Data Mining

Differential Privacy

डिफरेन्सियल गोपनीयता

A method that adds noise to data so individuals can't be identified while patterns remain useful.

Long Context

लामो सन्दर्भ

The capacity of a model to process very large amounts of input text at once.

Matrix Factorization

म्याट्रिक्स फ्याक्टराइजेसन

Breaking a large matrix into smaller ones to uncover patterns, common in recommenders.

Active Learning

सक्रिय सिकाइ

A method where the model selects the most useful examples for humans to label.

Anomaly Detection

असामान्यता पहिचान

Spotting data points that deviate sharply from normal patterns, e.g. fraud or faults.

Autoencoder

अटोइन्कोडर

A network that compresses data into a compact code and reconstructs it, useful for learning features.

Big Data

बिग डाटा / विशाल तथ्याङ्क

Extremely large and complex datasets that need special tools to store and analyze.

Collaborative Filtering

सहयोगी छनोट

Recommending items by finding users with similar tastes and patterns.