 |
Center for Nonlinear Studies |
Massive simulations and arrays of sensing devices, in combination with increasing computing resources, have generated in the recent decades large, complex, high-dimensional datasets used to study phenomena across numerous fields in science and engineering such as climate modeling, clean, efficient energy production or creation of new materials. Visualization plays an important role in exploring datasets in such fields although high dimensional information presents unique challenges in terms of data management and processing as well as presentation to the user. In particular, the curse of dimensionality refers to the intrinsic need to use large data collections and consume vast amounts of processing resources when analyzing phenomena that are well represented only by a large number of dimensions. In simple terms, the curse of dimensionality implies that software tools are intrinsically not scalable when dealing with high dimensional data and aiming at interactivity becomes particular challenging. Moreover, the effectiveness of visualization methods depends heavily on the use of visual metaphors that are familiar to the user since they facilitate the quick, intuitive understanding that is typical of interactive data exploration. Unfortunately, user intuition tends to break down when dealing with high dimensional data since the user is presented with configurations that cannot be found in familiar 3D spaces. In this talk, I will discuss some of the limitations of visualization methods (even in low dimensions) and present recent advances in high-dimensional data visualization attained in the past decade. Promising results achieved in a number of application areas show the value of high dimensional data visualization despite the challenges and limitations. Moreover, the numerous lessons learned provide guidance for data practitioners in the selection of useful techniques and inspire the creation of new, effective techniques while identifying future opportunities for visualization research.
Host: Curt Canada, 505-665-7453, cvc@lanl.gov