Publications
AI enabling technologies: a survey
May 8, 2019
Journal Article
Published in:
https://arxiv.org/abs/1905.03592
Summary
Artificial Intelligence (AI) has the opportunity to revolutionize the way the United States Department of Defense (DoD) and Intelligence Community (IC) address the challenges of evolving threats, data deluge, and rapid courses of action. Developing an end-to-end artificial intelligence system involves parallel development of different pieces that must work together in order to provide capabilities that can be used by decision makers, warfighters and analysts. These pieces include data collection, data conditioning, algorithms, computing, robust artificial intelligence, and human-machine teaming. While much of the popular press today surrounds advances in algorithms and computing, most modern AI systems leverage advances across numerous different fields. Further, while certain components may not be as visible to end-users as others, our experience has shown that each of these interrelated components play a major role in the success or failure of an AI system. This article is meant to highlight many of these technologies that are involved in an end-to-end AI system. The goal of this article is to provide readers with an overview of terminology, technical details and recent highlights from academia, industry and government. Where possible, we indicate relevant resources that can be used for further reading and understanding.
Summary
Artificial Intelligence (AI) has the opportunity to revolutionize the way the United States Department of Defense (DoD) and Intelligence Community (IC) address the challenges of evolving threats, data deluge, and rapid courses of action. Developing an end-to-end artificial intelligence system involves parallel development of different pieces that must work together...
READ MORE
A billion updates per second using 30,000 hierarchical in-memory D4M databases
January 24, 2019
Conference Paper
Published in:
North East Database Day, NEDB Day, 24 January 2019.
Topic:
R&D area:
R&D group:
Summary
Analyzing large scale networks requires high performance streaming updates of graph representations of these data. Associative arrays are mathematical objects combining properties of spreadsheets, databases, matrices, and graphs, and are well-suited for representing and analyzing streaming network data. The Dynamic Distributed Dimensional Data Model (D4M) library implements associative arrays in a variety of languages (Python, Julia, and Matlab/Octave) and provides a lightweight in-memory database. Associative arrays are designed for block updates. Streaming updates to a large associative array requires a hierarchical implementation to optimize the performance of the memory hierarchy. Running 34,000 instances of a hierarchical D4M associative arrays on 1,100 server nodes on the MIT SuperCloud achieved a sustained update rate of 1,900,000,000 updates per second. This capability allows the MIT SuperCloud to analyze extremely large streaming network data sets.
Summary
Analyzing large scale networks requires high performance streaming updates of graph representations of these data. Associative arrays are mathematical objects combining properties of spreadsheets, databases, matrices, and graphs, and are well-suited for representing and analyzing streaming network data. The Dynamic Distributed Dimensional Data Model (D4M) library implements associative arrays in...
READ MORE
Hyperscaling internet graph analysis with D4M on the MIT SuperCloud
September 26, 2018
Conference Paper
Published in:
IEEE High Performance Extreme Computing Conf., HPEC, 25-27 September 2018.
Topic:
R&D area:
R&D group:
Summary
Detecting anomalous behavior in network traffic is a major challenge due to the volume and velocity of network traffic. For example, a 10 Gigabit Ethernet connection can generate over 50 MB/s of packet headers. For global network providers, this challenge can be amplified by many orders of magnitude. Development of novel computer network traffic analytics requires: high level programming environments, massive amount of packet capture (PCAP) data, and diverse data products for "at scale" algorithm pipeline development. D4M (Dynamic Distributed Dimensional Data Model) combines the power of sparse linear algebra, associative arrays, parallel processing, and distributed databases (such as SciDB and Apache Accumulo) to provide a scalable data and computation system that addresses the big data problems associated with network analytics development. Combining D4M with the MIT SuperCloud manycore processors and parallel storage system enables network analysts to interactively process massive amounts of data in minutes. To demonstrate these capabilities, we have implemented a representative analytics pipeline in D4M and benchmarked it on 96 hours of Gigabit PCAP data with MIT SuperCloud. The entire pipeline from uncompressing the raw files to database ingest was implemented in 135 lines of D4M code and achieved speedups of over 20,000.
Summary
Detecting anomalous behavior in network traffic is a major challenge due to the volume and velocity of network traffic. For example, a 10 Gigabit Ethernet connection can generate over 50 MB/s of packet headers. For global network providers, this challenge can be amplified by many orders of magnitude. Development of...
READ MORE
Large-scale Bayesian kinship analysis
September 25, 2018
Conference Paper
Published in:
IEEE High Performance Extreme Computing Conf., HPEC, 25-27 September 2018.
