project chronos presentation - machine learning italy meetup in turin

18
Project Chronos Machine Learning Italy Turin, 3 December 2014 projectchronos.org

Upload: jacopo-durandi

Post on 13-Jul-2015

211 views

Category:

Science


3 download

TRANSCRIPT

Project Chronos

Machine Learning Italy

Turin, 3 December 2014

projectchronos.org

Project Chronos’ Milestones

➔ 11 – 13 April '14: NASA's Hackathon SpaceApps 2014 the Group is born: 2nd Place, “Best Mission Concept”.

➔ 6th May '14: Global Award Finalists for the category “Best Mission Concept”

➔ October ‘14: First Chronos Knowledge Base and Xploration video-animation is realized

➔ ➔ Nov ‘14: first collaboration on Climate Change monitoring. Project Chronos partners with iSeeChange

Our Mission

Giving people new tools to research knowledge.

A Citizen Scientist.

Everyone can be a scientist.

Our Mission

What do we know about space?

How do we look at it?

Knowledge about the Universe relies mostly on the analysis of the electromagnetic radiation collected from Space. But also on particle detection, plasma

investigation, soil analysis of the celestial bodies of the Solar System, and so on.

What can a spacecraft see, and how?

Spacecraft house the vast majority of detectors and sensors capable of performing measurements on

physical entities in Space and around the Earth.

An example Rosetta’s OrbiterCredits: Rosetta Blog

Rosetta’s Orbiter and its instrumentation

Spectrometer: detects visible &

infrared light

Working mechanism?

Scientific Goals?

Online datasets?

Credits: ESA

MAIN CHALLENGES IN IMPLEMENTATION

HOW TO DESCRIBE AEROSPACE KNOWLEDGE:

Agencies’ Open-dataW3C Standards

HOW TO COLLECT AND STORE DATA:

No-SQLPython and libraries

Crowd-sourcing (citizen science)“machine-learned crawling”?

Semantic Web: Knowledge Base

NASA-STI Open-Data: a hierarchical taxonomy containing terms and concepts for aerospace technical and scientific papers. Divisions > Subjects > Scopes > Keywords

Chronos Domain: Missions, Data, URLs and News from Agencies websites.

Chronos Sensors: A Taxonomy For Sensors

AEROSPACE ACTIVITIES WITH W3C STANDARDS

Semantic Web: Ontology Development

Chronos Domain: Class and Relations designed for describing Missions, Objectives, Destinations, DBpedias’, etc.

Chronos Sensors: Taxonomy of Sensors, to describe astrophysics and other scientific activities in space

Semantic Web: Knowledge Base

RDF (w3c)SKOS (w3c)OWL (w3c)SCHEMA.ORG (schema)

TRIPLES-TURTLE (data format)XML (data format)

JSON-LD (data format)MONGOdb (datacloud)PYTHON3.3 (scripting)

Most meaningful semantic interactions (schema)

keywords subjects

Graph representation printed with NetworkX Python Library

XplorationApp: Mining the Cloud and learn-by-playing

https://www.youtube.com/watch?v=ZMFZv_yYCBI or look in YouTube for “XplorationApp Project Chronos”

creare delle procedure semi-automatizzate (crowd-sourcing dei beta-testers più algoritmi) per il crawling e l’ampliamento della knowledge base?

migliorare le procedure semi-automatizzate per il riconoscimento semantico dei termini nelle pagine Web e loro automatico salvataggio nella knowledge base con le corrette relazioni? (natural language processing)

HELP!

Come è possibile utilizzare le tecniche machine-learning per:

proposte per unsupervised machine-learning?

GRAZIE PER L’ATTENZIONE

>>>> cominceremo a breve un funding round per la fase di seeding. Potete richiedere di ricevere una copia del business plan.

>>>> stiamo pensando anche ad una campagna di crowd-funding su una delle maggiori piattaforme a livello globale

>>> Cerchiamo programmatori Python e Javascript Full-Stack, nonché data scientists che vogliano scommettere sul progetto e cresecere con noi

>>>> [email protected]

projectchronos.org

>>>> [email protected]