LabDocs Unlocked

The 2025 ADLM Data Science Challenge

Introduction

Laboratories generate vast amounts of documentation, ranging from protocols and package inserts to regulatory materials like 510K clearance documents and checklists. These critical resources require significant time and effort to navigate, presenting a challenge for efficient decision-making and compliance management.

The 2025 ADLM Data Science Challenge seeks to address this gap by leveraging the power of AI to create a tool capable of quickly and accurately extracting and presenting user-requested information from these complex document stores. The proposed solution would ensure that laboratory professionals can focus on impactful work rather than time-intensive document searches.

The Challenge

Develop a Tool for Extracting Information from a Laboratory Document Store

We want you to develop an AI tool to make looking up information in laboratory documentation faster and more efficient. We will provide a representative set of documents. Your tool should ingest these documents and then extract and present user-requested information from those documents.

Preliminary Competition and Scoring

Teams will participate in a preliminary competition where they will be scored based on the following:

Criterion	Points	Description
Accuracy	10	Responses to a test set of user‑requests will be graded using a standardized scoring rubric.
User experience	10	The tool and its output should be intuitive and easy to navigate.
Explainability	10	The tool should provide references or links to relevant sections in the document store.
Best practices	10	Coding: Comments Modularity Reusability (e.g. can ingest new document store) Version control (logical commit histories) Deployment Usage notes provided Dependencies defined Virtualization, containerization, or web hosting

Note

Source code and commit histories are required to receive best practices points

Final (Live) Competition

The top two teams from the preliminary competition will be invited to present how they built their solution in a live Webinar/competition event. The teams will then compete based on the speed and accuracy of their responses to a new set of questions presented "game show style." The team with the highest score will be declared the winner.

Significance

Efficient access to laboratory documentation is essential for maintaining compliance, ensuring quality, and accelerating decision-making. This challenge addresses the inefficiencies caused by the complexity and volume of such documents by leveraging AI to streamline information retrieval. The resulting tool has the potential to transform workflows, reduce errors, and set a new standard for document management in laboratory practice.

Timeline

Date	Event
July 25th, 2025	Competition Begins
November 15th, 2025	Preliminary Competition Ends
December 15th, 2025	Announcement of Finalists
January, 2026 (Anticipated)	Finalists Presentations Webinar and Live Competition

How to Participate

Tip

If GitHub or any of the instructions below are unfamiliar to you, please reach out to us using the contact information at the bottom of this page. We will help you get set up to participate.

Sign up for a GitHub.com account

1. Navigate to https://github.com/
2. Click Sign up
3. Follow the prompts to create your personal account

Fork the competition repository

1. Navigate to https://github.com/myADLM/ADLM-2025-Data-Challenge/
2. Click Fork
3. Select Create a new fork
4. Set your GitHub account as the owner (default)
5. Click Create fork

Download the dataset

Caution

This dataset contains synthetic laboratory protocols that were generated by a large language model (LLM). These protocols are not validated methods and must not be used for any clinical, diagnostic, therapeutic, or patient‑care activities. They are provided solely for educational purposes.

Note

This dataset is about 3.5 gigabytes in size.

Command-line

Linux / MacOS

curl -L -C - -o LabDocs.zip \
    "https://zenodo.org/records/16328490/files/LabDocs.zip?download=1"

unzip LabDocs.zip

Windows

curl -L -C - -o LabDocs.zip \
    "https://zenodo.org/records/16328490/files/LabDocs.zip?download=1"

tar -xf LabDocs.zip

Manual Download via Browser

1. Open this link in your browser: https://zenodo.org/records/16328490
2. Click Download (or Save As…) and save the file as LabDocs.zip.
3. Use your operating system’s built-in extraction tool or one of the command-line method above.

Build your solution

1. Clone the forked repository
   

       git clone https://github.com/myGitHubUsername/ADLM-2025-Data-Challenge.git

Note

Replace myGitHubUsername with your GitHub handle

2. Make a folder in the cloned repository with your team name
3. Build your solution within your team folder and track your development with commits to the repository

Submit your solution via a pull request

1. Push your commits to the forked repository

   git push

2. Navigate to https://github.com/myGitHubUsername/2025_ADLM_Data_Analytics_Challenge
   

   [!note] Replace myGitHubUsername with your GitHub handle

3. Click 'Contribute'
4. Select 'Open pull request'
5. Provide a contact email in the description for your pull request
6. Click 'Create pull request'

We will review your pull request to ensure that it contains everything needed to score your submission. We may contact you via your contact email, if needed.

Important

You must submit your solution by November 15th, 2025 to be eligible to win the competition

Need Help?

If you are unfamiliar with GitHub, need help getting starting, or have other questions, we are available to help! Please email mboyle@myadlm.org for assistance.