semantic technology for development: semantic web without the web?
TRANSCRIPT
Semantic Technology for Development: Semantic Web without the Web?
Victor de BoerVrije Universiteit Amsterdam
With input from: Anna Bon, Christophe Gueret, Stephane Boyera , Nana Baah Gyan, Chris van Aart, Max Froumentin, Aman Grewal, Mary Allen, Amadou Tangara, Etienne Barnard, Hans Akkermans, Julie
Ferguson, Marije Visscher,…
Victor de Boer
Web & Media Group, CS, VU University Amsterdam
Netherlands Institute for Sound and Vision
Web Science
Cultural Heritage and Digital History
ICT for Development
CAUTION! DIGITAL DIVIDE AHEAD
Internet users per continent
Source: International Telecommunication Union (ITU)
Linked Data for Development?
Can the (Semantic) Web (be made to) meansomething for knowledge sharing even undervery constraining conditions?
No internet, no computer, no electricity
Multitude of languages, levels of literacy
http://worldwidesemanticweb.org
Example: Landportal
http://landportal.info and http://landportal.sbc4d.com
Published by FAO
Example: Agrovoc/Agris
http://linked-development.org/
Example: Institute for Development Studies’ Eldis database
ELDIS Link to GeoNames
IDS: document 0002 Country:”Gambia”
Geonames:Gambia
Region: Africa
population : 1593256
N 13° 30' 0'' W 15° 30' 0'
Links to DBPedia
IDS: document 0001 Theme:”Food Security”
DBPedia:”Food Security”
Analysis of approaches to understanding and addressing food security issues; examination of the structural causes of food insecurity and different policy responses
Theme:” Food aid emergencies ”
Person:”David Pimentel”
Organisation:”FAO”
“Voedselzekerheid”@NL
IATI Sector:”Higher Education”
Organisation : UN Habitat
Activity: Multi donor fund to support civil society in democracy related issues
Aid received (USD)
Geonames:Gambia
Worldbank:Gambia
N 13° 30' 0'' W 15° 30' 0'
Example: Linked Data for International Aid Transparency Initiative (IATI)
Kasper Brandt
http://iati2lod.appspot.com/4. How does violent conflict in recipient countries affect aid activities?
5. How does aid spending as registered in the IATI standard compare to World Bank indicators?
Hm..
Information sharing should be made
1. usable on small, affordable, hardware deployed in various connectivity contexts;
2. accessible to individuals with varied cultural backgrounds / literacy levels;
3. relevant and directly useful to the target public they aim to empower.
Infrastructure
Interface Relevancy
Infrastructure
RelevancyInterface
https://www.flickr.com/photos/worldbank/7826373720
Mobile phones
Community radio
Low-literacy
One of the grand challenges of ICT4DPrevalent among the rural poor
Speech: The most natural user interface
http
://ww
w.h
olisticlc.co
m/sp
eech
-therap
y/
M-agro Use Case Context: Regreening in Africa
Tominian regionMali
Market Data store
GSM/Voice interfaceWeb Interface
Text-To-Speech / ASR
Radio users
Field operative
Users
Information / Communique Services
“Slot and Filler” Text-to-Speech
English:
Bambara:
15 liters of offered by Zakari Diarra
15_ba.wav L_ba.wav Of_ba.wav
Spoken Language Elements Repository
honey
Talking to LOD
Web applications
<VoiceXML> to SPARQL
Voice browser
RadioMarché Linked market data
‘Allo, Linked Data?
