Method

Imbalanced learning

Oversampling, SMOTE, class weights and focal loss for rare classes.

20 cases from 18 companies apply Imbalanced learning. 3 were checked in an online A/B test; 7 report results in production. Includes variants: Focal loss.

Typical tasks: Catch fraud and abuse, Moderate content at scale, Optimize decisions, Evaluate models and 1 more. Companies: Target, Netflix, LinkedIn, Etsy, CFT, Canva and 12 more. Often combined with: Focal loss, Gradient boosting, Image classification, Synthetic data generation (LLM), Text classification and 43 more.

TargetRun: AI-Powered Trip Prediction for Delivering Timely Offersretail2026-06
With Eyes Wide Shut: Detecting Sleeping People and Other Remote Verification Attacks Without Consentfintech2026-05
Email Text Classification Under Severe Class Imbalancemanufacturing2026-05
Building the incremental and online training platform at LinkedInsocial2025-10
Attributor, or How an AI Agent Identifies the Advertiser of Digital Campaignsadtech2025-05
A Data-Driven Approach to Subscriber Scoring in Telecomtelecom2025-05
Predictive Modeling for Availability of Inventoryretail2024-10
Machine Learning in Content Moderation at Etsymarketplace2024-08
Predictive Analysis of Infrastructure Risks on Railway Tracksmanufacturing2024-05
Analyzing Experiment Design Validity in ECG Diagnosis with Machine Learninghealthtech2024-05
Solving the Face Spoofing Problem in Photographsfintech2023-05
Deep Learning for Infinite (Multi-Lingual) Keywordsmarketplace2023-01
Identifying High-Intent Buyersclassifieds2022-11
How Uber Optimizes the Timing of Push Notifications using ML and Linear Programmingfood-delivery2022-11
Generation of three-dimensional synthetic data for training neural networksgeneral2022-10
Phishing detection + domain expertsgeneral2022-10
Machine Learning for Fraud Detection in Streaming Servicesmedia-streaming2022-09
Formulating ‘Out of Memory Kill’ Prediction on the Netflix App as a Machine Learning Problemmedia-streaming2022-07
Calibration of the Prediction Scores: How does it help?adtech2022-05
Using predictive technology to foster constructive conversationssocial2022-05