SlideShare a Scribd company logo
www.invid-project.eu
In Video Veritas – Verification of Social Media
Video Content for the News Industry
Lyndon Nixon, MODUL Technology GmbH
Video Retrieval for Multimedia Verification
of Breaking News on Social Networks
MuVer workshop, Mountain View CA, USA, October 2017
2
• Our goals are to:
• identify new news events occurring in social media
• rank best candidate social media video for the news event
• analyse selected video in preparation for verification
• Why?
• Social media like Twitter is quicker than the traditional news outlets in
breaking news stories
• The scale of social media is huge so we need to focus our social network
queries on retrieving content about the current news
• Journalists need quick and easy access to what are the current news stories
discussed on social media and what video is being posted around those
stories
Our role in the InVID project
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
3
• 1. How we detect stories from the social media stream
• 2. How we use story detection for information retrieval from
social networks
• 3. How guided video retrieval supports the news verification
process
Structure of this presentation
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
4
1. News story detection
The Louvain community
algorithm partitions the
keyword graph into sub-
graphs with optimized
modularity.
There are limitations in
smaller datasets which lead to
imperfect partitioning.
Work on improving the
disambiguation of stories:
- Merge based on overlap of
keywords in story clusters
with weighting
- Split through keyword co-
occurrence matrices
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
5
1. News story detection
Same stories are partitioned
into different clusters whenever
the connections between the
nodes are weak or there are
no connections at all when
those components are located
in different time range (e.g.
New report about a same
event across several days).
Those components need to be
merged.
Cluster labels are sufficient for
a first merge. We consider the
keywords as a bag of words
and merge iff 2 cluster labels
overlap by at least half.
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
6
1. News story detection
Within partitions, there may be keywords which associate together to
represent different stories, requiring an approach to split partitions.
First, we remove redundant keywords, i.e. repetitions or keywords which
are sub-sets of other keywords (e.g. “keys” together with “florida keys”)
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
7
1. News story detection
First experiments with
keywords of the clusters
where stories are merged
indicated that there is a
weaker co-occurrence
between the keywords
from different stories than
between the keywords
from the same story.
In the matrix, “finsbury park” is only associated with “attack” while “raqqa”
and “coalition” are only associated with “isis” (apart from one another) which
is also associated with “attack” and “cruise missile”.
Two stories are in this cluster: ATTACK + FINSBURY PARK, ATTACK +
ISIS + CRUISE MISSILE + RAQQA + COALITION
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
8
1. News story detection
We formalized the idea of
identifying keywords
which co-occur
significantly as triads.
These are pair-wise
associations between
keywords. Keywords not
part of a triad are leaf
nodes. Triads are
grouped when each triad
has at least one common
node with another triad in
the group.
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
9
1. News story detection
An example of a story
with the top triads
shown.
Note how here two triad
groups have formed:
POLLUTION + ACTION
+ CAMPAIGN
STUDY +
AUSTRALIAN +
AUSPOL + LESS
These match well the
two stories within this
cluster, as reflected in
the documents shown.
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
10
1. News story detection
After the triad groups are detected, they are split, keeping the leaf nodes
directly connected to the triads when they belong to a single group. Those
leaf nodes which are connected to triads in multiple groups are removed,
because it is hard to determine which group they might correctly belong to.
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
11
1. News story detection
After splitting, a content-based merge is performed to correct splits within a
single news story (if the keywords used for the story differ sufficiently within
the documents, the triad approach will likely split those partitions).
For each cluster the top n keywords (n is configurable) with the highest
frequency are chosen to represent the cluster. The remaining low frequency
keywords, which may contain noise, are not taken into consideration.
Overlapping rate is computed between each cluster pair. When one story
overlaps significantly with another story (threshold of overlapping rate is
configurable), these are merged.
Examples of merges:
“north korea,nuclear,president donald trump” is merged with “north
korea,donald trump,big deal”
“security council,north,nikki,haley” is merged with “north korea,security
council,sanctions,today,latest,moment,greatest,china,donald trump”
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
12
Evaluating the story detection
Story evaluation (a framework for evaluating the accuracy of the topic
detection algorithm)
• Ground truth evaluation. Comparison of stories occurring in
dashboard with stories on aggregation sites, e.g. Current News Portal
of Wikipedia (similar to SNOW2014 Challenge)
• Story correctness. Boolean test for whether a cluster
unambiguously identifies a newsworthy story
• Story distinctiveness. Boolean test for whether a cluster
uniquely refers to a distinct news story
• Story completeness. Measure of the extent to which
documents for a single news story are found within a single
cluster.
• Story homogeneity. Measure of the extent to which each
cluster contains only documents relevant to the same news
story.
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
13
Evaluating the story detection
Story evaluation set-up: Twitter Accounts in EN during 19-23 June, daily
stories from all tweets + 5 most popular topics
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
14
Evaluating the story detection
Correctness Distinctiveness Completeness Homogeneity
Aggregated 5-
day
0.90 0.60 0.93 0.93
Correctness could have been 100% if not for one Twitter account which
we have now removed.
Also the figures were similar across the individual days and the
individual topics.
We also tested the Twitter News source (which is UGC for “breaking
news”. While correctness was unsurprisingly lower it was still good
(0.76) and some stories were detected which did not appear in the
professional sources, reflecting “public interest” stories like ‘Two
elephants work together to save a baby elephant from drowning’.
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
15
2. News video retrieval from social media
How to surface the appropriate video content for each story?
- Using the story detection, we generate a set of new queries every 6
hours for the social network APIs (Twitter Video, YouTube, Vimeo,
DailyMotion)
- We create for the top max n results per query a document per response
with the video link and extracted metadata
- We also cluster the collected video documents around stories
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
16
Evaluating the news video retrieval
Findings suggest querying by topic and using stories labels
(instead of keyword associations) will significantly improve
the Video document collection (breadth of stories)
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
17
3. How this supports news video verification
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu
Public demo at https://invid.weblyzard.com
18
Thank you for your attention!
Find our dashboard for news story detection and video retrieval at
http://invid.weblyzard.com
Any questions?
Project consortium and funding agency
InVID is funded by the EU Horizon 2020 Programme under the project number ICT-2015-687786
Video Retrieval for Multimedia Verification of Breaking News on Social Networks
Lyndon Nixon nixon@modultech.eu

