| 0 | 0 | 16 |
| 下载次数 | 被引频次 | 阅读次数 |
研究生学位论文评审意见是研究生培养的重要反馈，利用方面级情感分析方法（Aspect-based Sentiment Analysis, ABSA）对研究生学位论文评审意见进行分析对研究生教育质量提升具有重要意义。方面级情感四元组预测（Aspect Sentiment Quad Prediction, ASQP）作为ABSA中最具挑战性的任务，旨在同时预测文本中的方面词、方面类别、观点词和情感极性。为利用ABSA方法对研究生学位论文评审意见进行结构化分析，本文构建了研究生学位论文评审意见ASQP数据集并设计了基于离散扩散模型的ASQP方法。传统生成式ASQP方法多基于自回归模型，模型生成时逐词依赖前文，不利于方面级情感四元组的预测。本文提出的离散扩散模型ASQP预测方法通过全局条件建模实现四元组预测，结合ASQP条件生成模板以利用预训练模型的语义知识，并引入蒙特卡洛采样、筛选与投票机制提升生成稳定性。实验结果显示，本文提出的方法在ASQP任务中在预测精确率方面表现出明显优势，同时在整体性能上与主流方法达到可比水平。
Abstract：Review comments on graduate theses and dissertations provide important feedback for postgraduate education, and analyzing these review comments using Aspect-Based Sentiment Analysis(ABSA) is of great significance for improving the quality of graduate education. As the most challenging task in ABSA, Aspect Sentiment Quad Prediction(ASQP) aims to simultaneously identify aspect terms, aspect categories, opinion terms, and sentiment polarities in text. To enable the structured analysis of review comments on graduate theses and dissertations using ABSA, this study constructs an ASQP dataset from such comments and proposes a discrete diffusion model for ASQP. Existing generative ASQP methods are predominantly based on autoregressive models, which generate tokens sequentially by relying on previously generated tokens, making the prediction of aspect sentiment quadruples less effective. The proposed discrete diffusion-based ASQP method performs quadruple prediction through global conditional modeling, incorporates an ASQP conditional generation template to exploit the semantic knowledge of pretrained language models, and further improves generation stability by introducing Monte Carlo sampling, filtering, and voting mechanisms. Experimental results demonstrate that the proposed method achieves a clear advantage in prediction precision on the ASQP task while maintaining overall performance comparable to that of mainstream methods.
[1]陆根书,李珍艳,王玺.大数据分析在研究生教育质量评价中的应用探析[J].北京航空航天大学学报(社会科学版),2020,33(3):118-125.
[2]宋晨虎,张玲.硕士研究生学位论文质量管理及对策研究[J].吉林工程技术师范学院学报,2019,35(4):39-41.
[3]高耀,杨佳乐.“存在问题”专业硕士学位论文中的典型问题--基于Y市论文抽检同行专家文字评审意见的分析[J].教育科学,2017,33(3):66-71.
[4]李敏,陈洪捷.不合格学术型硕士研究生学位论文的典型特征--基于论文抽检专家评阅意见的分析[J].学位与研究生教育,2017,(6):50-55.
[5]程仕平,陈明,殷悦.AHP层次分析与K-means聚类相结合的博士学位论文评价指标权重确定方法[J].创新与创业教育,2021,12(5):72-76.
[6]罗小丹.“差评”硕士学位论文的主要特点与管理对策研究:基于S大学盲审专家评语的内容分析[D].湖南师范大学,2020:29.
[7]Zhang W,Li X,Deng Y,et al.A survey on aspect-based sentiment analysis:tasks,methods,and challenges[J].IEEE Transactions on Knowledge and Data Engineering,2022,35(11):11019-11038.
[8]Nazir A,Rao Y,Wu L,et al.Issues and challenges of aspect-based sentiment analysis:a comprehensive survey[J].IEEE Transactions on Affective Computing,2022,13(2):845-863.
[9]陈壮,钱铁云,李万理,等.低资源方面级情感分析研究综述[J].计算机学报,2023,46(7):1445-1472.
[10]Cai H,Xia R,Yu J.Aspect-category-opinion-sentiment quadruple extraction with implicit aspects and opinions[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (volume 1:long papers).2021:340-350.
[11]Devlin J,Chang M W,Lee K,et al.Bert:pre-training of deep bidirectional transformers for language understanding[C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics:human Language Technologies,Volume 1 (Long and Short Papers).2019:4171-4186.
[12]彭文忠,夏家莉,万齐智,等.基于完形填空的方面级情感四元组预测[J].计算机学报,2024,47(8):1744-1768.
[13]Zhang W,Deng Y,Li X,et al.Aspect sentiment quad prediction as paraphrase generation[C]//Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing.2021:9209-9219.
[14]Mao Y,Shen Y,Yang J,et al.Seq2path:generating sentiment tuples as paths of a tree[C]//Findings of the association for computational linguistics:ACL 2022.2022:2215-2225.
