In this paper we consider the problem of collectively classifying entities where relational information is available across the entities. In practice inaccurate class distribution for each entity is often available from another (external) classifier. For example this distribution could come from a classifier built using content features or a simple dictionary. Given the relational and inaccurate external classifier information, we consider two graph based settings in which the problem of collective classification can be solved. In the first setting the class distribution is used to fix labels to a subset of nodes and the labels for the remaining nodes are obtained like in a transductive setting. In the other setting the class distributions of all nodes are used to define the fitting function part of a graph regularized objective function. We define a generalized objective function that handles both the settings. Methods like harmonic Gaussian field and local-global consistency (LGC) reported in the literature can be seen as special cases. We extend the LGC and weighted vote relational neighbor classification (WvRN) methods to support usage of external classifier information. We also propose an efficient least squares regularization (LSR) based method and relate it to information regularization methods. All the methods are evaluated on several benchmark and real world datasets. Considering together speed, robustness and accuracy, experimental results indicate that the LSR and WvRN-extension methods perform better than other methods.
Uncertainty Calibration for Deep Audio Classifiers
Ping An Technology (Shenzhen) Co., Ltd.（平安科技（深圳）有限公司）；University of Science and Technology of China（中国科学技术大学）
尽管深度神经网络（DNN）在音频分类任务中取得了巨大的成功，但其不确定性校准仍处于探索阶段。一个经过良好校准的模型在确定其预测时应该是准确的，而在可能不准确时应该显示出高度的不确定性。在这项工作中，我们研究了深度音频分类器的不确定性校准。特别是，我们经验性地研究了流行的校准方法的性能：（i）Monte Carlo Dropout，（ii）合奏，（iii）焦点损失和（iv）频谱归一化高斯过程（SNGP），在音频分类数据集上。为此，我们评估（i-iv）的环境声音和音乐流派分类的任务。结果表明，未校准的深度音频分类器可能过于自信，SNGP表现最好，在本文的两个数据集上非常有效。
Robust Event Classification Using Imperfect Real-world PMU Data
University of Nevada Reno（内华达大学里诺分校）；California State University San Bernardino（加州州立大学圣贝纳迪诺分校）；Virginia Polytechnic Institute and State University（弗吉尼亚理工大学）；IBM Thomas J. Watson Research Center（IBM托马斯·J·沃森研究中心）；Arizona State University（亚利桑那州立大学）
forgeNet: A graph deep neural network model using tree-based ensemble classifiers for feature extraction
Emory University（埃默里大学）
组学数据预测模型构建中的一个独特挑战是样本数量少（$(n)$），而特征数量多（$(p)$）。这种“$n\ll p$“属性给使用深度学习技术的疾病结果分类带来了困难。通过结合外部基因网络信息（例如图形嵌入式深度前馈网络（GEDFN）模型）进行稀疏学习是解决这个问题的一种方法。然而，这样的方法需要现有的特征图，并且特征图的潜在错误指定可能对分类和特征选择有害。为了解决这一限制，并开发一个强大的分类模型，而不依赖于外部知识，我们提出了一个\underline{for} est \underline{e}嵌入式深度前馈\underline{net}工作（forgeNet）模型，集成GEDFN架构与森林特征图提取器，使特征图可以在监督的方式学习，并专门为给定的预测任务构建。为了验证该方法的能力，我们用合成数据集和真实数据集对forgeNet模型进行了实验。由此产生的高分类精度表明，该方法是对组学数据稀疏深度学习模型的有价值的补充。
Virtual vs. Reality: External Validation of COVID-19 Classifiers using XCAT Phantoms for Chest Computed Tomography
Duke University School of Medicine（杜克大学医学院）；Duke University（杜克大学）；National Artificial Intelligence Institute, U.S. Dept. of Veterans Affairs（美国退伍军人事务部国家人工智能研究所）
Using a Classifier Ensemble for Proactive Quality Monitoring and Control: the impact of the choice of classifiers types, selection criterion, and fusion process
Université de Lorraine（洛林大学）；CNRS（法国国家科学研究中心）；ACTA Mobilier
Bayesian Time-Series Classifier for Decoding Simple Visual Stimuli from Intracranial Neural Activity
Isfahan University of Technology（伊斯法罕理工大学）；Worcester Polytechnic Institute（伍斯特理工学院）；Massachusetts General Hospital（麻省总医院）；Harvard Medical School（哈佛医学院）