More Related Content

Viewers also liked

MediaEval 2016 - Verifying Multimedia Use Task Overview
MediaEval 2016 - Verifying Multimedia Use Task OverviewMediaEval 2016 - Verifying Multimedia Use Task Overview
MediaEval 2016 - Verifying Multimedia Use Task Overview
multimediaeval
 
MediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep Models
MediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep ModelsMediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep Models
MediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep Models
multimediaeval
 
MediaEval 2016 - BUT Zero-Cost Speech Recognition
MediaEval 2016 - BUT Zero-Cost Speech RecognitionMediaEval 2016 - BUT Zero-Cost Speech Recognition
MediaEval 2016 - BUT Zero-Cost Speech Recognition
multimediaeval
 
The InVID Plug-in: Web Video Verification on the Browser
The InVID Plug-in: Web Video Verification on the BrowserThe InVID Plug-in: Web Video Verification on the Browser
The InVID Plug-in: Web Video Verification on the Browser
InVID Project
 
MediaEval 2016: LAPI at Predicting Media Interestingness Task
MediaEval 2016: LAPI at Predicting Media Interestingness TaskMediaEval 2016: LAPI at Predicting Media Interestingness Task
MediaEval 2016: LAPI at Predicting Media Interestingness Task
multimediaeval
 
Media REVEALr: A social multimedia monitoring and intelligence system for Web...
Media REVEALr: A social multimedia monitoring and intelligence system for Web...Media REVEALr: A social multimedia monitoring and intelligence system for Web...
Media REVEALr: A social multimedia monitoring and intelligence system for Web...
Symeon Papadopoulos
 
MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...
MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...
MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...
multimediaeval
 
MediaEval 2016 - Simula Team @ Context of Experience Task
MediaEval 2016 - Simula Team @ Context of Experience TaskMediaEval 2016 - Simula Team @ Context of Experience Task
MediaEval 2016 - Simula Team @ Context of Experience Task
multimediaeval
 
MediaEval 2015 - JRS at Synchronization of Multi-user Event Media Task
MediaEval 2015 - JRS at Synchronization of Multi-user Event Media TaskMediaEval 2015 - JRS at Synchronization of Multi-user Event Media Task
MediaEval 2015 - JRS at Synchronization of Multi-user Event Media Task
multimediaeval
 
MediaEval 2016 - Emotion in Music Task: Lessons Learned
MediaEval 2016 - Emotion in Music Task: Lessons LearnedMediaEval 2016 - Emotion in Music Task: Lessons Learned
MediaEval 2016 - Emotion in Music Task: Lessons Learned
multimediaeval
 
MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...
MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...
MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...
multimediaeval
 
MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015
MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015
MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015
multimediaeval
 
MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...
MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...
MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...
multimediaeval
 
MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...
MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...
MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...
multimediaeval
 
MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...
MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...
MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...
multimediaeval
 
MediaEval 2016 - IR Evaluation: Putting the User Back in the Loop
MediaEval 2016 - IR Evaluation: Putting the User Back in the LoopMediaEval 2016 - IR Evaluation: Putting the User Back in the Loop
MediaEval 2016 - IR Evaluation: Putting the User Back in the Loop
multimediaeval
 
MediaEval 2016: A Multimodal System for the Verifying Multimedia Use Task
MediaEval 2016: A Multimodal System for the Verifying Multimedia Use TaskMediaEval 2016: A Multimodal System for the Verifying Multimedia Use Task
MediaEval 2016: A Multimodal System for the Verifying Multimedia Use Task
multimediaeval
 
MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015
MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015
MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015
multimediaeval
 
