Skip to the content.

Call for papers

Scope of the workshop

Machine learning models for personalized medicine aim at tailoring diagnostic and therapeutic strategies to individual patient characteristics, which often means that these models have to be trained on data that is not only high-dimensional but also sparse and heterogeneous. Typical approaches assume the availability of a joint dataset of multiple patients for training a global model, from which patient-specific variants can be built. In practice, difficulties associated to data collection and governance arise when compiling comprehensive datasets needed for training robust machine learning models. Moreover, working on historical data is complicated by the fact that this data often exhibits a domain shift in the data distribution, resulting from the use of different machinery and diagnostic techniques over time.

At the same time, the paradigms of incremental and continual learning in machine learning have garnered significant attention as a means to develop models that can adapt to new information over time without forgetting previously learned knowledge. This capability would perfectly suit the requirements of personalized medicine, assuming a scenario where machine learning models for healthcare could be trained incrementally on data from few or even a single patient at a time. Such an approach would not only respect the privacy and data ownership of individuals but would also allow for the continual updating of models for previous patients as new data becomes available.

The workshop on Personalized Incremental Learning in Medicine (PILM) aims to bridge the gap between incremental learning research and its application to personalized medicine. The objectives of this workshop are multi-fold:

Topics

The workshop aims to attract novel and original contributions at the intersection between incremental learning and personalized medicine. Expected submissions should cover, but are not limited to, the following topics:

Workshop Proceedings

Workshop proceedings will be published in the Lecture Notes in Computer Science (LNCS) series by Springer.