Contextual Search and Bilateral Trade with a Trace-Square-Root Potential

Get email updates

Romain Cosson, Federico Fusco, Anupam Gupta, Stefano Leonardi, Renato Paes Leme, Matteo Russo

For fixed buyer and seller parameter vectors in the unit Euclidean ball and adaptively chosen contexts of norm at most 1, the paper gives a randomised one-price algorithm whose expected cumulative gain-from-trade regret is O(d)O(\sqrt d) in dimension d. It uses only the trade bit and needs no prior or time horizon.

What was already known

Earlier work by the authors gave an O(d6d)O(d6^d) guarantee in this setting. The new order matches the known lower bound up to a constant.

AI involvement AI-centered

The authors credit GPT-6 Astra with suggesting the first proof of the adversarial bilateral-trade bound. They simplified the argument and developed the interpretation and related results.

Importance Notable

The result settles the dimensional order of worst-case regret for this one-bit model.

Technical verification Reported

The paper gives the theorem and proof. The available sources do not document a completed check of this argument.

Human understanding Explained

The authors explain how the potential tracks uncertainty in the two valuation vectors and how a lottery over prices guarantees enough expected decrease to cover the loss.

Correction and publication history