MediaEval 2016 - TUD-MMC Predicting media Interestingness Task
MediaEval 2016 - TUD-MMC Predicting media Interestingness TaskMediaEval 2016 - TUD-MMC Predicting media Interestingness Task
MediaEval 2016 - TUD-MMC Predicting media Interestingness Task
multimediaeval
 

Viewers also liked (19)

MediaEval 2016 - Verifying Multimedia Use Task Overview
MediaEval 2016 - Verifying Multimedia Use Task OverviewMediaEval 2016 - Verifying Multimedia Use Task Overview
MediaEval 2016 - Verifying Multimedia Use Task Overview
 
MediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep Models
MediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep ModelsMediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep Models
MediaEval 2016 - HUCVL Predicting Interesting Key Frames with Deep Models
 
MediaEval 2016 - BUT Zero-Cost Speech Recognition
MediaEval 2016 - BUT Zero-Cost Speech RecognitionMediaEval 2016 - BUT Zero-Cost Speech Recognition
MediaEval 2016 - BUT Zero-Cost Speech Recognition
 
The InVID Plug-in: Web Video Verification on the Browser
The InVID Plug-in: Web Video Verification on the BrowserThe InVID Plug-in: Web Video Verification on the Browser
The InVID Plug-in: Web Video Verification on the Browser
 
MediaEval 2016: LAPI at Predicting Media Interestingness Task
MediaEval 2016: LAPI at Predicting Media Interestingness TaskMediaEval 2016: LAPI at Predicting Media Interestingness Task
MediaEval 2016: LAPI at Predicting Media Interestingness Task
 
Media REVEALr: A social multimedia monitoring and intelligence system for Web...
Media REVEALr: A social multimedia monitoring and intelligence system for Web...Media REVEALr: A social multimedia monitoring and intelligence system for Web...
Media REVEALr: A social multimedia monitoring and intelligence system for Web...
 
MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...
MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...
MediaEval 2015 - Synchronization of Multi-User Event Media at MediaEval 2015:...
 
MediaEval 2016 - Simula Team @ Context of Experience Task
MediaEval 2016 - Simula Team @ Context of Experience TaskMediaEval 2016 - Simula Team @ Context of Experience Task
MediaEval 2016 - Simula Team @ Context of Experience Task
 
MediaEval 2015 - JRS at Synchronization of Multi-user Event Media Task
MediaEval 2015 - JRS at Synchronization of Multi-user Event Media TaskMediaEval 2015 - JRS at Synchronization of Multi-user Event Media Task
MediaEval 2015 - JRS at Synchronization of Multi-user Event Media Task
 
MediaEval 2016 - Emotion in Music Task: Lessons Learned
MediaEval 2016 - Emotion in Music Task: Lessons LearnedMediaEval 2016 - Emotion in Music Task: Lessons Learned
MediaEval 2016 - Emotion in Music Task: Lessons Learned
 
MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...
MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...
MediaEval 2016 - Placing Images with Refined Language Models and Similarity S...
 
MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015
MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015
MediaEval 2015 - GTM-UVigo Systems for Person Discovery Task at MediaEval 2015
 
MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...
MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...
MediaEval 2015 - CERTH at MediaEval 2015 Synchronization of Multi-User Event ...
 
MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...
MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...
MediaEval 2016 - COSMIR and the OpenMIC Challenge: A Plan for Sustainable Mus...
 
MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...
MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...
MediaEval 2016 - LAPI @ 2016 Retrieving Diverse Social Images Task: A Pseudo-...
 
MediaEval 2016 - IR Evaluation: Putting the User Back in the Loop
MediaEval 2016 - IR Evaluation: Putting the User Back in the LoopMediaEval 2016 - IR Evaluation: Putting the User Back in the Loop
MediaEval 2016 - IR Evaluation: Putting the User Back in the Loop
 
MediaEval 2016: A Multimodal System for the Verifying Multimedia Use Task
MediaEval 2016: A Multimodal System for the Verifying Multimedia Use TaskMediaEval 2016: A Multimodal System for the Verifying Multimedia Use Task
MediaEval 2016: A Multimodal System for the Verifying Multimedia Use Task
 
MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015
MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015
MediaEval 2015 - Verifying Multimedia Use at MediaEval 2015
 
MediaEval 2016 - TUD-MMC Predicting media Interestingness Task
MediaEval 2016 - TUD-MMC Predicting media Interestingness TaskMediaEval 2016 - TUD-MMC Predicting media Interestingness Task
MediaEval 2016 - TUD-MMC Predicting media Interestingness Task
 

Similar to Video Retrieval for Multimedia Verification of Breaking News on Social Networks

Classifying Crisis Information Relevancy with Semantics (ESWC 2018)
Classifying Crisis Information Relevancy with Semantics (ESWC 2018)Classifying Crisis Information Relevancy with Semantics (ESWC 2018)
Classifying Crisis Information Relevancy with Semantics (ESWC 2018)
Prashant Khare
 
