Wadii Boulila

Continuous geographic modeling reveals that predicting Arabic dialect origins can achieve impressive accuracy while exposing significant room for improvement in zero-shot scenarios.

Stop punishing your model for disagreeing with corrupted data – Trust-SSL learns better representations by treating alignment with degraded views as a residual learning problem, not a hard constraint.

A unified benchmark reveals the trade-offs between pixel-wise accuracy and perceptual realism in state-of-the-art image super-resolution techniques.