%202.png)
.svg)
Renaissance Learning, an educational technology company, needed to efficiently align Common Core aligned test items to candidate state standards. The process of manually aligning each item to the most appropriate state standard was time-consuming and resource-intensive.
Renaissance Learning partnered with Add Value Machine (AVM) to leverage AI, specifically Large Language Models (LLMs) on AWS, to automate parts of the alignment process. By using LLMs to predict the most likely state standard alignment for each item, Renaissance Learning could significantly reduce the manual effort required while maintaining alignment accuracy.
AVM developed a pipeline using AWS AI/ML services that uses outputs from multiple pre-trained LLMs to auto-align Common Core items to candidate state standards. The pipeline, built using AWS SageMaker JumpStart, includes:
1. Prompt templates that frame the alignment task for the LLMs in an optimal way
2. An ensemble of LLMs from the JumpStart model zoo that predict the alignment for each item/standard pair
3. Alignment methods that synthesize the LLM outputs into final alignment decisions
4. Auto-alignment of high confidence items to avoid manual review
5. Identification of lower confidence alignments for expert review
This allows Renaissance Learning to auto-align items where the LLMs have high agreement, while flagging more ambiguous cases for human review. The pipeline is flexible and can be tuned for different subjects, standards, prompt templates, LLMs, and alignment methods. AWS SageMaker JumpStart enables rapid prototyping and deployment of this solution.
AVM's AI-assisted alignment pipeline, powered by AWS AI/ML services, significantly reduces the manual effort required for educational technology companies like Renaissance Learning to align test items to state standards. By auto-aligning high confidence items and identifying which lower confidence alignments truly require expert review, the process becomes much more efficient.
This efficiency enables educational technology companies to more easily adapt their item banks to multiple state standards. It also frees up subject matter experts to focus their review efforts on the alignments that truly require their judgment and expertise. The use of AWS services ensures the solution is scalable, reliable, and can be quickly adapted to each company's specific needs.
For Renaissance Learning, AVM's AI-assisted alignment pipeline on AWS provides a competitive advantage by reducing the time and cost to align their item banks to state standards. The use of AWS services allows for rapid implementation and scaling of the solution. This efficiency allows Renaissance Learning to update their alignments more frequently and better serve their customers.
For Renaissance Learning's customers (schools and districts), the key benefits are more up-to-date and state-specific assessment content, potentially at a lower cost due to the reduced manual alignment effort. Educators can also have confidence in the alignment quality due to the combination of AI and expert review. The use of AWS ensures the reliability and security of the alignment process.
In summary, AVM's AI-assisted alignment pipeline, built on AWS AI/ML services like SageMaker JumpStart, leverages the power of large language models to drive significant efficiency gains for Renaissance Learning, while still incorporating human judgment where needed. The use of AWS ensures a scalable, reliable, and adaptable solution. This provides major benefits to both Renaissance Learning and the schools and districts they serve.
.svg)