Sensorpedia
SensorpediaSensorpedia
Sensorpedia
Franciel
 
Understanding User-Community Engagement by Multi-faceted Features: A Case ...
Understanding User-Community Engagement by Multi-faceted Features: A Case ...Understanding User-Community Engagement by Multi-faceted Features: A Case ...
Understanding User-Community Engagement by Multi-faceted Features: A Case ...
Artificial Intelligence Institute at UofSC
 
Social Media and the News: Approaches to the Spread of (Mis)information
Social Media and the News: Approaches to the Spread of (Mis)informationSocial Media and the News: Approaches to the Spread of (Mis)information
Social Media and the News: Approaches to the Spread of (Mis)information
Axel Bruns
 
CybersecurityTFReport2016 PRINT
CybersecurityTFReport2016 PRINTCybersecurityTFReport2016 PRINT
CybersecurityTFReport2016 PRINT
Aimee Shuck
 
Strategic scenarios in digital content and digital business
Strategic scenarios in digital content and digital businessStrategic scenarios in digital content and digital business
Strategic scenarios in digital content and digital business
Marco Brambilla
 
Distributed defense against disinformation: disinformation risk management an...
Distributed defense against disinformation: disinformation risk management an...Distributed defense against disinformation: disinformation risk management an...
Distributed defense against disinformation: disinformation risk management an...
Sara-Jayne Terp
 
2021-02-10_CogSecCollab_UBerkeley
2021-02-10_CogSecCollab_UBerkeley2021-02-10_CogSecCollab_UBerkeley
2021-02-10_CogSecCollab_UBerkeley
Sara-Jayne Terp
 
IRJET- Identification of Prevalent News from Twitter and Traditional Media us...
IRJET- Identification of Prevalent News from Twitter and Traditional Media us...IRJET- Identification of Prevalent News from Twitter and Traditional Media us...
IRJET- Identification of Prevalent News from Twitter and Traditional Media us...
IRJET Journal
 
EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...
EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...
EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...
Symeon Papadopoulos
 
Collaborative video WRITING WITH VIDEO SCMS workshop 2013
Collaborative video WRITING WITH VIDEO SCMS workshop 2013Collaborative video WRITING WITH VIDEO SCMS workshop 2013
Collaborative video WRITING WITH VIDEO SCMS workshop 2013
Vicki Callahan
 
Tags, Networks, Narrative
Tags, Networks, NarrativeTags, Networks, Narrative
Tags, Networks, Narrative
Bruce Mason
 
KNOW4DRR ws_polimi_bolzano_2013_introduction
KNOW4DRR  ws_polimi_bolzano_2013_introductionKNOW4DRR  ws_polimi_bolzano_2013_introduction
KNOW4DRR ws_polimi_bolzano_2013_introduction
know4drr
 
Mongoose H4D 2021 Lessons Learned
Mongoose H4D 2021 Lessons LearnedMongoose H4D 2021 Lessons Learned
Mongoose H4D 2021 Lessons Learned
Stanford University
 
Government Without Walls
Government Without WallsGovernment Without Walls
Government Without Walls
John Newton
 
Computational Verification Challenges in Social Media
Computational Verification Challenges in Social MediaComputational Verification Challenges in Social Media
Computational Verification Challenges in Social Media
Symeon Papadopoulos
 
OSINT Forum – Week 5For this forum, you are to answer one or mor.docx
OSINT Forum – Week 5For this forum, you are to answer one or mor.docxOSINT Forum – Week 5For this forum, you are to answer one or mor.docx
OSINT Forum – Week 5For this forum, you are to answer one or mor.docx
loganta
 
Reveal - Social Media Verification - poster
Reveal - Social Media Verification - posterReveal - Social Media Verification - poster
Reveal - Social Media Verification - poster
REVEAL - Social Media Verification
 
5 Musts of Social Media Events
5 Musts of Social Media Events5 Musts of Social Media Events
5 Musts of Social Media Events
Mallard Holliday
 
final_nlp
final_nlpfinal_nlp
final_nlp
aphex34
 

Similar to Video Retrieval for Multimedia Verification of Breaking News on Social Networks (20)

Classifying Crisis Information Relevancy with Semantics (ESWC 2018)
Classifying Crisis Information Relevancy with Semantics (ESWC 2018)Classifying Crisis Information Relevancy with Semantics (ESWC 2018)
Classifying Crisis Information Relevancy with Semantics (ESWC 2018)
 
Sensorpedia
SensorpediaSensorpedia
Sensorpedia
 
Understanding User-Community Engagement by Multi-faceted Features: A Case ...
Understanding User-Community Engagement by Multi-faceted Features: A Case ...Understanding User-Community Engagement by Multi-faceted Features: A Case ...
Understanding User-Community Engagement by Multi-faceted Features: A Case ...
 
