上海品茶

三个皮匠报告
上海品茶
最新
合集
英文
会议
图表
入驻
付费
new
报告
报告 合集 图表

探索图可解释性中的分布外泛化问题.pdf

编号：144840 PDF 35页 2.41MB 下载积分：VIP专享
下载报告请您先登录！

探索图可解释性中的分布外泛化问题.pdf

1、探索图可解释性中的分布外泛化问题方俊峰 中国科学技术大学 博士DataFunSummit#2023当前的可解释评估指标真的“公平”吗？可解释算法为何会引入OOD问题？如何实现网络-数据的联合解释？1)避开公式 2)中英混杂2Background3How to define explainability?(1)Miller,Tim.“Explanation in artificial intelligence:Insights from the social sciences.”arXiv Preprint arXiv:1706.07269.(2017).Explainability is th

2、e degree to which a human can understand the models result 1Find which fractions are most influential to the GNNs predictionFind important subgraphInputModelCycleCycleGridHouseBA3-motifExplanationMethodExisting methodsSAGradCAMGNNExplainerPGExplainerCF-GNNExGraphMaskGSATDIR*GREA*m=0.03m=0.94GemPGMEx

3、6Evaluation metricNIPS 2023 Oral(2%)Evaluating Post-hoc Explanations for Graph Neural Networks via Robustness Analysis.Existing Evaluation metrics1.Human supervision seeks to justify whether the explanations align with human knowledge,but it is often too subjective,thus hardly providing quantifiable

4、 assessments.2.Measuring the agreement between the generated and ground-truth explanations,such as Precision and Recall.Unfortunately,access to the ground truth is usually unavailable and labor-extensive.CycleGridExisting Evaluation metrics#Caveat of RM:as the after-removal subgraphs are likely to l

5、ie off the distribution of full graphs,the GNN is forced to handle these off-manifoldinputs and easily gets erroneous predictions.3.Feature Removal(RM):first remove the unimportant features and feed the remaining part(i.e.,explanatory subgraph)into the GNN,and then observe how the prediction changes

6、.Existing Evaluation metrics4.Generation-based metrics:Instead of directly feeding,they use a generative model to generate a new full graph conditioned on the subgraph.#Caveat of Generation-based metrics:the generation-based metrics show respectto the data distribution somehow but couldbe inconsiste

7、nt with GNNs behavior andlose control of the infilling part.Existing Evaluation metricsCan we design a metric that respects the data distribution and GNN behavior simultaneously?STEP 1:Formulation of Adversarial Robustness.通俗点说：如何判定你给定的解释好不好？如果好，那么我随便扰动除解释以外的边，输入不应该翻转。因此，解释的鲁棒性可以定义为：致使输出翻转的最小扰动。STEP

8、 2:Finding a Tractable Objective.Two observations:Perturbations may fail to find any adversarial example.Current attack methods typically turn to find an alternative sub-optimal solution.Lagrange duality:通俗点：那么我每次只扰动部分除解释外的边，看输出变化STEP 3:OOD Reweighting BlockExplanatory Subgraph（May violate the data

9、distribution）The degree of Violationhow to ensure that the aforementioned evaluation process respects the data distribution?通俗点：重建越失败，违反原始数据流行越大Overall Framework1234SimOARRemove grey dots!通俗点（KEY POINT）：不要直接输入子图，而是：固定子图，按比例k 扰动其他边，看平均输出的变化。ResultsResults19Explanation Method Cooperative Explanations

10、of Graph Neural Networks WSDM 2023Revealing the OOD Limitation通俗点：直接将Gs输入模型不正确，那么能不能找到一个适合Gs的模型呢？Revealing the OOD Limitation通俗点：直接将Gs输入受训于G的模型不正确，那么能不能找到一个受训于（适合）Gs的模型呢？可是，GsffGs，死循环了CGE：CGE：Theoretical BasisCGE：Lottory ticket hypothesisExplainerLTH in GNNModelMaskMask ModeltrainMask.How to do:gene

11、rate mask train rewindFor what:iteratively select subnetwork Explainer:GNNExplainer.How to do:generate mask train rewindFor what:iteratively select subgraph Incorporate LTH and GNNExplainerGraph：generate mask train for MMI（not label）rewindNetwork：generate mask train for MMI（not label）rewindTrain for

12、 MMI Cooperative GNN Explanation(CGE)Employing CGE to optimize existing methodsexisting methods:generate importance scores for each edgeTrain for MMI existing methodsgenerate importance scores M1Incorporate LTH and ExplainerTrickParameters simplification:we force the masks of neurons in one convolutional unit to the same value.Early stopping:Algo30ExperimentQuantitative&Qualitative analysisQuantitative analysisQualitative analysis33Theory in GNN ExplanationOn Regularization for Explaining Graph Neural Networks:An Information Theory Perspective.https:/ Why OOD?Thanks!

