This paper introduces a novel deep learning framework called Domain Adaptive Hashing (DAH) for unsupervised domain adaptation. DAH leverages labeled source data and unlabeled target data to learn informative hash codes, which can be used for accurate classification of unseen target data. The framework includes a unique loss function that combines supervised hash loss for source data, unsupervised entropy loss for target data, and a multi-kernel Maximum Mean Discrepancy (MK-MMD) loss to minimize the distribution difference between the source and target domains. The authors also introduce a new dataset, Office-Home, which contains images from multiple domains and categories. Extensive experiments on multiple transfer tasks demonstrate the effectiveness of DAH in learning efficient hash codes that outperform existing baselines for unsupervised domain adaptation. The paper also includes a detailed analysis of the feature representations learned by DAH, showing improved domain alignment and clustering compared to other domain adaptation methods.This paper introduces a novel deep learning framework called Domain Adaptive Hashing (DAH) for unsupervised domain adaptation. DAH leverages labeled source data and unlabeled target data to learn informative hash codes, which can be used for accurate classification of unseen target data. The framework includes a unique loss function that combines supervised hash loss for source data, unsupervised entropy loss for target data, and a multi-kernel Maximum Mean Discrepancy (MK-MMD) loss to minimize the distribution difference between the source and target domains. The authors also introduce a new dataset, Office-Home, which contains images from multiple domains and categories. Extensive experiments on multiple transfer tasks demonstrate the effectiveness of DAH in learning efficient hash codes that outperform existing baselines for unsupervised domain adaptation. The paper also includes a detailed analysis of the feature representations learned by DAH, showing improved domain alignment and clustering compared to other domain adaptation methods.