Social Media and the News: Approaches to the Spread of (Mis)information
Social Media and the News: Approaches to the Spread of (Mis)informationSocial Media and the News: Approaches to the Spread of (Mis)information
Social Media and the News: Approaches to the Spread of (Mis)information
 
CybersecurityTFReport2016 PRINT
CybersecurityTFReport2016 PRINTCybersecurityTFReport2016 PRINT
CybersecurityTFReport2016 PRINT
 
Strategic scenarios in digital content and digital business
Strategic scenarios in digital content and digital businessStrategic scenarios in digital content and digital business
Strategic scenarios in digital content and digital business
 
Distributed defense against disinformation: disinformation risk management an...
Distributed defense against disinformation: disinformation risk management an...Distributed defense against disinformation: disinformation risk management an...
Distributed defense against disinformation: disinformation risk management an...
 
2021-02-10_CogSecCollab_UBerkeley
2021-02-10_CogSecCollab_UBerkeley2021-02-10_CogSecCollab_UBerkeley
2021-02-10_CogSecCollab_UBerkeley
 
IRJET- Identification of Prevalent News from Twitter and Traditional Media us...
IRJET- Identification of Prevalent News from Twitter and Traditional Media us...IRJET- Identification of Prevalent News from Twitter and Traditional Media us...
IRJET- Identification of Prevalent News from Twitter and Traditional Media us...
 
EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...
EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...
EventSense: Capturing the Pulse of Large-scale Events by Mining Social Media ...
 
Collaborative video WRITING WITH VIDEO SCMS workshop 2013
Collaborative video WRITING WITH VIDEO SCMS workshop 2013Collaborative video WRITING WITH VIDEO SCMS workshop 2013
Collaborative video WRITING WITH VIDEO SCMS workshop 2013
 
Tags, Networks, Narrative
Tags, Networks, NarrativeTags, Networks, Narrative
Tags, Networks, Narrative
 
KNOW4DRR ws_polimi_bolzano_2013_introduction
KNOW4DRR  ws_polimi_bolzano_2013_introductionKNOW4DRR  ws_polimi_bolzano_2013_introduction
KNOW4DRR ws_polimi_bolzano_2013_introduction
 
Mongoose H4D 2021 Lessons Learned
Mongoose H4D 2021 Lessons LearnedMongoose H4D 2021 Lessons Learned
Mongoose H4D 2021 Lessons Learned
 
Government Without Walls
Government Without WallsGovernment Without Walls
Government Without Walls
 
Computational Verification Challenges in Social Media
Computational Verification Challenges in Social MediaComputational Verification Challenges in Social Media
Computational Verification Challenges in Social Media
 
OSINT Forum – Week 5For this forum, you are to answer one or mor.docx
OSINT Forum – Week 5For this forum, you are to answer one or mor.docxOSINT Forum – Week 5For this forum, you are to answer one or mor.docx
OSINT Forum – Week 5For this forum, you are to answer one or mor.docx
 
Reveal - Social Media Verification - poster
Reveal - Social Media Verification - posterReveal - Social Media Verification - poster
Reveal - Social Media Verification - poster
 
5 Musts of Social Media Events
5 Musts of Social Media Events5 Musts of Social Media Events
5 Musts of Social Media Events
 
final_nlp
final_nlpfinal_nlp
final_nlp
 

More from InVID Project

InVID at IFCN Global Fact V
InVID at IFCN Global Fact VInVID at IFCN Global Fact V
InVID at IFCN Global Fact V
InVID Project
 
Presentation of the InVID technologies for image forensics analysis
Presentation of the InVID technologies for image forensics analysisPresentation of the InVID technologies for image forensics analysis
Presentation of the InVID technologies for image forensics analysis
InVID Project
 
Presentation of the InVID verification technologies at IPTC 2018
Presentation of the InVID verification technologies at IPTC 2018Presentation of the InVID verification technologies at IPTC 2018
Presentation of the InVID verification technologies at IPTC 2018
InVID Project
 
IUT-Toulouse-InVID-2018
IUT-Toulouse-InVID-2018IUT-Toulouse-InVID-2018
IUT-Toulouse-InVID-2018
InVID Project
 
A state of the art in journalism about fake image and video detection
A state of the art in journalism about fake image and video detectionA state of the art in journalism about fake image and video detection
A state of the art in journalism about fake image and video detection
InVID Project
 
Presentation of the InVID tool for video fragmentation and reverse keyframe s...
Presentation of the InVID tool for video fragmentation and reverse keyframe s...Presentation of the InVID tool for video fragmentation and reverse keyframe s...
Presentation of the InVID tool for video fragmentation and reverse keyframe s...
InVID Project
 
Presentation of the InVID Verification Plugin
Presentation of the InVID Verification PluginPresentation of the InVID Verification Plugin
Presentation of the InVID Verification Plugin
InVID Project
 
Presentation of the InVID tools for image forensics analysis
Presentation of the InVID tools for image forensics analysisPresentation of the InVID tools for image forensics analysis
Presentation of the InVID tools for image forensics analysis
InVID Project
 
