The intelligent question-generation system based on large Language models aims to address the issues of low efficiency and susceptibility to subjective factors in traditional manual question-setting, as well as the limited scale of existing question banks and potential copyright risks. This system integrates multi-turn dialogue mechanisms, Retrieval-Augmented Generation (RAG) technology, and the Agent framework to automatically generate questions based on teaching syllabi and knowledge points. The working principle of the system involves first constructing a knowledge base through an electronic resource parsing module. It then continuously optimizes the accuracy of the generated questions through multi-turn dialogue prompt learning. Meanwhile, the Agent generation module breaks down the question-setting process into extraction, generation, and evaluation stages, with specialized agents monitoring and optimizing each stage. Finally, the question evaluation module assesses the questions from multiple dimensions, such as difficulty and accuracy. After practical testing, the system generated 1000 questions for the “Wireless Communication” course, and a random inspection of 100 questions revealed an accuracy rate of 94

This paper proposes a novel educational system that integrates Large Language Models (LLMs) with Knowledge Graphs (KGs) to generate personalized learning paths. By leveraging the structured knowledge representation of KGs and the generative capabilities of LLMs, the system aims to enhance the learning experience in the field of Artificial Intelligence Programming. Our contributions include the construction of a comprehensive KG for AI programming, the development of a prompt engineering framework for LLMs, and the implementation of a web-based learning system. The system demonstrates significant potential in providing context-aware and user-friendly learning experiences, paving the way for more intelligent educational tools.

Translating abstract concepts from text into visual imagery presents a significant challenge for generative AI. Although text-to-image models have made remarkable progress, they often struggle to capture the profound cultural nuances, emotional depth, and era-specific details inherent in classical poetry. To address this, we propose a novel closed-loop framework that synergistically integrates Large Language Models (LLMs), Diffusion Models, and a Multi-modal Large Language Model (MLLM) to provide evaluative feedback. Our system is built around a core pipeline: "Text Semantic Parsing → Structured Prompt Construction → Image Generation → Image-Text Alignment Analysis → Multi-round Optimization." Specifically, we employ Baidu's ERNIE model for in-depth semantic parsing of the text, construct culturally-aware structured prompts for diffusion models (such as SDXL, Playground v2, and Kandinsky), and introduce a tailored evaluation system using the MoE-LLaVA model to assess the generated images. Automatic multi-round optimization is then triggered based on this assessment to refine the outputs. Experimental results show that our method significantly improves both the semantic alignment (CLIP Score) and human preference score (HPS) of the generated images compared to baseline approaches, with the highest improvements reaching 34.44% and 28.87%, respectively. This work provides a robust technical pathway for the digital preservation and innovative interpretation of classical poetry.

Predicting unseen relations that cannot be observed during the training phase is a challenging task in relation extraction. Previous works have made progress by matching the semantics between input instances and label descriptions. However, fine-grained matching often requires laborious manual annotation, and rich interactions between instances and label descriptions come with significant computational overhead. In this work, we propose an efficient multi-grained matching approach that uses virtual entity matching to reduce manual annotation cost, and fuses coarse-grained recall and fine-grained classification for rich interactions with guaranteed inference speed. Experimental results show that our approach outperforms the previous State Of The Art (SOTA) methods, and achieves a balance between inference efficiency and prediction accuracy in zero-shot relation extraction tasks. Our code is available at https://github.com/longls777/EMMA.

Cross-domain few-shot Relation Extraction (RE) aims to transfer knowledge from a source domain to a different target domain to address low-resource problems. Previous work utilized label descriptions and entity information to leverage the knowledge of the source domain. However, these models are prone to confusion when directly applying this knowledge to a target domain with entirely new types of relations, which becomes particularly pronounced when facing similar relations. In this work, we propose a relation-aware prompt learning method with pre-training. Specifically, we empower the model to clear confusion by decomposing various relation types through an innovative label prompt, while a context prompt is employed to capture differences in different scenarios, enabling the model to further discern confusion. Two pre-training tasks are designed to leverage the prompt knowledge and paradigm. Experiments show that our method outperforms previous sota methods, yielding significantly better results on cross-domain few-shot RE tasks.

Mitigating linguistic bias and attaining modal equilibrium in Visual Question Answering (VQA) tasks constitute a pivotal concern. Previous work has mainly focused on data augmentation or a uni-modal approach, which is insufficient to fully utilize bi-modal information. In this work, we propose a new causal modal equilibrium framework CausalME, addressing the issue from a causal perspective. CausalME utilizes a question-only branch to capture the linguistic bias of the textual modality and mitigate its causal effect with a newly designed adaptive paradigm. Additionally, CausalME employs counterfactual generation to enhance the causal effect of visual modality. By optimizing the objective function of the entire VQA model, CausalME balances the causal effects of bi-modalities and explicitly guides the model to align text and image information. We conducted extensive experiments and the results show that CausalME brings significant improvements and achieves competitive performance on the bias-sensitive VQA-CP v2 dataset.

