To improve uncertainty quantification of variance networks, we propose a novel tree-structured local neural network model that partitions the feature space into multiple regions based on uncertainty heterogeneity. A tree is built upon giving the training data, whose leaf nodes represent different regions where region-specific neural networks are trained to predict both the mean and the variance for quantifying uncertainty. The proposed Uncertainty-Splitting Neural Regression Tree (USNRT) employs novel splitting criteria. At each node, a neural network is trained on the full data first, and a statistical test for the residuals is conducted to find the best split, corresponding to the two sub-regions with the most significant uncertainty heterogeneity. USNRT is computationally friendly because very few leaf nodes are sufficient and pruning is unnecessary. On extensive UCI datasets, in terms of both calibration and sharpness, USNRT shows superior performance compared to some recent popular methods for variance prediction, including vanilla variance network, deep ensemble, dropout-based methods, tree-based models, etc. Through comprehensive visualization and analysis, we uncover how USNRT works and show its merits.
translated by 谷歌翻译
增加片上光子神经网络(PNN)的层数对于改善其模型性能至关重要。但是,网络隐藏层的连续级联导致更大的集成光子芯片区域。为了解决此问题,我们提出了光学神经常规微分方程(ON-ON-ON-OD-ON-OD-ON-OD-ON-OD-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ODINE),该架构用光ODE求解器参数化了隐藏层的连续动力学。 On-Ode包括PNN,然后是光子积分器和光反馈回路,可以配置为代表残留的神经网络(RESNET)和复发性神经网络,并有效地降低了芯片面积占用率。对于基于干扰的光电非线性隐藏层,数值实验表明,单个隐藏层ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ON-ONE表示与图像分类任务中的两层光学重新系统大致相同。此外,Onode提高了基于衍射的全光线性隐藏层的模型分类精度。 On-Eod的时间依赖性动力学属性进一步应用于高精度的轨迹预测。
translated by 谷歌翻译
尽管条件变异自动编码器(CVAE)模型比传统的SEQ2SEQ模型可以产生更多的多样化响应,但响应通常与输入词的相关性低或与问题不合逻辑。进行因果分析以研究背后的原因,并提供了一种寻找调解人并减轻对话中混杂偏见的方法。具体而言,我们建议预测调解人,以保留相关信息,并自动将调解人纳入生成过程中。此外,动态主题图指导条件变异自动编码器(TGG-CVAE)模型用于补充语义空间并减少响应中的混杂偏置。广泛的实验表明,所提出的模型能够产生相关和信息性的响应,并且在自动指标和人类评估方面优于最先进的响应。
translated by 谷歌翻译
经济学和医疗保健方面的许多实际决策问题寻求从观察数据中估算平均治疗效果(ATE)。双重/辩护的机器学习(DML)是观察性研究中估计吃量的普遍方法之一。但是,DML估计器可能会遇到错误的问题,甚至在倾向分数被弄错或非常接近0或1时进行极端估计。现有文献从理论的角度解决了这个问题。在本文中,我们提出了一种健壮的因果学习(RCL)方法,以抵消DML估计量的缺陷。从理论上讲,RCL估计量i)与DML估计器一样一致且双重稳健,ii)可以摆脱错误混合问题。从经验上讲,全面的实验表明,i)RCL估计器比DML估计器给出了因果参数的稳定估计,ii)RCL估计器在模拟和基准标准数据集上应用不同的机器学习模型时,RCL估计器优于传统估计器及其变体。 。
translated by 谷歌翻译
采样是原始点云数据处理的重要组成部分,例如在流行的PointNet ++方案中。最远的点采样(FPS)是最流行的采样方案之一,最远的点采样(FPS)是最远的点并执行距离更新。不幸的是,它的效率低,并且可能成为点云应用的瓶颈。我们提出了由M参数化的可调节FPS(AFP),以积极地降低FPS的复杂性,而不会损害采样性能。具体而言,它将原始点云分为M小点云,并同时将样品M点分为M点。它利用了大约分类点云数据的尺寸局部性,以最大程度地减少其性能降解。 AFPS方法可以在原始FPS上实现22至30倍的速度。此外,我们提出了最近的点距离级别(NPDU)方法,以将距离更新数限制为常数数字。 AFPS方法上的NPDU组合可以在具有2K-32K点的点云上实现34-280X的加速,其算法性能与原始FPS相当。例如,对于Shapenet部件分割任务,它可以达到0.8490实例平均MIOU(联合平均交叉点),与原始FPS相比,它仅下降0.0035。
translated by 谷歌翻译
多光谱和全型图像的融合始终被称为pansharpening。大多数可用的基于深度学习的pan-sharpening方法通过一步方案增强了多光谱图像,这在很大程度上取决于网络的重建能力。但是,遥感图像总是具有很大的变化,因此,这些一步方法容易受到误差积累的影响,因此无法保留空间细节以及光谱信息。在本文中,我们提出了一个新型的两步模型,用于泛叠式模型,该模型通过空间和光谱信息的进行性补偿来锐化MS图像。首先,深层多尺度引导的生成对抗网络用于初步增强MS图像的空间分辨率。从粗糙域中的预交换MS图像开始,我们的方法随后逐步完善了具有反向体系结构的几个生成对抗网络(GAN)的空间和光谱残差。整个模型由三重gan组成,基于特定的架构,关节补偿损失函数旨在使三重甘族能够同时训练。此外,本文提出的空间谱系残留补偿结构可以扩展到其他泛伴式方法,以进一步增强其融合结果。在不同的数据集上进行了广泛的实验,结果证明了我们提出的方法的有效性和效率。
translated by 谷歌翻译