The InVID Vefirication Plugin at #IFRAExpo #DCXExpo
The InVID Vefirication Plugin at #IFRAExpo #DCXExpoThe InVID Vefirication Plugin at #IFRAExpo #DCXExpo
The InVID Vefirication Plugin at #IFRAExpo #DCXExpo
InVID Project
 
The InVID Vefirication Plugin at DisinfoLab
The InVID Vefirication Plugin at DisinfoLabThe InVID Vefirication Plugin at DisinfoLab
The InVID Vefirication Plugin at DisinfoLab
InVID Project
 
Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...
Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...
Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...
InVID Project
 
TRECVID 2016 Ad-hoc Video Seach task, CERTH-ITI
TRECVID 2016 Ad-hoc Video Seach task, CERTH-ITITRECVID 2016 Ad-hoc Video Seach task, CERTH-ITI
TRECVID 2016 Ad-hoc Video Seach task, CERTH-ITI
InVID Project
 
ELLA LC algorithm presentation in ICIP 2016
ELLA LC algorithm presentation in ICIP 2016ELLA LC algorithm presentation in ICIP 2016
ELLA LC algorithm presentation in ICIP 2016
InVID Project
 
InVID verification application presentation at SMVW16
InVID verification application presentation at SMVW16InVID verification application presentation at SMVW16
InVID verification application presentation at SMVW16
InVID Project
 
InVID platform presentation at SMVW16
InVID platform presentation at SMVW16InVID platform presentation at SMVW16
InVID platform presentation at SMVW16
InVID Project
 
InVID project presentation at SMVW16
InVID project presentation at SMVW16InVID project presentation at SMVW16
InVID project presentation at SMVW16
InVID Project
 
InVID at Computational journalism workshop in Rennes, France
InVID at Computational journalism workshop in Rennes, FranceInVID at Computational journalism workshop in Rennes, France
InVID at Computational journalism workshop in Rennes, France
InVID Project
 
InVID Project Presentation 3rd release March 2016
InVID Project Presentation 3rd release March 2016InVID Project Presentation 3rd release March 2016
InVID Project Presentation 3rd release March 2016
InVID Project
 

More from InVID Project (18)

InVID at IFCN Global Fact V
InVID at IFCN Global Fact VInVID at IFCN Global Fact V
InVID at IFCN Global Fact V
 
Presentation of the InVID technologies for image forensics analysis
Presentation of the InVID technologies for image forensics analysisPresentation of the InVID technologies for image forensics analysis
Presentation of the InVID technologies for image forensics analysis
 
Presentation of the InVID verification technologies at IPTC 2018
Presentation of the InVID verification technologies at IPTC 2018Presentation of the InVID verification technologies at IPTC 2018
Presentation of the InVID verification technologies at IPTC 2018
 
IUT-Toulouse-InVID-2018
IUT-Toulouse-InVID-2018IUT-Toulouse-InVID-2018
IUT-Toulouse-InVID-2018
 
A state of the art in journalism about fake image and video detection
A state of the art in journalism about fake image and video detectionA state of the art in journalism about fake image and video detection
A state of the art in journalism about fake image and video detection
 
Presentation of the InVID tool for video fragmentation and reverse keyframe s...
Presentation of the InVID tool for video fragmentation and reverse keyframe s...Presentation of the InVID tool for video fragmentation and reverse keyframe s...
Presentation of the InVID tool for video fragmentation and reverse keyframe s...
 
Presentation of the InVID Verification Plugin
Presentation of the InVID Verification PluginPresentation of the InVID Verification Plugin
Presentation of the InVID Verification Plugin
 
Presentation of the InVID tools for image forensics analysis
Presentation of the InVID tools for image forensics analysisPresentation of the InVID tools for image forensics analysis
Presentation of the InVID tools for image forensics analysis
 
The InVID Vefirication Plugin at #IFRAExpo #DCXExpo
The InVID Vefirication Plugin at #IFRAExpo #DCXExpoThe InVID Vefirication Plugin at #IFRAExpo #DCXExpo
The InVID Vefirication Plugin at #IFRAExpo #DCXExpo
 
The InVID Vefirication Plugin at DisinfoLab
The InVID Vefirication Plugin at DisinfoLabThe InVID Vefirication Plugin at DisinfoLab
The InVID Vefirication Plugin at DisinfoLab
 
Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...
Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...
Comparison of Fine-tuning and Extension Strategies for Deep Convolutional Neu...
 
