Generative AI Lab - Human-in-the-Loop AI Annotation, Tuning & Testing

Generative AI Lab (previously known as NLP Lab and Annotation Lab) is an End-to-End No-Code platform for annotating text and training AI/ML models. It enables domain experts to extract meaningful facts from text documents, images or PDFs and train models to automatically predict those facts on new documents. By offering out-of-the-box support for Large Language Model prompting, Zero-Shot prompting, Rules and state-of-the-art John Snow Labs pre-trained models, Generative AI Lab helps domain experts efficiently prepare training data for tuning custom AI models for specific tasks and use-cases.

The annotation tool also supports human-in-the-loop workflows. In industries like healthcare, in which regulatory-grade accuracy is a requirement, human validation is often a critical requirement. The tool supports task management, full audit trails, custom review and approval workflows, versioning, tuning, testing and analytics, fully supporting the human-in-the-loop needs of high-compliance industries.

About the offer: Based on an auto-scaling architecture powered by Kubernetes, the annotation tool can scale to many teams and projects. Enterprise-grade security is included, with support for air-gap environments, zero data sharing, role-based access, full audit trails, MFA, and identity provider integrations. Generative AI Lab allows powerful experiments for model training and finetuning, testing, and deployment as API endpoints. There is no limitation on the number of users, projects, tasks, models, or trainings that can be run with this subscription. This product includes a Pay-As-You-Go license key for John Snow Labs libraries and models, that offers access to 40.000+ models and pipelines for healthcare, legal, finance, downloadable from the NLP Models Hub and with access to OCR and Visual Document understanding features. Designed to take advantage of GPU architecture, the product offers a boost in performance for model training and preannotation tasks - https://nlp.johnsnowlabs.com/docs/en/CPUvsGPUbenchmark_healthcare  You will be charged ONLY as long as you use the product. Simply stop your instance and restart it when needed it so you get charged only based on what you consume.

Included Features:

• Prompt engineering for Large Language and Zero-Shot Models - entity recognition, relation extraction, classification.
• AI-Assisted Annotation: never start from scratch but reuse existing resources to pre-annotate tasks with the latest models for classification, NER, assertion status, entity resolution, relation detection;
• High productivity annotation UI with keyboard shortcuts and pre-annotations;
• Annotation support for Text, Image, Audio, Video and HTML;
• Text annotation in 250+ languages;
• Projects and teams: 30+ project templates; unlimited projects and users, project import, export , cloning, grouping;
• Task assignment, tagging, and comments; deduplication, searching and filtering;
• Inter Annotator Agreement charts;
• Enterprise-level security and privacy: role-based views and access control, annotation versioning, full audit trail, SSO;
• Full NLP Models Hub integration: explore and download models and embeddings, to reuse those in your projects.
• Train Classification, NER, and Assertion Status models: use default parameters, tune them on the UI for your experiments;
• Active Learning automatically trains new model versions once new annotations are available;
• Playground - deploy, test, and update prompts, rules and models before including them in your project;
• API access to all features for easy integration into custom pipelines;

Who is this offer for

• Domain experts (e.g. nurses, doctors, lawyers, accountants, investors, etc.) who want to test DL models on their data or/and tune/train new models via an easy-to-use UI, without writing a line of code;
• Data labeling teams who want to optimize the efficiency and speed of their day-to-day work with preannotations;
• Machine Learning engineers who need to test/train/tune NLP models;
• Researchers who need to extract meaning from unstructured, natural language documents;
• And anyone else interested in text and image analysis, image digitization, data extraction, document labeling and/or NLP model training.

Target verticals Its integration with the NLP Models Hub facilitates access to over 40k pre-trained models for general-purpose text documents as well as 2000+ pre-trained models covering 400+ clinical and biomedical entity types.

Technical Specifications Operating System:Ubuntu 20.04

3 Easy Steps to get started Subscribe to the product on the AWS Marketplace. Deploy it on a new machine. Access the login page for a guided experience on http://INSTANCE_IP. For the first login use the following credentials: Username: admin Password: INSTANCE_ID