As one of the most active research topics, multimedia big data becomes increasingly popular especially for web and cloud based applications. With rich information contained in various modalities and dimensions, effective data analytics and interpretation becomes the bottleneck. To tackle this challenging problem, smart mining is desirable to refine the quality and characterise the data for improved understanding of the data. Relevant applications can be widely found in energy, manufacturing, health, digital economy, security, remote sensing and other fields. With emerging techniques in machine learning, artificial intelligence, IT, data management and cloud computing, new trends are arisen to combine such techniques for more robust and effective mining of multimedia big data. This will include not only new models and algorithms but also new applications where quality, efficiency and efficacy should be focused. As a result, in this Workshop we aim to gathering most up to date contributions in this area, which will provide a useful forum for the community to report the progress and exchange their findings. We are particularly interested in fundamental models, algorithms, integrated solutions and novel applications as well as data and methods for performance assessment. Researchers in all the areas of multimedia data mining are welcome to attend this Workshop, where both original research and review articles in relevant fields are overed.