The continued progress in digital data acquisition, storage and communication technology generates huge amounts of multimedia data (i.e., text, images, audio, video, among others), which are made widely available thanks to internet supported sharing platforms. This kind of information is a valuable source of knowledge, that can only be properly exploited with the development of effective tools that integrate the complementary information obtained from the different perspectives of this heterogeneous kind of data.
The main goal is to investigate effective and efficient methods to combine complementary evidence and model the relationships between multiple modalities (or views) of multimedia data in order to obtain valuable insights about the data and
improve the performance in multiple tasks, such as content-based retrieval, exploration and automatic annotation and classification, among others.
Representation is a fundamental process for machine learning. Its goal
is to extract useful features from training data which are later fed to a
learning algorithm. Traditional approaches are
based on standard or hand-crafted feature detectors which are manually
selected to fit the problem at hand using expert knowledge in the
domain. A main drawback in hand-crafted features is the high cost of
such expert intervention. Experts usually have to design a different set
of features for each problem. Representation learning tackle this
problem from a different perspective. Instead of designing custom
feature detectors, representation learning learns them from data. More...
growing on variety, volume and velocity of public biomedical databases
in the last years have generate an explosion of big data in biology and
medicine. Most of these databases comprise structural, molecular and
genetic information from different kind of images acquisition modalities
and associated metadata having a great potential, not yet exploited, as
a source of information and knowledge which could impact biomedical research in different application fields.
address the problem of analyzing medical information, using
computational tools, to automatically discover patterns and inherent structures within such databases. More...
Classical latent semantic embedding models can boost the performance in tasks like content-based retrieval, exploration and automatic annotation and classification, among others. Nevertheless, one important limitation of most of these approaches is that they assume linear dependencies. However, it is reasonable to expect that there is an inherent non-linearity in the nature of the data. Therefore, this kind of linear models can only give a limited and restricted approximation of the nature of the data and the complex relationships. This important issue, motivates this research area.