Dimension Reduction: A Powerful Principle for Automatically Finding Concepts in Unstructured Data

Dimension reduction techniques have been a successful avenue for automatically extracting the “concepts” underlying unstructured data, a task that naturally arises in fields as diverse as information retrieval, image processing, social science, etc. It is surprising how much can be achieved for this task using only the raw data itself, without resorting to any additional knowledge or intelligence. We will survey the most important schemes contributed from the various communities to date, by commenting on the following aspects: optimization techniques, the role of normalizations, setting the parameters, computing time, quality of results, and the integration of external knowledge.