视频显着对象检测模型在像素密集注释上训练有素的训练有素,已经达到了出色的性能,但获得像素逐像素注释的数据集很费力。尚未探索几项作品,试图使用涂鸦注释来缓解这个问题,但是尚未探讨点监督作为一种更节省劳动的注释方法(即使是对密集预测的手动注释方法中最多的劳动方法)。在本文中,我们提出了一个基于点监督的强基线模型。为了使用时间信息来推断显着性图,我们分别从短期和长期角度挖掘了框架间的互补信息。具体而言,我们提出了一个混合令牌注意模块,该模块将光流和图像信息从正交方向混合在一起,自适应地突出了关键的光流信息(通道维度)和关键令牌信息(空间维度)。为了利用长期提示,我们开发了长期的跨框架注意模块(LCFA),该模块有助于当前框架基于多框架代币推断出显着对象。此外,我们通过重新标记Davis和DavSod数据集来标记两个分配的数据集P-Davis和P-Davsod。六个基准数据集的实验说明了我们的方法优于先前的最先进的弱监督方法,甚至与某些完全监督的方法相当。源代码和数据集可用。
translated by 谷歌翻译
少量学习(FSL)旨在解决数据稀缺问题。标准FSL框架由两个组件组成:(1)预先火车。采用基础数据以生成基于CNN的特征提取模型(FEM)。 (2)元测试。应用训练有素的有限元素以获取新的数据的特征并识别它们。 FSL严重依赖于FEM的设计。然而,各种有限元有明显的重点。例如,若干可以更关注轮廓信息,而其他人可以特别强调纹理信息。单个头功能只是样本的单面表示。除了跨域的负影响(例如,训练有素的有限元件无瑕疵地适应新颖的类),与地面真理分布相比,新型数据的分布可能具有一定程度的偏差,如分配转移 - 问题(DSP)。为了解决DSP,我们提出了多头功能协作(MHFC)算法,该算法试图将多头特征(例如,从各种FEM中提取的多个功能)投影到统一空间并融合它们以捕获更多辨别信息。通常,首先,我们介绍子空间学习方法来转换多头特征以对准低维表示。它通过学习具有更强大的歧视的功能来纠正DSP,并克服了来自不同头部特征的不一致测量尺度的问题。然后,我们设计注意力块以自动更新每个头部功能的组合权重。它全面考虑各种观点的贡献,进一步提高了特征的歧视。我们评估了五个基准数据集(包括跨域实验)的提出方法,与最先进的情况下实现了2.1%-7.8%的显着改善。
translated by 谷歌翻译
Causal learning is the key to obtaining stable predictions and answering \textit{what if} problems in decision-makings. In causal learning, it is central to seek methods to estimate the average treatment effect (ATE) from observational data. The Double/Debiased Machine Learning (DML) is one of the prevalent methods to estimate ATE. However, the DML estimators can suffer from an \textit{error-compounding issue} and even give extreme estimates when the propensity scores are close to 0 or 1. Previous studies have overcome this issue through some empirical tricks such as propensity score trimming, yet none of the existing works solves it from a theoretical standpoint. In this paper, we propose a \textit{Robust Causal Learning (RCL)} method to offset the deficiencies of DML estimators. Theoretically, the RCL estimators i) satisfy the (higher-order) orthogonal condition and are as \textit{consistent and doubly robust} as the DML estimators, and ii) get rid of the error-compounding issue. Empirically, the comprehensive experiments show that: i) the RCL estimators give more stable estimations of the causal parameters than DML; ii) the RCL estimators outperform traditional estimators and their variants when applying different machine learning models on both simulation and benchmark datasets, and a mimic consumer credit dataset generated by WGAN.
translated by 谷歌翻译
Siamese network based trackers formulate tracking as convolutional feature cross-correlation between a target template and a search region. However, Siamese trackers still have an accuracy gap compared with state-of-theart algorithms and they cannot take advantage of features from deep networks, such as ResNet-50 or deeper. In this work we prove the core reason comes from the lack of strict translation invariance. By comprehensive theoretical analysis and experimental validations, we break this restriction through a simple yet effective spatial aware sampling strategy and successfully train a ResNet-driven Siamese tracker with significant performance gain. Moreover, we propose a new model architecture to perform layer-wise and depthwise aggregations, which not only further improves the accuracy but also reduces the model size. We conduct extensive ablation studies to demonstrate the effectiveness of the proposed tracker, which obtains currently the best results on five large tracking benchmarks, including OTB2015, VOT2018, UAV123, LaSOT, and TrackingNet. Our model will be released to facilitate further researches.
translated by 谷歌翻译