Multiple-choice question answering (MCQA) requires the model to select the correct answer from a set of candidate options when given a passage and a question. Previous research has achieved promising results with the assistance of Pre-trained Language Models(PrLMs). However, it has been observed that these models heavily rely on text similarity for inference. In this study, we approach the MCQA task from a causal perspective, treating text similarity as a confounder and establishing a structural causal model(SCM) for the MCQA task. To mitigate the impact of this confounder, we propose a novel Causal Dual Multi-head Co-Attention (CDUMA) model. CDUMA reduces the influence of text similarity between questions and options on model inference and enhances the model’s generalization ability during routine testing. Experimental results on two widely used datasets demonstrate the effectiveness of our approach.

Despite the remarkable advancements in Visual Question Answering (VQA), the challenge of mitigating the language bias introduced by textual information remains unresolved. Previous approaches capture language bias from a coarse-grained perspective. However, the finer-grained information within a sentence, such as context and keywords, can result in different biases. Due to the ignorance of fine-grained information, most existing methods fail to sufficiently capture language bias. In this paper, we propose a novel causal intervention training scheme named CIBi to eliminate language bias from a finer-grained perspective. Specifically, we divide the language bias into context bias and keyword bias. We employ causal intervention and contrastive learning to eliminate context bias and improve the multi-modal representation. Additionally, we design a new question-only branch based on counterfactual generation to distill and eliminate keyword bias. Experimental results illustrate that CIBi is applicable to various VQA models, yielding competitive performance.

Cross-domain Relation Extraction aims to transfer knowledge from a source domain to a different target domain to address low-resource challenges. However, the semantic gap caused by data bias between domains is a major challenge, especially in few-shot scenarios. Previous work has mainly focused on transferring knowledge between domains through shared feature representations without analyzing the impact of each factor that may produce data bias based on the characteristics of each domain. This work takes a causal perspective and proposes a new framework CausalGF. By constructing a unified structural causal model, we estimating the causal effects of factors such as syntactic structure, label distribution,and entities on the outcome. CausalGF calculates the causal effects among the factors and adjusts them dynamically based on domain characteristics, enabling adaptive gap filling. Our experiments show that our approach better fills the domain gap, yielding significantly better results on the cross-domain few-shot relation extraction task.

Artificial intelligence-based virtual assistants make people's daily life more convenient. However, the utterances of language-impaired people are limited and different in characteristics and domains from that of ordinary people. So it is difficult for language-impaired people to benefit from standard data-driven artificial intelligence algorithms. In this paper, we propose a multi-task training method for the dialogue state tracking (DST) task in dialogue systems that make up virtual assistants, improving the performance of T5 on the few-shot cross-domain DST task. Specifically, we consider two ways of handling DST task: predicting the dialogue state from the beginning or updating the dialogue state every turn, and accordingly design the main task and auxiliary task for the model. Experiments show that our method outperforms most previous works on the MultiWOZ 2.0 and 2.1 datasets for the few-shot cross-domain DST task. For the artificial-crafted language-impaired dataset, our method can effectively improve the few-shot cross-domain performance of the model. Additionally, we analyzed the possible reason why this multitasking approach works well.

Multiple-choice question answering(MCQA) is one of the most challenging tasks in machine reading comprehension. MCQA task requires selecting the most appropriate answer from several relevant options for a given question. In recent years, many works have concentrated on designing models from the perspective of using the information of the question and options at a large granularity level. However, few studies have explored how the model uses the information to find the correct answer at a fine granularity level or a multi-granularity level. This paper proposed a multi-granularity representation enhancement method to use information from different granularities. The method introduces large-grained candidate option information into the question to guide the selection of fine-grained critical information and facilitate the information interaction between the answer and the question which is in line with the human reasoning processes. Experimental results show that the method proposed in this paper can effectively improve the accuracy of MCQA tasks without introducing external knowledge.

Multiple-choice question answering (MCQA) is a challenging task that requires selecting the correct answer from a set of options based on a given question. There is a trend to use pre-trained encoder-decoder models to solve MCQA. Previous works concentrate on the decoder and adopt the generated text to enhance model performance. However, few studies have optimized the use of encoders for the characteristics of MCQA. In this work, we propose a dynamic exclusion model for MCQA named ExcMC, which mimics human thinking in selection. It dynamically eliminates several incorrect options to optimize the encoder usage. ExcMC outperforms existing comparable works on two widely-used MCQA datasets, demonstrating the effectiveness of our model.

Machine Reading Comprehension (MRC) is a challenging task in natural language understanding, especially multi-hop question answering (QA) in long texts. One of the challenges in multi-hop QA requires models to produce interpretable answers based on evidence that is selected from a given long text. Based on the Retriever-Reader architecture, existing work tackles this problem by using different methods to exploit various evidence information. To better use evidence information, we propose a loss function considering answer groups, which improves the reasoning ability of the reader in the Retriever-Reader architecture. Besides, we introduce the relevance constraint factor containing evidence information to improve the reader’s ability of locating key sentences. Evaluated on the HotpotQA dataset, the proposed methods achieve improvement, demonstrating the effectiveness of our methods and the importance of evidence information.