TRECVID 2016 Ad-hoc Video Seach task, CERTH-ITI
TRECVID 2016 Ad-hoc Video Seach task, CERTH-ITITRECVID 2016 Ad-hoc Video Seach task, CERTH-ITI
TRECVID 2016 Ad-hoc Video Seach task, CERTH-ITI
 
ELLA LC algorithm presentation in ICIP 2016
ELLA LC algorithm presentation in ICIP 2016ELLA LC algorithm presentation in ICIP 2016
ELLA LC algorithm presentation in ICIP 2016
 
InVID verification application presentation at SMVW16
InVID verification application presentation at SMVW16InVID verification application presentation at SMVW16
InVID verification application presentation at SMVW16
 
InVID platform presentation at SMVW16
InVID platform presentation at SMVW16InVID platform presentation at SMVW16
InVID platform presentation at SMVW16
 
InVID project presentation at SMVW16
InVID project presentation at SMVW16InVID project presentation at SMVW16
InVID project presentation at SMVW16
 
InVID at Computational journalism workshop in Rennes, France
InVID at Computational journalism workshop in Rennes, FranceInVID at Computational journalism workshop in Rennes, France
InVID at Computational journalism workshop in Rennes, France
 
InVID Project Presentation 3rd release March 2016
InVID Project Presentation 3rd release March 2016InVID Project Presentation 3rd release March 2016
InVID Project Presentation 3rd release March 2016
 

Recently uploaded

Portugal Dreamin 24 - How to easily use an API with Flows
Portugal Dreamin 24  - How to easily use an API with FlowsPortugal Dreamin 24  - How to easily use an API with Flows
Portugal Dreamin 24 - How to easily use an API with Flows
Thierry TROUIN ☁
 
Girls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in City
Girls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in CityGirls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in City
Girls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in City
dilbaagsingh0898
 
High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...
High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...
High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...
shamrisumri
 
Rent remote desktop server mangohost .net
Rent remote desktop server mangohost .netRent remote desktop server mangohost .net
Rent remote desktop server mangohost .net
pdfsubmission50
 
Use of Ontologies in Chemical Kinetic Database CHEMCONNECT
Use of Ontologies in Chemical Kinetic Database CHEMCONNECTUse of Ontologies in Chemical Kinetic Database CHEMCONNECT
Use of Ontologies in Chemical Kinetic Database CHEMCONNECT
Edward Blurock
 
202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧
202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧
202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧
ffg01100
 
Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...
Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...
Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...
paridubey2024#G05
 
Web development Platform Constraints.pptx
Web development Platform Constraints.pptxWeb development Platform Constraints.pptx
Web development Platform Constraints.pptx
ssuser2f6682
 
Chennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai Available
Chennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai AvailableChennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai Available
Chennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai Available
shamrisumri
 
Trading Strategy for London silver bullet
Trading Strategy for London silver bulletTrading Strategy for London silver bullet
Trading Strategy for London silver bullet
OkgatoSemadi1
 
Information Systems Auditing, Controls and Assurance , tanapat limsaiprom
Information Systems Auditing, Controls and Assurance , tanapat limsaipromInformation Systems Auditing, Controls and Assurance , tanapat limsaiprom
Information Systems Auditing, Controls and Assurance , tanapat limsaiprom
TanapatLimsaiprom1
 
Team Cymru Community Services,Overview of all public services
Team Cymru Community Services,Overview of all public servicesTeam Cymru Community Services,Overview of all public services
Team Cymru Community Services,Overview of all public services
Bangladesh Network Operators Group
 
IPv6 Deployment Planning and Security Considerations
IPv6 Deployment Planning and Security ConsiderationsIPv6 Deployment Planning and Security Considerations
IPv6 Deployment Planning and Security Considerations
Bangladesh Network Operators Group
 
Understanding Threat Intelligence | What is Threat Intelligence
Understanding Threat Intelligence | What is Threat IntelligenceUnderstanding Threat Intelligence | What is Threat Intelligence
Understanding Threat Intelligence | What is Threat Intelligence
Lumiverse Solutions Pvt Ltd
 
Career Development Advice for Network Engineers across the Pacific, presented...
Career Development Advice for Network Engineers across the Pacific, presented...Career Development Advice for Network Engineers across the Pacific, presented...
Career Development Advice for Network Engineers across the Pacific, presented...
APNIC
 
Build a Professional Resume using Canva , Tanapat Limsaiprom
Build a Professional Resume using Canva , Tanapat LimsaipromBuild a Professional Resume using Canva , Tanapat Limsaiprom
Build a Professional Resume using Canva , Tanapat Limsaiprom
TanapatLimsaiprom1
 
Network Security version1.0 - Module 3.pptx
Network Security version1.0 - Module 3.pptxNetwork Security version1.0 - Module 3.pptx
Network Security version1.0 - Module 3.pptx
Infotainmentforall
 
Maximizing Network Efficiency with Large Language Models (LLM)
Maximizing Network Efficiency with Large Language Models (LLM)Maximizing Network Efficiency with Large Language Models (LLM)
Maximizing Network Efficiency with Large Language Models (LLM)
Bangladesh Network Operators Group
 
optimized green synthesis characterization and evaluation
optimized green synthesis characterization and evaluationoptimized green synthesis characterization and evaluation
optimized green synthesis characterization and evaluation
ManojKumarr75
 
@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...
@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...
@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...
shamrisumri
 

Recently uploaded (20)

Portugal Dreamin 24 - How to easily use an API with Flows
Portugal Dreamin 24  - How to easily use an API with FlowsPortugal Dreamin 24  - How to easily use an API with Flows
Portugal Dreamin 24 - How to easily use an API with Flows
 
Girls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in City
Girls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in CityGirls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in City
Girls Call Shimla 000XX00000 Provide Best And Top Girl Service And No1 in City
 
High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...
High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...
High Profile Girls Call ServiCe Chennai XX00XXX00X Tanisha Best High Class Ch...
 