Summary
Kinship prediction in forensics is limited to first degree relatives due to the small number of short tandem repeat loci characterized. The Genetic Chain Rule for Probabilistic Kinship Estimation can leverage large panels of single nucleotide polymorphisms (SNPs) or sets of sequence linked SNPs, called haploblocks, to estimate more distant relationships between individuals. This method uses allele frequencies and Markov Chain Monte Carlo methods to determine kinship probabilities. Allele frequencies are a crucial input to this method. Since these frequencies are estimated from finite populations and many alleles are rare, a Bayesian extension to the algorithm has been developed to determine credible intervals for kinship estimates as a function of the certainty in allele frequency estimates. Generation of sufficiently large samples to accurately estimate credible intervals can take significant computational resources. In this paper, we leverage hundreds of compute cores to generate large numbers of Dirichlet random samples for Bayesian kinship prediction. We show that it is possible to generate 2,097,152 random samples on 32,768 cores at a rate of 29.68 samples per second. The ability to generate extremely large number of samples enables the computation of more statistically significant results from a Bayesian approach to kinship analysis.
Summary
Kinship prediction in forensics is limited to first degree relatives due to the small number of short tandem repeat loci characterized. The Genetic Chain Rule for Probabilistic Kinship Estimation can leverage large panels of single nucleotide polymorphisms (SNPs) or sets of sequence linked SNPs, called haploblocks, to estimate more distant...
READ MORE
Interactive supercomputing on 40,000 cores for machine learning and data analysis
September 25, 2018
Conference Paper
Published in:
IEEE High Performance Extreme Computing Conf., HPEC, 25-27 September 2018.
Topic:
R&D area:
R&D group:
Summary
Interactive massively parallel computations are critical for machine learning and data analysis. These computations are a staple of the MIT Lincoln Laboratory Supercomputing Center (LLSC) and has required the LLSC to develop unique interactive supercomputing capabilities. Scaling interactive machine learning frameworks, such as TensorFlow, and data analysis environments, such as MATLAB/Octave, to tens of thousands of cores presents many technical challenges – in particular, rapidly dispatching many tasks through a scheduler, such as Slurm, and starting many instances of applications with thousands of dependencies. Careful tuning of launches and prepositioning of applications overcome these challenges and allow the launching of thousands of tasks in seconds on a 40,000-core supercomputer. Specifically, this work demonstrates launching 32,000 TensorFlow processes in 4 seconds and launching 262,000 Octave processes in 40 seconds. These capabilities allow researchers to rapidly explore novel machine learning architecture and data analysis algorithms.
Summary
Interactive massively parallel computations are critical for machine learning and data analysis. These computations are a staple of the MIT Lincoln Laboratory Supercomputing Center (LLSC) and has required the LLSC to develop unique interactive supercomputing capabilities. Scaling interactive machine learning frameworks, such as TensorFlow, and data analysis environments, such as...
READ MORE
GraphChallenge.org: raising the bar on graph analytic performance
September 25, 2018
Conference Paper
Published in:
IEEE High Performance Extreme Computing Conf., HPEC, 25-27 September 2018.
Summary
The rise of graph analytic systems has created a need for new ways to measure and compare the capabilities of graph processing systems. The MIT/Amazon/IEEE Graph Challenge has been developed to provide a well-defined community venue for stimulating research and highlighting innovations in graph analysis software, hardware, algorithms, and systems. GraphChallenge.org provides a wide range of preparsed graph data sets, graph generators, mathematically defined graph algorithms, example serial implementations in a variety of languages, and specific metrics for measuring performance. Graph Challenge 2017 received 22 submissions by 111 authors from 36 organizations. The submissions highlighted graph analytic innovations in hardware, software, algorithms, systems, and visualization. These submissions produced many comparable performance measurements that can be used for assessing the current state of the art of the field. There were numerous submissions that implemented the triangle counting challenge and resulted in over 350 distinct measurements. Analysis of these submissions show that their execution time is a strong function of the number of edges in the graph, Ne, and is typically proportional to N4=3 e for large values of Ne. Combining the model fits of the submissions presents a picture of the current state of the art of graph analysis, which is typically 108 edges processed per second for graphs with 108 edges. These results are 30 times faster than serial implementations commonly used by many graph analysts and underscore the importance of making these performance benefits available to the broader community. Graph Challenge provides a clear picture of current graph analysis systems and underscores the need for new innovations to achieve high performance on very large graphs.
Summary
The rise of graph analytic systems has created a need for new ways to measure and compare the capabilities of graph processing systems. The MIT/Amazon/IEEE Graph Challenge has been developed to provide a well-defined community venue for stimulating research and highlighting innovations in graph analysis software, hardware, algorithms, and systems...
READ MORE
Measuring the impact of Spectre and Meltdown
September 25, 2018
Conference Paper
Published in:
IEEE High Performance Extreme Computing Conf., HPEC, 25-27 September 2018.
Topic:
R&D area:
R&D group:
Summary
The Spectre and Meltdown flaws in modern microprocessors represent a new class of attacks that have been difficult to mitigate. The mitigations that have been proposed have known performance impacts. The reported magnitude of these impacts varies depending on the industry sector and expected workload characteristics. In this paper, we measure the performance impact on several workloads relevant to HPC systems. We show that the impact can be significant on both synthetic and realistic workloads. We also show that the performance penalties are difficult to avoid even in dedicated systems where security is a lesser concern.
