A Survey of Deep Active Learning 1.Introduction DeepAL产生的背景: DL: DL具有很强的学习能力: 源于其复杂的结构 需要大量的已标注数据去完成 相比很多传统机器学习算法,DL在大部分应用领域都有绝对的优势。 Limitation: 因为需要大量的标注样本,而有些专业领域的标注成本很高 AL: focuses on the study of data...
Continual Lifelong Learning and Catastrophic Forgetting 6. Future Prospective and Unanswered Questions 当前医疗图像分割目的在于用尽可能少量的带标注数据,获取尽可能好的效果。 Active Learning 假定存用于标注的用户接口,但仅与要标注的数据有关。Refinement 假设我们能与当前模型进行迭代式的交互,生成更为准确的标注。
A Survey of Deep Active Learning. ACM Comput. Surv. 2022 paper bib Pengzhen Ren, Yun Xiao, Xiaojun Chang, Po-Yao Huang, Zhihui Li, Brij B. Gupta, Xiaojiang Chen, Xin Wang A Survey of Deep Learning for Data Caching in Edge Network. Informatics 2020 paper bib Yantong Wang, Vasilis Frid...
论文(Improving information extraction by acquiring external evidence with reinforcement learning) 将信息提取任务建模为马尔科夫决策过程(MDP)该过程动态的使用了实体预测任务, 并提供了一组自动生成的替代方案中选择下一个 query 的方法. 模型流程包含从 发出搜索查询, 从新来源中提取, 识别获得的特征, 然后重复该过...
Due to the recent advancements in computing power, deep learning-based algorithms have become most effective and efficient choice of algorithms for recognizing and solving HAR problems. In this survey, we categorize recent research work with respect to various factors and measures to investigate the ...
2.A Survey on Self-supervised Learning: Algorithms, Applications, and Future Trends.Jie Gui, Tuo Chen, Jing V. R. de Sa, “Learning classification with unlabeled data,” inNeural Inf. Process. Syst., pp. 112–119, 1994 Devlin, Jacob et al. “BERT:Pre-trainingof Deep Bidirectional Transf...
Journal of Artificial Intelligence Research 2018 paper bib Albert Gatt,Emiel Krahmer Information Extraction A Survey of Deep Learning Methods for Relation Extraction. arXiv 2017 paper bib Shantanu Kumar A Survey of Event Extraction From Text. IEEE 2019 paper bib Wei Xiang, Bang Wang A Survey of ...
Named Entity Recognition is the research foundation of many Natural Language Processing sub-tasks. Named Entity Recognition for Chinese social media is to identify entity nouns such as person names, place names, and organization names in Chinese Social M
2. Overview of deep learning methods 本部分的目的是对我们在本survey的医学图像分析论文中发现的深度学习概念,技术和体系结构进行正式介绍和定义。 2.1. Learning algorithms 机器学习方法虽然有很多细微差别,但通常分为有监督的学习算法和无监督的学习算法。 在监督学习中,模型由输入数据\boldsymbol{x}和标签对y构...
deep learning; fusion method; multimodal emotion recognition; survey1. Introduction Emotion recognition is an important area of research because it enables computers to accurately comprehend human emotions and provide intelligent responses to meet human requirements. In educational settings, emotions can ...