With the rapid adoption of machine learning techniques for large-scale applications in science and engineering comes the convergence of two grand challenges in visualization. First, the utilization of black box models (e.g., deep neural networks) calls for advanced techniques in exploring and interpreting model behaviors. Second, the rapid growth in computing has produced enormous datasets that require techniques that can handle millions or more samples. Although some solutions to these interpretability challenges have been proposed, they typically do not scale beyond thousands of samples, nor do they provide the high-level intuition scientists are looking for. Here, we present the first scalable solution to explore and analyze high-dimensional functions often encountered in the scientific data analysis pipeline. By combining a new streaming neighborhood graph construction, the corresponding topology computation, and a novel data aggregation scheme, namely topology aware datacubes, we enable interactive exploration of both the topological and the geometric aspect of high-dimensional data. Following two use cases from high-energy-density (HED) physics and computational biology, we demonstrate how these capabilities have led to crucial new insights in both applications.

Training deep neural networks on large scientific data is a challenging task that requires enormous compute power, especially if no pre-trained models exist to initialize the process. We present a novel tournament method to train traditional as well as generative adversarial networks built on LBANN, a scalable deep learning framework optimized for HPC systems. LBANN combines multiple levels of parallelism and exploits some of the worlds largest supercomputers.We demonstrate our framework by creating a complex predictive model based on multi-variate data from high-energy-density physics containing hundreds of millions of images and hundreds of millions of scalar values derived from tens of millions of simulations of inertial confinement fusion. Our approach combines an HPC workflow and extends LBANN with optimized data ingestion and the new tournament-style training algorithm to produce a scalable neural network architecture using a CORAL-class supercomputer. Experimental results show that 64 trainers (1024 GPUs) achieve a speedup of 70.2× over a single trainer (16 GPUs) baseline, and an effective 109% parallel efficiency.

MOTIVATION: Deep metagenomic sequencing of biological samples has the potential to recover otherwise difficult-to-detect microorganisms and accurately characterize biological samples with limited prior knowledge of sample contents. Existing metagenomic taxonomic classification algorithms, however, do not scale well to analyze large metagenomic datasets, and balancing classification accuracy with computational efficiency presents a fundamental challenge. RESULTS: A method is presented to shift computational costs to an off-line computation by creating a taxonomy/genome index that supports scalable metagenomic classification. Scalable performance is demonstrated on real and simulated data to show accurate classification in the presence of novel organisms on samples that include viruses, prokaryotes, fungi and protists. Taxonomic classification of the previously published 150 giga-base Tyrolean Iceman dataset was found to take <20 h on a single node 40 core large memory machine and provide new insights on the metagenomic contents of the sample. AVAILABILITY: Software was implemented in C++ and is freely available at http://sourceforge.net/projects/lmat CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Conditional random fields (CRFs), which are popular supervised learning models for many natural language processing (NLP) tasks, typically require a large collection of labeled data for training. In practice, however, manual annotation of text documents is quite costly. Furthermore, even large labeled training sets can have arbitrarily limited performance peaks if they are not chosen with care. This paper considers the use of multi-criterion active learning for identification of a small but sufficient set of text samples for training CRFs. Our empirical results demonstrate that our method is capable of reducing the manual annotation costs, while also limiting the retraining costs that are often associated with active learning. In addition, we show that the generalization performance of CRFs can be enhanced through judicious selection of training examples

We discuss parallel performance of structured adaptive mesh refinement calculations using the SAMRAI library. We focus on fundamental aspects of adaptive gridding and dynamic computation of changing data dependencies. Previous analysis of performance of large-scale parallel adaptive calculations revealed poor scaling in these operations. Specifically, we found that these operations are inexpensive for small problems, but that their costs can become unacceptable for problems run on large numbers of processors. This paper describes subsequent developments involving graph- and tree-based algorithms that reduce runtime complexity and substantially increase scalability. We characterize performance on realistic adaptive problems using up to 512 processors of an IBM SP system and up to 1024 processors of a Linux cluster.

We report the development of a parallel algorithm for computing ILU preconditioners. The algorithm attains a high degree of parallelism through employmentofatwo-level ordering strategy, coupled with a subdomain graph constraint that regulates the location of nonzeros in the Schur complement. Experimental results include timings on four parallel platforms, for problems with up to 20 million unknowns running on up to 216 processors. The results support our theoretic analysis that the algorithm is highly scalable, for both preconditioner computation #factorization# and application #triangular solve# stages.