DBpediaGeoNames
Agrovoc
rm:offering0001
rm:shea_butter
rm:1000
rdfs:label
rdfs:label “Amande de Karité”@fr
“Shea Nuts”@en
speakle:voicelabel_ba
rm:audio_shea_nl.wav
rm:audio_shea_ba.wav
speakle:voicelabel_nl
rdfs:label“1000”
speakle:voicelabel_ba
rm:audio_1000_nl.wav
rm:audio_1000_ba.wavspeakle:voicelabel_nl
rm:Mazankuy_Diarra
rm:kilo
rdfs:label “kilo”@en
speakle:voicelabel_ba
rm:audio_kilo_nl.wav
rm:audio_kilo_ba.wav
speakle:voicelabel_nl
rm:has_contact
Speakle voice labels
RadioMarché leads to new questions, new actions, new knowledge
Honey sales doubled in Tominian since RadioMarche;Sheabutter sales lagged behind, despite the same approach in RadioMarche
Source: Sahel Eco Amadou Tangara internalEvaluation report EVP – Radio Marché
Infrastructure
Interface Relevancy
Weather Service (Ghana, Mali, Burkina Faso)
DigiVet (N-Ghana)
Radio message service (Mali)
Poultry vaccination service (Mali)
Milk service (Mali
Seed market (Mali and Burkina Faso)
Veterinary service (N-Ghana)
Knowledge base and diagnosis system
CommonKADS
Information sharing across locations
DigiVet
Mr. Meteo
Mali Milk
Linked Data’s flexible data model fits the ``downscaling’’ aspect of CS4D well
Data is locally relevant……but might be shared across regions……across locally co-developed applications……and aggregated to country/global level
Interface Relevancy
Infrastructure
http://www.nytimes.com/2008/12/16/health/16incubators.html?pagewanted=1&_r=3&partner=permalink&exprod=permalink
Appropriate, sustainable hardware
affordable,
low-power-consuming,
repairable,
durable,
ownable,
Total Cost of Ownership (TCO)
Example: Efficient Knowledge sharing with SemanticXO and ERS
Christophe Guéret
ENTITY REGISTRY SYSTEM (ERS)
• Fully decentralised Linked Data publication platform• Works under any kind of connectivity context
• Tracks back individual edits back to their authors• Simple and versatile• Open Source https://github.com/ers-devs• Low resource demanding
... and open for contributions so don'thesitate to fork it!
Voice service: Multiple solutions
SIP over Ethernet
HTTP
Officerouterunning Asterisk
Orange Emerginov Platform
GSM dongle
DO
WN
SCA
LIN
G
Cheap, small hardware
Foroba Blon 2.0
Kasadaka
Low-resource knowledge sharing platform
Low-power, ubiquitous, cheap hardware FLOSS components
Rapid prototyping + deployment of (knowledge-intensive) services
User InterfacesVoice servicesSMS-based Visual
Wifi, 3g or GSM network
RDF Data store (Linked Data) allows for flexible data integration across applications, deployments
Kasadaka (“talking box”)
Andre Baart
Machine to machine information exchange
GSM network
SPARQL in an SMS
Enable (Semantic) Web data exchange over GSM networks.
Practical differences HTTP and SMS:SMS works with phone number, HTTP works with URLs.SMS has a size restriction, HTTP practically has none.SMS is one-way messaging, HTTP follows request-response.
Basic M2M communication based on SPARQL.
Onno Valkering
SPARQL in an SMSConverters to translate SPARQL HTTP request to SMS message (140 or 160 chars) and vice versa
CONSTRUCT, INSERT/DELETE DATA
ChallengesBlending synchronous and asynchronous messaging
SPARQL/ RDF compression
Unpredictable query result sizes
Compression for small datasets experiments
Strategies
Different serializations: RDF/XML, N-triples, Turtle, HDT1, EXI2
Compression (zip)
Assume shared vocabularies (top 20 from prefix.cc) and remove redundant (inferenced) triples (RDFS reasoning)
