(PopTika/Shutterstock)
Snorkel AI recently announced updates to its flagship Snorkel Flow platform aimed at enhancing the customization of AI and ML models while minimizing the time, cost, and complexity involved in data preparation for enterprises.
The new functionality includes the ability to utilize GenAI for use-case-specific benchmarks, advanced named entity recognition (NER) for PDFs, fine-tuning workflows, AI ecosystem integrations, and more.
With the introduction of the new capabilities, Snorkel Flow seeks to provide comprehensive support for the entire AI data development life cycle, from data preparation and labeling to model evaluation and deployment.
AI-ready data is critical for AI development, governance, and model validation. Enterprises must ensure they have not only sufficient data but also data that accurately reflects relevant patterns, errors, outliers, and unexpected elements from real-world applications. This approach enables them to leverage their unstructured and proprietary data effectively.
“AI is at the top of every enterprise leader’s priority list, but the deep work needed for consistent, repeatable AI development is daunting, costly, and manual,” said Alex Ratner, CEO and co-founder of Snorkel AI.
“Data underpins successful adoption of AI in large enterprises, which is why these updates to our AI data development platform are so important. They’re fundamental to helping enterprises accelerate and optimize the delivery of AI solutions.”
Snorkel AI is a four-year-old startup based in Palo Alto, CA, that emerged from a data labeling project at the Stanford AI Lab. Researchers at the lab identified a market need for tools to automate data preparation, leading to the creation of the Snorkel framework and the eventual founding of Snorkel AI to commercialize this technology.
Since its launch, Snorkel AI has been implemented in production environments by several Fortune 500 companies, including Wayfair, Chubb, and BNY Mellon.
While competitors like Clarifai and Scale AI have introduced similar tools for managing the lifecycle of AI models, the main feature of SnorkelFlow is its capability for programmatic data labeling and development.
Snorkel Flow claims that its latest update provides enterprises with a powerful platform for implementing and scaling AI data development practices, thereby speeding up the production of highly accurate, specialized AI models.
Among the key updates to the platform are LLM evaluation tools which allow users to customize assessments for specific domains and derive in-depth insights about performance, error modes, and data development needs.
The new release comes at a time when organizations face challenges with GenAI’s accuracy and scalability. Across sectors, there’s growing recognition of the need for high-quality, use-case-specific data to enhance the performance of LLMs.
Traditional AI systems often struggle with retrieval accuracy and relevance, leading to inefficient responses and increased development time. Snorkel Flow aims to solve this by introducing enhanced RAG tuning workflows that improve retrieval accuracy through advanced chunking, metadata extraction, and embedding model fine-tuning.
Extracting relevant information from complex PDF documents can be quite challenging and often involves considerable manual effort. Snorkel Flow tackles this issue with its new Named Entity Recognition (NER) feature, which simplifies and accelerates the process of information extraction from PDFs.
via Shutterstock
Additionally, this NER capability can be utilized for virtual assistants and GenAI chatbots by improving their ability to understand and retrieve specific information from documents.
According to Snorkel AI, the new update offers a more efficient workflow for subject matter experts, allowing them to annotate data more easily. It also provides deeper integration with popular AI development tools like Amazon SageMaker and Databricks.
The new Spotlight Mode helps users identify model prediction errors in sequence tagging workflows, while class-level metrics offer deeper insights into performance. Other improvements include user experience (UI) enhancements to allow for smoother navigation and data labeling.
Earlier this year, Snorkel AI announced a major update to Snorkel Flow focused on the integration of enterprise data with AI models. Building on that update, Snorkel Flow’s latest release addresses critical aspects of data readiness and model evaluation. It provides enterprises with a useful tool to leverage unstructured data effectively while improving workflows.
As enterprises focus on tailored AI solutions, the advancements in Snorkel Flow could impact how they approach the development and deployment of specialized models for specific business requirements.
Related Items
Snorkel AI Heads into 2023 with Record Momentum and Growth
Analyze This: How to Prepare Your Unstructured Data in Six Steps
Snorkel AI Launches Application Studio, Raises $35M Led by Lightspeed Venture Partners
Source link
lol
