BISCDSS - PowerPoint PPT Presentation

Loading...

PPT – BISCDSS PowerPoint presentation | free to download - id: 113647-M2E1M



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

BISCDSS

Description:

How to transform Earth Science data into transactions? ... Knowledge Mobilization and Intelligent Augmentation (KnowMInA) Application to Health Care and ... – PowerPoint PPT presentation

Number of Views:52
Avg rating:3.0/5.0
Slides: 35
Provided by: masoudni
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: BISCDSS


1
Fuzzy Set 1965 Fuzzy Logic 1973 Soft
Decision 1981 BISC 1990 Human-Machine
Perception 2000 -
BISC Program Masoud Nikravesh BISC Program,
EECS-UCB Imaging and Informatics Life
Sciences Lawrence Berkeley National Laboratory
(LBNL) http//www-bisc.cs.berkeley.edu/ Email
Nikravesh_at_cs.berkeley.edu Tel (510) 643-4522
Fax (510) 642-5775 Member of Executive
Committee UC Discovery (Appointed by Provost and
Senior Vice President) Member of Research
Council UC Discovery (Appointed by Provost and
Senior Vice President) LBNL-NERSC Representative
in UC Discovery Program (Appointed by NERSC
Director)
BISC
2
BISC-UCB EECS-CS Division-UCB
  • BISC Program is the world-leading center for
    basic and applied research in soft computing.
  • The principal constituents of soft computing (SC)
    are fuzzy logic (FL), neural network theory (NN),
    Evolutionary computing and probabilistic
    reasoning (PR), including DNA computing, chaos
    theory and parts of machine learning theory.

3
BISC Program Special Interest Groups
  • BISC SIG in Anticipation
  • BISC SIG in Computational Intelligence for
    Bioinformatics
  • BISC SIG in Biotechnology
  • BISC SIG in Communication and Networking
    (Wireless)
  • BISC SIG in Data Mining
  • BISC SIG in Decision Analysis and Support System
  • BISC SIG in Earth Sciences
  • BISC SIG in Economics
  • BISC SIG in Entertainments
  • BISC SIG in Environment Management Systems
  • BISC SIG in Fuzzy Logic and the Internet
  • BISC SIG in Granular Computing
  • BISC SIG in Healthcare
  • BISC SIG in Homeland Security
  • BISC SIG in Information Technology
  • BISC SIG on Intelligent Agents in Complex Systems
  • BISC SIG on Intelligent Manufacturing and Fault
    Diagnosis
  • BISC SIG in Life Sciences
  • BISC SIG in Medicine

4
EVOLUTION OF FUZZY LOGIC
generality
nl-generalization
computing with words and perceptions (CWP)
f.g-generalization
f-generalization
classical bivalent
time
1965
1973
1999
1965 crisp sets fuzzy sets 1973 fuzzy
sets granulated fuzzy sets (linguistic
variable) 1999 measurements perceptions
Fuzzy Set 1965 Fuzzy Logic 1973 Soft
Decision 1981 BISC 1990 Human-Machine
Perception 2000 -
5
(No Transcript)
6
Factual Information About the Impact of Fuzzy
Logic
FACTs
7
  • January 26, 2005
  • Factual Information About the Impact of Fuzzy
    Logic
  •  
  •  PATENTS
  • -- Number of fuzzy-logic-related patents applied
    for in Japan 17,740
  • -- Number of fuzzy-logic-related patents issued
    in Japan  4,801
  • -- Number of fuzzy-logic-related patents issued
    in the US around 1,700

8
Applications and Challenges
Applications
New Challenges
9
Homeland Security
Internet
NLC Search Engine
Cybersecurity CIP
Planetary Sciences
Earth Sciences
Genome
DNA Genetic
Geophysical Data Oil Industry
Satellite Images NASA
10
Center for Computational Machine Intelligence and
Systems Science
11
Center for Computational Machine Intelligence and
Systems Science
12
Center for Computational Machine Intelligence and
Systems Science
  • Scientific instruments and numerical simulation
    models generate massive amounts of multi-spectral
    spatio-temporal image data
  • Difficult to visualize and summarize
  • With meaningful high level representation

