MACHINE LEARNING

Transfer Learning

A machine learning technique where a model trained on one task is repurposed as the starting point for a model on a different but related task.

BasaltHQ leverages transfer learning to dramatically reduce the time and data required to deploy industry-specific AI. Rather than training a legal document classifier from scratch (which would require millions of labeled examples), we start with a model pre-trained on general legal corpus and fine-tune it on your specific contract library. The model inherits broad legal reasoning capabilities and quickly learns your company's unique terminology, clause structures, and risk factors. This reduces deployment time from months to days.