JoinDiscord

🌊 Simplify annotation with machine feedback workflows

Besides direct manual annotation, we also support a few annotation workflows that can be used to smoothen the annotation for your records. These workflows are: active learning, weak supervision, semantic search and job scheduling.

🧑‍🏫 Active Learning

A human-in-the-loop framework for iteratively building training sets and models.

👮 Weak Supervision

An approach to machine learning in which high-level rules and queries are used to label training data.

🔦 Semantic Search

Find similar and dissimilar records in your dataset by using semantic vectors.

⏲️ Job Scheduling and Callbacks

Schedule jobs to run at a specific time or periodically.

📇 Text Descriptives as Metadata

Add text descriptives to your metadata to simplify the data annotation and filtering process.