The spread of rumors along with breaking events seriously hinders the truth in the era of social media. Previous studies reveal that due to the lack of annotated resources, rumors presented in minority languages are hard to be detected. Furthermore, the unforeseen breaking events not involved in yesterday's news exacerbate the scarcity of data resources. In this work, we propose a novel zero-shot framework based on prompt learning to detect rumors falling in different domains or presented in different languages. More specifically, we firstly represent rumor circulated on social media as diverse propagation threads, then design a hierarchical prompt encoding mechanism to learn language-agnostic contextual representations for both prompts and rumor data. To further enhance domain adaptation, we model the domain-invariant structural features from the propagation threads, to incorporate structural position representations of influential community response. In addition, a new virtual response augmentation method is used to improve model training. Extensive experiments conducted on three real-world datasets demonstrate that our proposed model achieves much better performance than state-of-the-art methods and exhibits a superior capacity for detecting rumors at early stages.

Traditional document summarization models perform less satisfactorily on dialogues due to the complex personal pronouns referential relationships and insufficient modeling of conversation. To address this problem, we propose a novel end-to-end Transformer-based model for abstractive dialogue summarization with Relation Enhanced method based on BART named RE-BART. Our model leverages local relation and global relation in a conversation to model dialogue and to generate better summaries. In detail, we consider that the verb and related arguments in a single utterance contribute to the local event for encoding the dialogue. And coreference information in a whole conversation represents the global relation which helps to trace the topic and information flow of the speakers. Then we design a dialogue relation enhanced model for modeling both information. Experiments on the SAMsum dataset show that our model outperforms various dialogue summarization approaches and achieves new state-of- the-art ROUGE results.

Despite the impressive improvements of Visual Question Answer (VQA), it still remains a challenge of how to avoid the suffering of spurious correlations from textual content to answer. Previous researches have shown that due to the existence of language bias in the VQA dataset, VQA models may tend to capture superficial statistical correlation and suffer from the poor generalization capability in the out-of-distribution data. To alleviate the biases caused by language modality, we propose a method of context augmentation and adaptive loss adjustment, which can alleviate shortcut learning behavior of VQA models. Specifically, the existence of language bias is due to the high co-occurrence frequency of categories and the words in “Question”, therefore, we propose to use “Paraphrase Generation” to produce paraphrases with diverse contexts, so as to mitigate such correlation. Secondly, we use adaptive loss adjustment to adjust the importance of samples, that is, reduce the importance of bias-aligned samples and improve the importance of bias-conflicting samples, so as to guide the model to capture the intrinsic attributes that are beneficial to generalization. The experiments have demonstrated the feasibility and validity of our method on a variety of VQA models.

在自然语言处理领域,全局注意力机制通过考虑编码器的所有隐藏状态来捕获信息,从而帮助预测结果.然而在理解汉语成语这种复杂的语言现象时,模型往往会因特定语境的干扰而产生错误的决定和认知.因此,为了使模型更好地感知成语在不同语境下的语法功能,该文提出了一种增强型的全局注意力机制,通过对每个位置空间产生额外的注意因子来调整原始的全局注意力,最终提高了模型对特定语义的学习能力.该文将增强型全局注意力机制与BERT语言模型相结合,设计了一个用于完形填空任务的模型,并在最近发布的中文成语完形填空数据集ChID上进行了实验.结果表明,相比于传统的BERT模型和全局注意模型,该模型取得的效果更优.

论文针对答案选择任务进行研究,利用深层神经网络结合外部知识库信息的方法,提出基于知识库关系信息的双向长短时记忆网络(Knowledge Based Relation-Bidirectional Long Short Term Memory,KBR-BiLSTM),引入知识库中实体信息及实体关系信息去优化基准模型中的注意力机制;并利用知识库关系信息结合上下文丰富了问答的句子编码信息,提升模型效果.在维基问答(Wiki QA)数据集和TREC QA数据集上进行对比实验,证明了KBR-BiLSTM模型的有效性.

Multi-label text classification is an essential task in natural language processing. Existing multi-label classification models generally consider labels as categorical variables and ignore the exploitation of label semantics. In this paper, we view the task as a correlation-guided text representation problem: an attention-based two-step framework is proposed to integrate text information and label semantics by jointly learning words and labels in the same space. In this way, we aim to capture high-order label-label correlations as well as context-label correlations. Specifically, the proposed approach works by learning token-level representations of words and labels globally through a multi-layer Transformer and constructing an attention vector through word-label correlation matrix to generate the text representation. It ensures that relevant words receive higher weights than irrelevant words and thus directly optimizes the classification performance. Extensive experiments over benchmark multi-label datasets clearly validate the effectiveness of the proposed approach, and further analysis demonstrates that it is competitive in both predicting low-frequency labels and convergence speed.