data
Multi-level representation
For example, simulated sea surface temperature
data at a typical ocean model resolution of 100km
(horizontal grid size of 384x320) from a 100 year
simulation run with 24 hour sampling interval
reveals a matrix of 36,500-by-36,500. (through
singular value decomposition), which is roughly
1.3 Tera points. Even at this resolution,
standard principal component analysis (PCA) may
require sampling in high performance computing
environment.
Predictive models
Feature-based query
Quantitative analysis
Feature-based visualization
Comparative analysis
B. Parvin and Masoud Nikravesh vision.lbl.gov
13
Center for Computational Machine Intelligence and
Systems Science
Feature-based Representation of Spatio-temporal
data
Feature-based representation of Oceanography
simulation data
Feature-based representation of core collapse
super nova simulation data
B. Parvin and Masoud Nikravesh vision.lbl.gov
14
Center for Computational Machine Intelligence and
Systems Science
  • Tera-Flop computation of scientific data are
    going to be routine
  • For example, simulated sea surface temperature
    data at a typical ocean model resolution of 100km
    (horizontal grid size of 384x320) from a 100 year
    simulation run with 24 hour sampling interval
    reveals a matrix of 36,500-by-36,500. (through
    singular value decomposition), which is roughly
    1.3 Tera points. Even at this resolution,
    standard principal component analysis (PCA) may
    require sampling in high performance computing
    environment.
  • An example is CCSM3, which was used to provide a
    suite of simulations for the Fourth Assessment
    report of the Intergovernmental Panel on Climate
    Change (IPCC). CCSM3 provided simulations with
    the unprecedented atmosphere resolution of 180km.
    Over 7.5 TB of data are produced in each
    100year simulation of this model, with most data
    output only at monthly intervals. Several
    100-year integrations are required to simulate
    all of the IPCC future emission scenarios.

Astrophysics Now and 5 yrs Can soak up
anything! Fusion Now 100Mbytes/15min 5 yrs
1000Mbytes/2 min
Climate Now 20-40TB per simulated year 5 yrs
100TB/yr 5-10PB/yr
15
Center for Computational Machine Intelligence and
Systems Science
  • Algorithmic Complexity
  • Calculate Means O(n)
  • Calculate FFT O(n log(n))
  • Calculate PCA O(r c)
  • Algorithmic Complexity
  • Kernelized CCA O(n m2)
  • Hierarchical Clust. O(n2)
  • Decision Tree Induction O(mn log n) O(n (log
    n)2)

