• Aucun résultat trouvé

How NASA experiments with knowledge discovery

N/A
N/A
Protected

Academic year: 2022

Partager "How NASA experiments with knowledge discovery"

Copied!
21
0
0

Texte intégral

(1)

SAS founded in 2013 in Paris | http://linkurio.us | @linkurious

How NASA

experiments with

knowledge discovery

(2)

Graph visualization and analysis startup founded in 2013.

40+ clients in 20+ countries (NASA, Cisco, French Ministry of Finances).

Linkurious Enterprise and Linkurious SDK.

Who we are

(3)

PUBLIC DATA INTERNAL DATA

PROPRIETARY DATA

Companies painfully connect the dots in complex data

(4)

?

But non-technical analysts cannot find key connections

BUSINESS ANALYSTS

(5)

We’ve solved this

problem for the...

(6)

FIND INSIGHTS IN CONNECTED DATA

Linkurious Enterprise helps analysts

(7)

What a fraud ring looks like

(8)

NASA:

How to provide efficient

access to critical data?

(9)

NASA: How to provide efficient access to critical data?

Problem

Leverage an existing lessons learned database to improve

the success of projects.

Benefit

More natural way to search content results in better insights derived from data.

Background

NASA is the United States government agency responsible for the civilian

space program.

(10)
(11)

Public database of lessons learned - llis.nasa.gov

(12)

Example of a lesson

(13)

Search engines are not efficient

David Meza,

Chief Knowledge Architect

A project engineer asked me if we could search our lessons learned using a list of 22 key terms the team was interested in. Our current keyword search

engine would require him to search our entire corpus of 20 million URLs. He would have to enter each

term individually, select the link and save the

document for review.

(14)

Relevance is highly context-dependent

(15)

Approach: visual search of connected documents

Topic modelling

Extract main topic of each lesson

Graph database

Neo4j graph database Easy integration in IT systems

Visualization

Full-text search Visual exploration of document correlations

(16)

Mapping knowledge with topic modelling

2,000 lessons extracted from the public NASA Engineering Network lesson learned database.

(17)

Graph model of the NASA lessons learned database

(18)

Linkurious reveals connections among lessons

(19)

Conclusion

Status

Still an experiment for NASA’s Strategy for Critical Data Visibility Through KM. Next:

user study & large-scale deployment.

Expectations

“A more effective search experience, reducing time to

find answers and allowing users to start their project on

the right foot.”

Background

NASA needs to provide efficient access to critical

data such as the lessons learned database.

(20)

Question?

(21)

Resources

● Linkurious: https://linkurio.us

● NASA public lessons learned database: https://llis.nasa.gov/

● http://km.nasa.gov/on-developing-better-magnets-for-finding-needles-in-haystacks/

● https://linkurio.us/how-nasa-experiments-with-knowledge-discovery/

● http://neo4j.com/blog/nasa-lesson-learned-database-using-neo4j-linkurious/

● NASA Strategy for Critical Data Visibility Through KM https://www.youtube.com/watch?v=vwJyU9vsfmU

● https://github.com/davidmeza1/doctopics

Références

Documents relatifs

Therefore, beyond checking fea- tures which are mainly related to the IEEE802.15.4 TSCH MAC [17], during the Plugtests event, other tests were performed for checking the RPL

In the following the dimensions of social capital are studied in three different contexts; business, virtual, and higher education in order to illuminate

[3] Jean-Andr´e Benvenuti, Laure Berti–quille & ´ Eric Jacopin, The Sabre Project: an Intelligent Tu- toring System for Military Behaviours, Poster at the CNRS Summer

In 2003 and 2004, Mohammed Zaki and myself organised the Frequent Itemset Mining Implementations (FIMI) workshops in an attempt to characterize and understand the

Nevertheless, the reinforcement learning networks (RL networks) found their application in the other compo- nent of AlphaGo, the value network, which is used to approximate the

It requires linking equipment effectiveness (OEE). Although largely technology-driven, this challenge also comprises KM procedures e.g. for drawing the right conclusions from

Another chal- lenge will be to deal with the long, transient climate simulations (thousands of years of model data) generated by the PMIP4 experi- ments (deglaciation, the Eemian

The following sections present each of the facets of this deliberative layer: how knowledge is produced asynchronously from geometric reasoning, how the robot conducts