友情提示

1、下载报告失败解决办法
2、PDF文件下载后，可能会被浏览器默认打开，此种情况可以点击浏览器菜单，保存网页到桌面，就可以正常下载了。
3、本站不支持迅雷下载，请使用电脑自带的IE浏览器，或者360浏览器、谷歌浏览器下载即可。
4、本站报告下载后的文档和图纸-无水印,预览文档经过压缩，下载后原文更清晰。

本文（探索图可解释性中的分布外泛化问题.pdf）为本站 （2200） 主动上传，三个皮匠报告文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。 若此文所含内容侵犯了您的版权或隐私，请立即通知三个皮匠报告文库（点击联系客服），我们立即给予删除！

温馨提示：如果因为网速或其他原因下载失败请重新下载，重复下载不扣分。
客服
商务合作
小程序
服务号
会员动态
会员动态 会员动态：

 152**94...  升级为高级VIP wei**n_...  升级为标准VIP

wei**n_... 升级为标准VIP  185**27... 升级为标准VIP  

135**37... 升级为至尊VIP 159**71... 升级为高级VIP

 139**27... 升级为至尊VIP wei**n_... 升级为高级VIP  

wei**n_...  升级为高级VIP  188**66... 升级为标准VIP 

wei**n_... 升级为至尊VIP   wei**n_...  升级为高级VIP

wei**n_...   升级为至尊VIP wei**n_... 升级为高级VIP 

wei**n_... 升级为高级VIP    wei**n_... 升级为至尊VIP

 177**81... 升级为标准VIP  185**22... 升级为标准VIP 

 138**26... 升级为至尊VIP  军歌 升级为至尊VIP

 159**75...  升级为至尊VIP wei**n_... 升级为标准VIP 

  wei**n_... 升级为至尊VIP wei**n_...  升级为高级VIP

su2**62... 升级为至尊VIP  wei**n_...  升级为至尊VIP

wei**n_... 升级为至尊VIP    186**35... 升级为高级VIP 

 186**21... 升级为标准VIP wei**n_... 升级为标准VIP

 wei**n_... 升级为标准VIP   wei**n_...  升级为标准VIP

137**40... 升级为至尊VIP  wei**n_...  升级为至尊VIP

186**37... 升级为至尊VIP 177**05...  升级为至尊VIP

 wei**n_...  升级为高级VIP  wei**n_... 升级为至尊VIP 

wei**n_...  升级为至尊VIP  wei**n_... 升级为标准VIP

 wei**n_... 升级为高级VIP  155**91...  升级为至尊VIP

 155**91... 升级为标准VIP  177**25... 升级为至尊VIP

139**88... 升级为至尊VIP    wei**n_... 升级为至尊VIP

wei**n_...  升级为高级VIP  wei**n_... 升级为标准VIP

 135**30... 升级为标准VIP  wei**n_... 升级为高级VIP

138**62...  升级为标准VIP 洛宾 升级为高级VIP 

wei**n_...  升级为标准VIP   wei**n_... 升级为高级VIP

wei**n_... 升级为标准VIP  180**13... 升级为高级VIP 

wei**n_... 升级为至尊VIP  152**69... 升级为标准VIP 

152**69...  升级为标准VIP  小**...  升级为标准VIP

 wei**n_... 升级为标准VIP 138**09... 升级为标准VIP 

wei**n_...   升级为至尊VIP  邓** 升级为标准VIP

 wei**n_... 升级为标准VIP wei**n_...   升级为至尊VIP

186**22... 升级为标准VIP  微**...  升级为至尊VIP

wei**n_... 升级为至尊VIP  zhh**_s... 升级为标准VIP 

 wei**n_... 升级为至尊VIP  wei**n_... 升级为至尊VIP

 wei**n_... 升级为高级VIP  wei**n_... 升级为至尊VIP 

 131**00... 升级为高级VIP wei**n_... 升级为高级VIP

188**05...  升级为至尊VIP  139**80... 升级为至尊VIP 

wei**n_... 升级为高级VIP 173**11...   升级为至尊VIP

 152**71... 升级为高级VIP 137**24... 升级为至尊VIP  

wei**n_...  升级为高级VIP  185**31...  升级为至尊VIP 

186**76... 升级为至尊VIP   wei**n_... 升级为标准VIP

wei**n_... 升级为标准VIP    138**50...  升级为标准VIP

wei**n_... 升级为高级VIP  wei**n_... 升级为高级VIP 

  wei**n_... 升级为标准VIP wei**n_...  升级为至尊VIP

 Bry**-C... 升级为至尊VIP   151**85... 升级为至尊VIP

 136**28... 升级为至尊VIP  166**35...  升级为至尊VIP

狗**... 升级为至尊VIP  般若 升级为标准VIP

wei**n_...  升级为标准VIP 185**87... 升级为至尊VIP