16
Mining Associations in Earth Science Data
Challenges
  • How to transform Earth Science data into
    transactions?

17
Center for Computational Machine Intelligence and
Systems Science
Vipin Kumar
18
Center for Computational Machine Intelligence and
Systems Science
19
Center for Computational Machine Intelligence and
Systems Science
Organization Dr. Bahram Parvin (LBNL, Adj.
Professor of Electrical Engineering U.C.
Riverside) will be the lead PI. He has a
long-standing experience in scientific data
understanding and information management. Dr.
Raymond Bair (Co-PI, ANL, University of Chicago)
is the Director of the Laboratory Computing
Resource Center and manages the ANL climate
group. Professor Shankar Sastry (Co-PI, U.C.
Berkeley) is the Director of CITRUS (Center for
Information Technology Research in the Interest
of Society) with a long- standing research
credential in the area of system theory,
non-linear control and machine learning. He will
coordinate his research with Dr. Masoud Nikravesh
(Co-PI, LBNL and U.C. Berkeley, Executive
director of Berkeley Institute of Soft Computing)
and Professor Michael Jordon (U.C.B.), who is
well-established in the are area of machine
learning and kernel-based methods. Dr. Jacob
Barhen (Co-PI, ORNL, University of Tennessee) is
the Director of Center for Engineering Science
with a strong background on machine learning and
uncertainty analysis. Dr. Will Schroeder (Co-PI)
is the founder and chief technical officer at
Kitware, Inc. Each member of this team brings
forth a unique dimension to the proposed
enterprise from Computer Science, Computational
Science, Parallel Processing, Machine Learning,
and uncertainty analysis.
20
Center for Computational Machine Intelligence and
Systems Science
  • Specific Aims. These components form a tightly
    orchestrated set of software capability
    developments by the consortium members
  • LBNL has two specific aims
  • Development of geometric techniques for
    feature-based representation of spatio-temporal
    data that are generated from science applications
    such as Climate modeling and observational
    platforms. These representations will enable both
    short and long term cyclical events for higher
    level analysis and visualization.
  • Development of an informatics system for
    multidimensional representation of summary data
    computed from each data variable of
    spatio-temporal data. Such a system can offer
    information visualization through different
    dimensions of computed representations, enable
    reasoning, and causal analysis. The informatics
    module is one facet of the interaction with
    Kitware, Inc. and other members of the
    consortium.
  • UCB has three specific aims
  • Application and extension of generalized
    principal component analysis for saliency
    detection in each variable field (e.g.,
    temperature, pressure) and its motion field. The
    focus will be incorporating time-varying events
    into a generalized PCA (GPCA) framework.
  • Application, extension, and development of
    kernel-based methods for detection and
    localization of coupled events from multiple data
    streams for dimensionality reduction as well as
    clustering. While GPCA of aim 2a operates on a
    single variable space of raw data, kernel methods
    will aim at coupled feature-base representation.
  • Development of a framework for reasoning and
    association for coupled data sets. This framework
    will offer an improved factual, evidential
    reasoning, and knowledge organization, which can
    potentially contribute to knowledge discovery.
  • ORNL has three specific aims
  • Develop innovative methods and tools for
    quantification, propagation, and reduction of
    uncertainty in combining information from
    multiscale, spatio-temporal simulations and
    corresponding experimental data. This paradigm
    will be referred to as Multiscale Uncertainty
    Analysis (MUA).
  • Develop the mathematical and software
    capabilities enabling transparent interface of
    MUA to the feature representation framework
    developed by LBNL, and apply these methods in
    conjunction with the machine learning techniques
    developed at UCB.
  • Demonstrate and assess MUA on a model of interest
    to the SciDAC effort (e.g., climate).
  • Argonne has four specific aims
  • Improve performance of computational modules and
    algorithms developed in aims 1 and 2. This will
    be in collaboration with Kitware, Inc.
  • Provide climate model output data for testing and
    construct the Grid interface for querying and
    streaming pertinent data from data servers.
  • Integrate, validate, and compare new
    feature-based informatics systems with
    traditional climate model diagnostics and
    evaluation methods.
  • Examine the feasibility of runtime generation of
    features and associated statistics as a way to
    reduce total output from high resolution
    simulations.
  • Kitware has one specific aim and one core
    functionality
  • (aim) Leverage and extend the existing VTK and
    ITK toolkits to support advanced methods for
    visualization, data analysis, distributed
    parallel processing. New template libraries will
    be developed within ITK to support (i) data
    analysis in spherical coordinate system (as
    dictated by climate data) and (ii) MPI
    implementation of a certain class of algorithms.

21
Center for Computational Machine Intelligence and
Systems Science
22
Energy Earth Sciences Intelligent Reservoir
Characterization
IRESC
23
(No Transcript)
24
IRESC
25
Transaction and Support Definitions
  • time series gt Tables

Vipin Kumar
26
Mining Associations in Earth Science Data
Challenges
27
BISC DSS Components and Structure
Model and Data Visualization
  • Model Management
  • Query
  • Aggregation
  • Ranking
  • Fitness Evaluation

Evolutionary Kernel Genetic Algorithm, Genetic
Programming, and DNA
  • Selection
  • Cross Over
  • Mutation

Input From Decision Makers
Experts Knowledge
Model Representation Including Linguistic
Formulation
Data Management
  • Functional Requirements
  • Constraints
  • Goals and Objectives
  • Linguistic Variables Requirement

28
BISC-DSS Tool i.e. Intelligent Managment
Systems i.e. Customer Satisfaction Software
System Medical Diagnosis Prognasis Media
Directed Advertising Automated Sensory Inspection
System
29
Natural Language Computing
  • FC-DNA as a basis for
  • Common Sense Knowledge, Human Reasoning and
    Deduction
  • Next Generation of Concept-Based Search Engine

PRUF C-DNA gt Foundation for Chinese NLP
PRUF C-DNA gt Human Reasoning and Deduction
Capability
Chinese-DNA gt M(x) Fuzzy Logic gt Human
Reasoning Fuzzy Logic gt Reasoning with
C-DNA Z-Compact gt Fine Tune C-DNA
Tree Z-Compact gt C-DNA M(x) C-DNA
Z-Compact gt Chinese NLC Rules
30
KnowMInA
Knowledge Mobilization and Intelligent
Augmentation (KnowMInA) Application to Health
Care and Defense Security (Homeland Security)
KnowMInA
31
KnowMInA
32
Berkeley International Institute for System and
Computational Intelligence
33
(No Transcript)
34
(No Transcript)
About PowerShow.com