Toward Relational Learning with Misinformation Liang Wu * , Jundong Li * , Fred Morstatter + , Huan Liu * * Arizona State University + University of Southern California {wuliang, jundongl, huanliu}@asu.edu, morstatt@usc.edu Arizona State University Data Mining and Machine Learning Lab
Classification in Social Media • Relational learning aims to classify linked nodes in a graph (social networks) • Task: Classification • Feature: Attributes, Links Arizona State University Data Mining and Machine Learning Lab
Classification in Social Media: Our Task • Relational learning aims to classify linked nodes in a graph (social networks) • Task: Classification • Feature: Attributes, Links • Challenge: Data is Inaccurate Arizona State University Data Mining and Machine Learning Lab
Social Media Data is Inaccurate and Noisy • Attacks of content polluters – Node attributes cannot reveal the identity • Colloquial language of regular users – Misinformation, inaccurate data Arizona State University Data Mining and Machine Learning Lab
Classification with Noisy Data • Weighting Nodes Node Weights Classifier Weighted Learning ? • Anomalous points are lower weighted – Larger loss leads to smaller weights Arizona State University Data Mining and Machine Learning Lab
Classification with Noisy Social Media Data • Attacks of content polluters – Node attributes cannot reveal the identity • Colloquial language of regular users – Misinformation, inaccurate data Arizona State University Data Mining and Machine Learning Lab
Robust Classification with Network Information • Weighting Nodes with Centrality Node Weights Classifier Weighted Learning ? • Authoritative points are higher weighted – – Larger centrality leads to higher weights Arizona State University Data Mining and Machine Learning Lab
Denoising with Social Networks? • Links can be noisy • Obtaining all links (complete graph) is difficult Arizona State University Data Mining and Machine Learning Lab
Community Structures are More Robust Malicious User Arizona State University Data Mining and Machine Learning Lab
Community Structures are More Robust Community Detection Malicious Malicious User User Arizona State University Data Mining and Machine Learning Lab
Denoise with Community Structures Arizona State University Data Mining and Machine Learning Lab
Community Candidate Generation + Community Selection Arizona State University Data Mining and Machine Learning Lab
Community Candidate Generation + Community Selection 𝑶 𝟑 + λ 1 || w | | 2 𝐧𝐣𝐨 2 d n i ||𝐝 G j + λ 2 σ i=0 σ j=1 i || 2 c i 𝐲 𝐣 𝐱 − 𝐳 𝒋 𝐱,𝐝 𝒋=𝟐 𝑏𝑤𝑝𝑗𝑒 𝑝𝑤𝑓𝑠𝑔𝑗𝑢𝑢𝑗𝑜 𝑠𝑝𝑣𝑞 𝑀𝑏𝑡𝑡𝑝 𝐓𝐯𝐜𝐤𝐟𝐝𝐮 𝐮𝐩 𝒅 𝒋 = 𝑳 L 1 1 nor norm on on th the 𝒋 inter-group in p le level d: depth of hierarchy of Louvain method L 2 norm on on th the n i : number of groups on layer i in intra-group le level 𝐝 G j i : nodes of group j on layer i Arizona State University Data Mining and Machine Learning Lab
Optimization Optimize w 𝒏 c i 𝐲 𝐣 𝐱 − 𝐳𝐣 𝟑 + λ 1 || w | | 2 𝐧𝐣𝐨 2 𝐱 𝒋=𝟐 Optimize c 𝒏 𝐧𝐣𝐨 d n i ||𝐝 G j c i 𝑢 𝒋 + λ 2 σ i=0 σ j=1 i || 2 𝐱,𝐝 𝒋=𝟐 𝐓𝐯𝐜𝐤𝐟𝐝𝐮 𝐮𝐩 𝒅 𝒋 = 𝟐 𝒋 Arizona State University Data Mining and Machine Learning Lab
Evaluation Results Macro- and Micro-average of F 1 -measures with increasing ratio of misinformation Flickr Arizona State University Data Mining and Machine Learning Lab
More Results BlogCatalog Effectiveness of identifying mislabeled instances BlogCatalog Flickr Arizona State University Data Mining and Machine Learning Lab
Conclusions • A supervised learning method with inaccurate networked data – Focusing on community structures instead of links – Can be integrated to other algorithms – Efficient to solve Arizona State University Data Mining and Machine Learning Lab
Recommend
More recommend