilized uniformity increase during training. We conducted extensive experiments on three real-world datasets, and the proposed RAU resulted in significant performance improvements compared to current state-of-the-art CF methods, which confirms the advantages of the two proposed regularization methods.
Recently, the focus of entity linking research has centered on the zero-shot scenario, where the entity mention to be labeled at the time of testing was never observed during the training phase or may belong to a different domain than the source domain. Current studies have used BERT as the base encoder as it effectively establishes distributional links between source and target domains. The currently available negative sampling methods all use an extractive approach, which makes it difficult for the models to learn diverse and more challenging negative samples. To address this problem, we propose a generative negative sampling method, Adaptive_mixup_hard, which generates more difficult negative entities by fusing the features of both positive and negative samples on top of hard negative sampling and introduces a transformable adaptive parameter W to increase the diversity of negative samples. Next, we fuse our method with the Biencoder architecture and evaluate its performance under three different score functions. Ultimately, experimental results on the standard benchmark dataset Zeshel demonstrate the effectiveness of our method.