Evaluated on real-world datasets LOD Laundromat3
232,822 small datasets (1-1,000 triples)
[1] Fernández et al. “Compact Representation of Large RDF Data Sets for publishing and exchange” (ISWC 2010)[2] Käbisch et al. “Standardized and Efficient RDF Encoding for Constrained Embedded Networks” (ESWC2015)[3] http://lodlaundromat.org/ and Rietveld et al. “LOD Lab: Experiments at LOD Scale “ (ISWC2015)
TriplesN-triples +gzip
RDF/ XML
RDF/XML+Gzip Turtle
Turtle+Gzip HDT
HDT+ Gzip EXI
EXI+ Gzip
Best + vocabulary-based
1-10 50.7 103.8 77 102 70.3 495.5 180.1 57.5 65.9 44.2
11-20 22.5 62 27.1 50.5 24.2 122.2 47 23.3 24.9 18.9
21-30 16.2 58.2 18.5 48.7 16.3 79.5 31.1 16.5 17.5 13.6
31-40 28.3 69.1 30.9 62.1 28.6 86.5 40.7 28.2 29.1 23.5
41-50 9.8 51.2 10.2 42.3 8.6 38.1 14.8 9.3 9.7 8
51-60 17.2 59.2 17.5 50.1 15.9 50.5 22.8 15.8 16.3 8.7
61-70 11.8 58.5 12.4 42.4 10 43 17.7 11.1 11.6 6
71-80 8.8 54.8 8.5 40.9 7 31.6 11.2 7.5 7.8 6.4
81-90 6.7 52 6.3 40.6 5.1 25.4 9.1 5.8 6 4.4
91-100 8.1 54.9 7.6 40.4 6.2 26.9 9.7 6.8 7 5.7
101-200 8.8 62 8.3 39.2 6.7 24.7 10.1 7.6 7.9 5.7
201-300 4.8 50.8 3.6 39 2.8 13.4 4 3.6 3.6 2.7
301-400 4.8 51.5 3.3 37.7 2.5 11.4 3.3 3 3.1 2.5
401-500 4.4 51.5 2.9 37.4 2.2 10.4 2.7 2.6 2.7 2.2
501-600 5 53.8 3.4 38.7 2.5 8.9 3 2.9 3 2.4
601-700 4.1 51 2.5 35.9 1.7 8.5 2.2 2.3 2.4 1.7
701-800 4.5 51.1 2.7 36.2 1.9 8.1 2.1 2.4 2.4 1.9
801-900 4.4 51.1 2.6 36.4 1.8 7.9 1.9 2.3 2.3 1.8
901-1000 4.1 50.9 2.4 36.5 1.7 7.7 1.7 2.1 2.1 1.7
801-900 4.4 51.1 2.6 36.4 1.8 7.9 1.9 2.3 2.3 1.8
901-1000 4.1 50.9 2.4 36.5 1.7 7.7 1.7 2.1 2.1 1.7
For very small datasets (<40 triples), n-triples + gzip works best For larger datasets Turtle+gzip compresses bestRemoving redundancies using shared vocabularies adds additional compression
Compression experiments resultsNumber of SMSes
Avg. number of triples
1 0
2 3
3 8
4 16
5 24
6 84
7 98
8 126
9 189
10 301
For very small datasets (<40 triples), n-triples + gzip works best For larger datasets Turtle+gzip compresses bestRemoving redundancies using shared vocabularies adds additional compression
0
10
20
30
40
50
60
70
80
Co
mp
ress
ion
(p
erc
en
tage
wrt
n-t
rip
les)
Size of dataset in triples (binned)
N-triples +gzip
Turtle+Gzip
Best + vocabulary-based
Evaluation: 4 scenarios in 2 cases
Digivet and RadioMarche applications
Four scenarios / SPARQL queries in total
Results
Linked Data over Sneakernet / Driveby wifi
Master project Fahad Ali
Semantic Web over SMS is feasible using data compression + semantic background knowledge
economically feasible for ICT4D services for small datasets
Semantic Web without the Web is possible cf. IOT
Radio Data System (RDS), LoRa,..
Dev. countries can leapfrog directly into the information age,
jumping many phases of immature technologies
Img: flickr/n3v3rv0id
From ICT4D to Development Informatics
Img: flickr/n3v3rv0id
KR4D, LD4D, HCI4D, …
Test Computer Science hypotheses in domains/environments
One more example
Aid accountability in Nepal
Marije VisscherSocial Scientist
Aske RobenhagenComputer Scientist
Citizens
Community frontline associates (CFAs)
Accountability Lab Office
Donors, Journalists
Surveys, focus groups. Collected in forms.
Reports from the field, solutions.
Aggregated reports, accountability
Example data
Citizens
Community frontline associates (CFAs)
Accountability Lab Office
Donors, Journalists
Smartphone - Voice System - SMS
Report1
हसकल्पनाम्यामआन्जिननयरहरुद्वारागाब
geo:Kavra
Report2
Knowledge graph