Method
Autoencoder
Reconstruction-based representation learning, incl. VAE.
9 cases from 7 companies apply Autoencoder.
Typical tasks: Learn user and item embeddings, Detect anomalies, Explain a metric change. Companies: Walmart, Zillow and 5 more. Often combined with: Gradient boosting, Outlier handling, Autoencoder anomaly detection, Feature importance, LSTM and 36 more.
Virtual Flow Metering: Forecasting Oil Production with Machine Learning![]() | ![]() | manufacturing | 2025-05 |
| Single AI View of Customer: A retailers guide to Know Your Customer Better using Customer Embeddings and Customer Persona | retail | 2025-03 | |
| Geo Embeddings: Universal Representations for Locations | banking | 2024-05 | |
| Neural Network Architectures in Credit Scoring | — | banking | 2024-05 |
| Imputing Data for the Zestimate | marketplace | 2023-11 | |
| Explaining Autoencoder Anomalies in Industrial Control Systems | manufacturing | 2023-10 | |
A Guide to Active Learning in CV: Saving on Labeling and Getting a Quality Model![]() | ![]() | general | 2023-05 |
| A Trial ML System Design Interview | — | retail | 2023-05 |
| Preprocessing and Anomaly Detection in Time SeriesWT | WT | manufacturing | 2023-05 |

