Skip to main content
Uber logo

Schedule rides in advance

Reserve a rideReserve a ride

Schedule rides in advance

Reserve a rideReserve a ride
AI, Engineering

Fraud Detection: Using Relational Graph Learning to Detect Collusion

May 21, 2021 / Global
Featured image for Fraud Detection: Using Relational Graph Learning to Detect Collusion
Figure 1: An illustrated graph of connected drivers and riders. Red nodes represent fraud users, blue nodes represent legitimate users. Users are connected via shared information.
Figure 2 Modeling flow of RGCN: inputs of users’ node features and edge types are passed into several RGCN layers to generate node scores. The edge colors represent different edge types.
Figure 3: Fraud score modeling flow: score is learned via a binary cross-entropy loss. Model outputs are two scores, one for driver and one for rider.
Figure 4: Feature importance of the features used in the downstream risk model: user score1 and user score0 learned from RGCN are ranked to be the 4th and the 39th respectively.
Figure 5: Data pipeline (top row) and training pipeline (bottom row) used to learn fraud scores for improving fraud detection.
Xinyu Hu

Xinyu Hu

Xinyu Hu is a Senior Research Scientist at Uber AI focused on large-scale machine learning applications in spatial-temporal problems and causal inference. Xinyu holds a Ph.D. in Biostatistics from Columbia University.

Chengliang Yang

Chengliang Yang

Chengliang is a machine learning engineer in the Uber Risk engineering team. He primarily focuses on building machine learning based solutions to identify fraudulent behaviors on the Uber platform. Chengliang holds a PhD in Computer Science from the University of Florida with specialization in machine learning.

Ankur Sarda

Ankur Sarda

Ankur is a former software engineer at the Uber Risk engineering team focussing on graph applications at Uber.

Ankit Jain

Ankit Jain

Ankit Jain is a former research scientist of Uber AI.

Piero Molino

Piero Molino

Piero is a Staff Research Scientist in the Hazy research group at Stanford University. He is a former founding member of Uber AI where he created Ludwig, worked on applied projects (COTA, Graph Learning for Uber Eats, Uber’s Dialogue System) and published research on NLP, Dialogue, Visualization, Graph Learning, Reinforcement Learning and Computer Vision.

Posted by Xinyu Hu, Chengliang Yang, Ankur Sarda, Ankit Jain, Piero Molino

Category: