Graph Learning in Financial Networks, Lecture notes of Financial Management

Download Graph Learning in Financial Networks and more Lecture notes Financial Management in PDF only on Docsity!

Graph Learning

in Financial Networks

Jiaxuan You

Stanford University

Collaboration with Tianyu Du, Fan-yun Sun, Jure Leskovec

Financial Networks

§ Financial Networks: Describe financial entities and their connections

Image credit: The Political Economy of Global Finance: A Network Model (^) Image credit: https://dailyblockchain.github.io/ International banking

• Nodes: Countries
• Edges: Capital flows Bitcoin transactions
  • Nodes: BTC wallets
  • Edges: Transactions

Why Graph Representation?

§ Transaction-based approach

§ “On 01/03, Client 𝐴 sends Company 𝐵 $500” § Build models based on transaction attributes § Issues: ignore the context of a transaction

§ Graph-based approach

§ Represent transactions as a dynamic graph § Predictions are made based on the entire graph § Benefits: § Represents a transaction with a broader context § Requires fewer feature-engineering $400, 01/ client (^) bank company $100, 01/ $500, 01/ $200, 01/ $200, 01/ $100, 01/

Overview of This Talk

ROLAND: GNN

for Financial

Networks

Results &

Insights from

ROLAND

Application:

Anomaly

Detection

Graph Neural Networks client (^) bank company NN NN NN NN NN NN

ROLAND Model: From Static to Dynamic GNN GNN Layer 1 Graph! Pred " GNN Layer 2

(")

($) a) A static GNN with modern architectural design options Skip connections GNN Layer MLP Layer MLP Layer Message passing layers Post- process layers GNN Layer GNN Layer GNN Layer MLP Layer MLP Layer Post- process layers MLP Layer MLP Layer Pre- process layers Linear BatchNorm Dropout Activation Attention Aggregation Aggregation Attention Activation Dropout BatchNorm Linear GNN Layer 1 Snapshot !%&" Pred "% GNN Layer 2 Embedding update Embedding update #$%&" (") #%&" (") #$%&" ($) #%&" ($) GNN Layer 1 Snapshot !% Pred "%'" GNN Layer 2 Embedding update Embedding update #$% (") #% (") #$% ($) #% ($) #%&" (") #%&" ($) b) Extend static GNNs to dynamic GNNs …^ … Hierarchical Node state #% = {#% " , … , #% (() } Question: Can we adapt a SOTA static GNN to dynamic prediction tasks? t+ t+ t+ This transaction is fraudulent SOTA Static GNN

ROLAND Model: From Static to Dynamic GNN

§ Idea: Recurrently update node

embeddings at each layer

§ Introduce a new module to a static GNN: § Benefits: § Simple and effective § Benefit from the SOTA designs of a static GNN Static GNN Dynamic GNN Embedding update Input: § Previous embeddings from the the same layer § Current embeddings from the previous layer Output: Updated embeddings GNN Layer 1 Graph! Pred " GNN Layer 2 GNN Layer 1 Snapshot !!"# Pred "! GNN Layer 2 Embedding update Embedding update GNN Layer 1 Snapshot !! Pred "!$# GNN Layer 2 Embedding update Embedding update …^ …

Overview of This Talk

ROLAND: GNN

for Financial

Networks

Results &

Insights from

ROLAND

Application:

Anomaly

Detection

Graph Neural Networks client (^) bank company NN NN NN NN NN NN

ROLAND Implementation: GraphGym

§ GraphGym offers a standardized pipeline for graph learning

§ Design Space of Graph Neural Networks (NeurIPS 2020 spotlight) § Now officially part of PyG 2.0! (Stay tuned for my next talk :) General GNN design space Clean interface for comparing designs J. You , R. Ying, J. Leskovec. Design Space of Graph Neural Networks, NeurIPS 2020

§ We implement the financial prediction pipeline with GraphGym

§ We adopt many optimal design choices from GraphGym

§ Task setting: Supervised learning

§ Unbalanced labels: 2% of all the transactions are fraudulent § Random dataset split: 80% training, 10% validation, 10% testing § Metric: AUC on test set (random guess has AUC 0.5, higher is better)

§ Best non-deep learning baseline

§ AUC: ~0.

§ Simple GNN implementation using out-of-the-box PyG

§ AUC: 0.

§ ROLAND framework works much better!

§ AUC: 0.

Task 1: Classify Fraud Transactions

Task 2: Forecast Future Transactions

§ Task setting: Self-supervised/Unsupervised learning

§ Rolling prediction: On each day, use all the historical information to predict the transactions on the next day § Metric: Mean reciprocal ranking (MMR) of ground-truth future transactions ROLAND significantly outperforms SOTA baselines

Analysis: Interpreting ROLAND’s Predictions

§ Question: What kind of transactions the

model paid attention to?

§ Method : Inspect edges' attention scores of

the last GNN layer.

§ Select edges with low/high attention scores. § Get the feature value distribution of these edges Domestic - > Abroad Non-financial à Household Higher attention to large amount transactions

Overview of This Talk

ROLAND: GNN

for Financial

Networks

Results &

Insights from

ROLAND

Application:

Anomaly

Detection

Graph Neural Networks client (^) bank company NN NN NN NN NN NN

Self-supervised objectives

Real world Indicators

(from baselgovernance.org)

Objectives for GNN

investment funds sent to countries of concern Predict country Large cash deposits into (company) accounts Predict amount unusual subject of transaction Predict subject of transaction (link prediction) high volume of transactions within a short period predict acculumated amount multiple individuals sending funds to the one beneficiary predict # transactions “Indicators and patterns of money laundering or terrorist financing”

Objectives for Anomaly Detection t+ t+ t+ Self-supervised Objectives ● Objective 1: predict amount ● Objective 2: predict recipient ● Objective 3: predict country ● … dim |N| Outlier Detection Objectives

§ Solution: ROLAND framework + diverse objectives