A few challenges related to exploiting multimedia collections

Media analytics refers to interactive exploration of multimedia collections to search for information, gain insight and knowledge on a topic of interest. Exploring large-scale collections requires an organization of the collection on top of the description of each item. We will review recent research at IRISA on the creation of links in media archives to offer graph-based navigation in large-scale collections, leveraging speech and natural language processing, statistical modeling, machine learning and graph theory.

Recent research has shown that image recognition is vulnerable to adversarial attacks. By adding to the query image an imperceptible amount of adversarial noise, it is highly likely that image recognition can be tricked, possibly returning with a high similarity confidence an image that we, as humans, find extremely dissimilar. Such adversarial behaviors have also been observed in the context of machine learning and deep learning. We will discuss these issues.