Integrated Quantum Launches $10K Kaggle Hackathon to ‘Pierce the VEIL’

Integrated Quantum Technologies has launched a hackathon, “Piece the VEIL: Hack It and Crack It Simulation,” on Kaggle. The competition will end May 24th, providing participants with a defined window to develop and submit their solutions, and a chance to win pretty substantial prizes.

DETAILS:

A total prize pool of USD $10,000 will be awarded, with $8,000 awarded to the top-performing solution and $2,000 awarded to a secondary qualifying submission, based on the competition’s validation criteria. 

The competition is a global, open challenge inviting data scientists, cryptographers, and AI researchers to attempt a full reconstruction of protected data from outputs generated by IQT’s proprietary VEIL™ technology, without access to the original inputs, structure, or encoding methods. Participants must submit algorithms that reconstruct original data records from unseen values, directly testing whether such transformations can be reversed.

Modern machine learning systems depend on sensitive data, but once that data leaves its source environment, it becomes vulnerable. Attackers today don’t just steal raw datasets; they intercept intermediate representations and attempt to reverse-engineer the original inputs.

This competition challenges participants to do exactly that. The task is to reconstruct the original data—or prove that it can be done.

Participants are given access to a batch of 4,096 one-dimensional values. These values were sufficient to power a high-performing machine learning pipeline, matching the predictive accuracy of a model trained on the original raw data. However, the original inputs, their dimensionality, feature names, structure, and generating process are completely hidden.

More details here: https://www.kaggle.com/competitions/pierce-the-veil 

About VEIL™

Integrated Quantum’s VEIL™ represents A new era of data security.

VEIL™ is a data security solution that sets a new standard for enterprises using machine learning by removing what’s valuable to adversaries before data ever enters ML pipelines, while still retaining what’s useful. The solution enables enterprises to secure data before it enters ML model pipelines using a new privacy-preserving framework that completely removes personally identifiable information (PII), while conserving and enhancing data utility — allowing for scalable innovation without tradeoffs.