Summary
The Spectre and Meltdown flaws in modern microprocessors represent a new class of attacks that have been difficult to mitigate. The mitigations that have been proposed have known performance impacts. The reported magnitude of these impacts varies depending on the industry sector and expected workload characteristics. In this paper, we...
READ MORE
Colorization of H&E stained tissue using deep learning
July 17, 2018
Conference Paper
Published in:
40th Int. Conf. of the IEEE Engineering in Medicine and Biology Society, EMBC, 17-21 July 2018.
Topic:
R&D area:
R&D group:
Summary
Histopathology is a critical tool in the diagnosis and stratification of cancer. Digital Pathology involves the scanning of stained and fixed tissue samples to produce high-resolution images that can be used for computer-aided diagnosis and research. A common challenge in digital pathology related to the quality and characteristics of staining, which can vary widely from center to center and also within the same institution depending on the age of the stain and other human factors. In this paper we examine the use of deep learning models for colorizing H&E stained tissue images and compare the results with traditional image processing/statistical approaches that have been developed for standardizing or normalizing histopathology images. We adapt existing deep learning models that have been developed for colorizing natural images and compare the results with models developed specifically for digital pathology. Our results show that deep learning approaches can standardize the colorization of H&E images. The performance as measured by the chi-square statistic shows that the deep learning approach can be nearly as good as current state-of-the art normalization methods.
Summary
Histopathology is a critical tool in the diagnosis and stratification of cancer. Digital Pathology involves the scanning of stained and fixed tissue samples to produce high-resolution images that can be used for computer-aided diagnosis and research. A common challenge in digital pathology related to the quality and characteristics of staining...
READ MORE
Lessons learned from a decade of providing interactive, on-demand high performance computing to scientists and engineers
June 28, 2018
Conference Paper
Published in:
ISC High Performance Int. Workshops, Lecture Notes in Computer Science (LNSC 11203), 28 June 2018, pp. 655-68.
Topic:
R&D area:
R&D group:
Summary
For decades, the use of HPC systems was limited to those in the physical sciences who had mastered their domain in conjunction with a deep understanding of HPC architectures and algorithms. During these same decades, consumer computing device advances produced tablets and smartphones that allow millions of children to interactively develop and share code projects across the globe. As the HPC community faces the challenges associated with guiding researchers from disciplines using high productivity interactive tools to effective use of HPC systems, it seems appropriate to revisit the assumptions surrounding the necessary skills required for access to large computational systems. For over a decade, MIT Lincoln Laboratory has been supporting interactive, on demand high performance computing by seamlessly integrating familiar high productivity tools to provide users with an increased number of design turns, rapid prototyping capability, and faster time to insight. In this paper, we discuss the lessons learned while supporting interactive, on-demand high performance computing from the perspectives of the users and the team supporting the users and the system. Building on these lessons, we present an overview of current needs and the technical solutions we are building to lower the barrier to entry for new users from the humanities, social, and biological sciences.
Summary
For decades, the use of HPC systems was limited to those in the physical sciences who had mastered their domain in conjunction with a deep understanding of HPC architectures and algorithms. During these same decades, consumer computing device advances produced tablets and smartphones that allow millions of children to interactively...
READ MORE
Detecting pathogen exposure during the non-symptomatic incubation period using physiological data
November 13, 2017
Journal Article
Published in:
bioRxiv preprint doi: https://doi.org/10.1101/218818, 13 November 2017.
Summary
Early pathogen exposure detection allows better patient care and faster implementation of public health measures (patient isolation, contact tracing). Existing exposure detection most frequently relies on overt clinical symptoms, namely fever, during the infectious prodromal period. We have developed a robust machine learning based method to better detect asymptomatic states during the incubation period using subtle, sub-clinical physiological markers. Starting with highresolution physiological waveform data from non-human primate studies of viral (Ebola, Marburg, Lassa, and Nipah viruses) and bacterial (Y. pestis) exposure, we processed the data to reduce short-term variability and normalize diurnal variations, then provided these to a supervised random forest classification algorithm and post-classifier declaration logic step to reduce false alarms. In most subjects detection is achieved well before the onset of fever; subject cross-validation across exposure studies (varying viruses, exposure routes, animal species, and target dose) lead to 51h mean early detection (at 0.93 area under the receiver-operating characteristic curve [AUCROC]). Evaluating the algorithm against entirely independent datasets for Lassa, Nipah, and Y. pestis exposures un-used in algorithm training and development yields a mean 51h early warning time (at AUCROC=0.95). We discuss which physiological indicators are most informative for early detection and options for extending this capability to limited datasets such as those available from wearable, non-invasive, ECG-based sensors.
Summary
Early pathogen exposure detection allows better patient care and faster implementation of public health measures (patient isolation, contact tracing). Existing exposure detection most frequently relies on overt clinical symptoms, namely fever, during the infectious prodromal period. We have developed a robust machine learning based method to better detect asymptomatic states...
READ MORE