Combining data, AI, network science, and psychology for social influence.

What you'll learn

How cognitive biases mold our view of the world, and how they can be leveraged to exert influence

How directed influence campaigns shape opinion in social networks

How AI can generate realistic data, and how that data can be used to deceive

How to build graph neural networks (GNN, GCN, GAT, Node2Vec, DeepWalk, & more)

How statistical analysis and hypothesis tests can be fudged to accept or reject any hypothesis

How to detect rising stars in social networks and root out botnets

Build a hate speech detector bot for Slack

Build a news recommendation website

Run Bayesian A/B tests in real time on your news recommendation website

Requirements

You should know the foundations of machine learning, statistics, and network science.

Intermediate Python and Docker skills are required for the projects. You should know how to use the following libraries: Numpy, Pytorch, Django, FastAPI

Some knowledge of linear algebra, psychology, and philosophy would be helpful.

Description

A new age has arrived. AI is sufficiently advanced to learn our opinions and what we care about, and craft text and media to influence our thoughts and opinions. It is likely that AI will soon be better able to influence us than other people. Individuals and organizations equipped with AI are now able to exert influence at a previously inconceivable scale, and they will become more successful at it over time.In this course, we will combine concepts from psychology, data science, and network science to describe how social influence can be exerted. We will consider how our thoughts are influenced by our social networks, and how our biases work. We will explore how an individual’s opinions impact social networks, and how the collective opinions of entire networks can change under the right conditions. You will see how statistical analysis can be manipulated and how AI can be used for deception. Ultimately, you will learn how to exert large scale social influence, using AI for leverage.This is not a course for beginners. Basic concepts in data science will not be explained. This is an interdisciplinary course that will challenge you to think for yourself. You will learn about powerful techniques and you will need to decide how to manage them ethically and morally.

Overview

Section 1: Introduction

Lecture 1 About this Course

Lecture 2 Are You Ready for this Course?

Lecture 3 Course Materials

Section 2: Psychology of Social Influence

Lecture 4 Psychology of Social Influence Intro

Lecture 5 A Perspective of Social Influence

Lecture 6 Cognitive Biases Part 1: Primers, Illusory Truth Effect, Availability Heuristic

Lecture 7 Cognitive Biases Part 2: Cognitive Dissonance

Lecture 8 Cognitive Biases Part 3: Conformity & Ostracism

Lecture 9 Behavior in Groups

Section 3: Influence in Social Networks

Lecture 10 Influence in Social Networks Intro

Lecture 11 Influence

Lecture 12 Influence Decay and the Network Horizon

Lecture 13 Information Spread in Social Networks

Lecture 14 Phase Transitions in the Ising Model

Lecture 15 The Rise of an Influencer + Demo of Detecting a Rising Star

Section 4: Graph Representation Learning

Lecture 16 Graph Representation Learning Intro

Lecture 17 Graph Feature Engineering

Lecture 18 Graph Spectral Properties & the Laplacian

Lecture 19 Graph Embeddings

Lecture 20 GNNs Part 1

Lecture 21 GNNs Part 2

Lecture 22 Graph Convolutions & GCNs

Lecture 23 Graph Embeddings & GNNs for Dynamic Graphs

Lecture 24 Evaluating Graph Representations

Lecture 25 Project Overview: Node Classification with GNNs

Lecture 26 Project: Node Classification with GNNs

Section 5: Data Manipulation

Lecture 27 Data Manipulation Intro

Lecture 28 How to Fake Statistical Analysis

Lecture 29 Bayesian A/B Testing

Lecture 30 How to Generate Realistic Data

Lecture 31 Demo: How to Break Benford's Law

Lecture 32 Fake News & Deepfakes

Lecture 33 How to Create a Deepfake & Leverage it for Social Influence

Lecture 34 Exploiting Data Visualization

Section 6: Media Bias & Propaganda

Lecture 35 Media Bias & Propaganda Intro

Lecture 36 Media Bias

Lecture 37 Propaganda

Lecture 38 Censorship

Lecture 39 Project Overview: Hate Speech Detection

Lecture 40 Project: Hate Speech Detector

Lecture 41 Project Overview: News Recommender

Lecture 42 Project: News Recommender

Section 7: Directed Influence Campaigns & Botnets

Lecture 43 Directed Influence Campaigns Intro

Lecture 44 Directed Influence

Lecture 45 Demo: Social Botnet Detection

Lecture 46 Project Overview: Directed Influence Campaign

Lecture 47 Project: Directed Influence Campaign

Section 8: Conclusion

Lecture 48 Where to Go From Here

Data Scientists, ML Engineers, and Data Analysts with a few years of work experience or higher education,This is not a course for beginners.