[15]Gao T,Fang J,Liu H,et al.LEGO-ABSA:a prompt-based task assemblable unified generative framework for multi-task aspectbased sentiment analysis[C]//Proceedings of the 29th International Conference on Computational Linguistics.2022:7002-7012.
[16]Gou Z,Guo Q,Yang Y.Mv P:multi-view prompting improves aspect sentiment tuple prediction[C]//Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers).2023:4380-4397.
[17]Vaswani A,Shazeer N,Parmar N,et al.Attention is all you need[C]//Advances in Neural Information Processing Systems,2017,30.
[18]Raffel C,Shazeer N,Roberts A,et al.Exploring the limits of transfer learning with a unified text-to-text transformer[J].Journal of Machine Learning Research,2020,21(140):1-67.
[19]Kalyan K S,Rajasekharan A,Sangeetha S.Ammus:a survey of transformer-based pretrained models in natural language processing[PP/OL].ar Xiv (2021-08-11)[2026-10-15].https://arxiv.org/abs/2108.05542.
[20]Bao X,Wang Z,Jiang X,et al.Aspect-based sentiment analysis with opinion tree generation[C]//Proceedings of the 31st International Joint Conference on Artificial Intelligence.Vienna:IJCAI,2022:4044-4050.
[21]Austin J,Johnson D D,Ho J,et al.Structured denoising diffusion models in discrete state-spaces[J].Advances in Neural Information Processing Systems,2021,34:17981-17993.
[22]Yi Q,Chen X,Zhang C,et al.Diffusion models in text generation:a survey[J].Peer J Computer Science,2024,10:e1905.
[23]Li T,Chen M,Guo B,et al.A survey on diffusion language models[PP/OL].ar Xiv (2025-08-01)[2026-10-15].https://arxiv.org/abs/2508.10875.
[24]Lin Z,Gong Y,Shen Y,et al.Text generation with diffusion language models:a pre-training approach with continuous paragraph denoise[C]//International Conference on Machine Learning.PMLR,2023:21051-21064.
[25]Zou H,Kim Z M,Kang D.A survey of diffusion models in natural language processing[PP/OL].ar Xiv (2023-05-23)[2026-10-15].https://arxiv.org/abs/2305.14671.
[26]Xiao Y,Wu L,Guo J,et al.A survey on non-autoregressive generation for neural machine translation and beyond[J].IEEETransactions on Pattern Analysis and Machine Intelligence,2023,45(10):11407-11427.
[27]Gao T,Fisch A,Chen D.Making pre-trained language models better few-shot learners[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the11th International Joint Conference on Natural Language Processing(Volume 1:Long Papers),2021:3816-3830.
[28]Hoogeboom E,Nielsen D,Jaini P,et al.Argmax flows and multinomial diffusion:learning categorical distributions[C]//Advances in Neural Information Processing Systems,2021,34:12454-12465.
[29]Zheng L,Yuan J,Yu L,et al.A reparameterized discrete diffusion model for text generation[PP/OL].ar Xiv(2023-02-11)[2026-10-15].https://arxiv.org/abs/2302.05737.
[30]He Z,Sun T,Tang Q,et al.Diffusionbert:improving generative masked language models with diffusion models[C]//Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers),2023:4521-4534.
[31]Pontiki M,Galanis D,Papageorgiou H,et al.Semeval-2015 task12:aspect based sentiment analysis[C]//Proceedings of the 9th International Workshop on Semantic Evaluation (Sem Eval 2015).Denver:Association for Computational Linguistics,2015:486-495.
[32]Pontiki M,Galanis D,Papageorgiou H,et al.Semeval-2016 task5:aspect based sentiment analysis[C]//Proceedings of the 10th International Workshop on Semantic Evaluation (Sem Eval-2016),2016:19-30.
[33]Zhang W,Li X,Deng Y,et al.Towards generative aspect-based sentiment analysis[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing(volume 2:short papers),2021:504-510.
[34]Hu M,Wu Y,Gao H,et al.Improving aspect sentiment quad prediction via template-order data augmentation[C]//Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing,2022:7889-7900.
基本信息:
中图分类号:TP391.1;TP18;G643.8
引用信息:
[1]周琛节,卢笑,周炫余,等.基于离散扩散模型的研究生学位论文评审意见方面级情感四元组预测方法[J].教育智能化前沿,2026,39(04):21-30+39.
基金信息:
2025年度湖南省自然科学基金面上项目“基于多模态大模型的中小学生篮球运动身体对抗能力评价与适应性反馈研究”（项目编号：2025JJ50362）; 2026年度湖南省自然科学基金面上资助项目“基于多模态数据支持的中小学生科学探究能力智能化测评与个性化反馈研究”（项目编号：2026JJ50046）; 2025年度湖南省教育厅资助项目“大模型赋能的科学职前教师科学探究与实验操作技能自动评价方法研究”（项目编号：25A0109）
2026-08-28
2026-08-28