Rent remote desktop server mangohost .net
Rent remote desktop server mangohost .netRent remote desktop server mangohost .net
Rent remote desktop server mangohost .net
 
Use of Ontologies in Chemical Kinetic Database CHEMCONNECT
Use of Ontologies in Chemical Kinetic Database CHEMCONNECTUse of Ontologies in Chemical Kinetic Database CHEMCONNECT
Use of Ontologies in Chemical Kinetic Database CHEMCONNECT
 
202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧
202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧
202254.com免费观看《长相思第二季》免费观看高清,长相思第二季线上看,《长相思第二季》最新电视剧在线观看,杨紫最新电视剧
 
Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...
Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...
Kolkata @Girls @Call WhatsApp Numbers 🫦0000XX0000🫦 List For Friendship Girls ...
 
Web development Platform Constraints.pptx
Web development Platform Constraints.pptxWeb development Platform Constraints.pptx
Web development Platform Constraints.pptx
 
Chennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai Available
Chennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai AvailableChennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai Available
Chennai Girls Call ServiCe X00XXX00XX Tanisha Best High Class Chennai Available
 
Trading Strategy for London silver bullet
Trading Strategy for London silver bulletTrading Strategy for London silver bullet
Trading Strategy for London silver bullet
 
Information Systems Auditing, Controls and Assurance , tanapat limsaiprom
Information Systems Auditing, Controls and Assurance , tanapat limsaipromInformation Systems Auditing, Controls and Assurance , tanapat limsaiprom
Information Systems Auditing, Controls and Assurance , tanapat limsaiprom
 
Team Cymru Community Services,Overview of all public services
Team Cymru Community Services,Overview of all public servicesTeam Cymru Community Services,Overview of all public services
Team Cymru Community Services,Overview of all public services
 
IPv6 Deployment Planning and Security Considerations
IPv6 Deployment Planning and Security ConsiderationsIPv6 Deployment Planning and Security Considerations
IPv6 Deployment Planning and Security Considerations
 
Understanding Threat Intelligence | What is Threat Intelligence
Understanding Threat Intelligence | What is Threat IntelligenceUnderstanding Threat Intelligence | What is Threat Intelligence
Understanding Threat Intelligence | What is Threat Intelligence
 
Career Development Advice for Network Engineers across the Pacific, presented...
Career Development Advice for Network Engineers across the Pacific, presented...Career Development Advice for Network Engineers across the Pacific, presented...
Career Development Advice for Network Engineers across the Pacific, presented...
 
Build a Professional Resume using Canva , Tanapat Limsaiprom
Build a Professional Resume using Canva , Tanapat LimsaipromBuild a Professional Resume using Canva , Tanapat Limsaiprom
Build a Professional Resume using Canva , Tanapat Limsaiprom
 
Network Security version1.0 - Module 3.pptx
Network Security version1.0 - Module 3.pptxNetwork Security version1.0 - Module 3.pptx
Network Security version1.0 - Module 3.pptx
 
Maximizing Network Efficiency with Large Language Models (LLM)
Maximizing Network Efficiency with Large Language Models (LLM)Maximizing Network Efficiency with Large Language Models (LLM)
Maximizing Network Efficiency with Large Language Models (LLM)
 
optimized green synthesis characterization and evaluation
optimized green synthesis characterization and evaluationoptimized green synthesis characterization and evaluation
optimized green synthesis characterization and evaluation
 
@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...
@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...
@Girls @Call Chennai 🛬 XXXXXXXXXX 🛬 available 24*7 cash payment book now pay ...
 

Video Retrieval for Multimedia Verification of Breaking News on Social Networks

  • 1. www.invid-project.eu In Video Veritas – Verification of Social Media Video Content for the News Industry Lyndon Nixon, MODUL Technology GmbH Video Retrieval for Multimedia Verification of Breaking News on Social Networks MuVer workshop, Mountain View CA, USA, October 2017
  • 2. 2 • Our goals are to: • identify new news events occurring in social media • rank best candidate social media video for the news event • analyse selected video in preparation for verification • Why? • Social media like Twitter is quicker than the traditional news outlets in breaking news stories • The scale of social media is huge so we need to focus our social network queries on retrieving content about the current news • Journalists need quick and easy access to what are the current news stories discussed on social media and what video is being posted around those stories Our role in the InVID project Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 3. 3 • 1. How we detect stories from the social media stream • 2. How we use story detection for information retrieval from social networks • 3. How guided video retrieval supports the news verification process Structure of this presentation Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 4. 4 1. News story detection The Louvain community algorithm partitions the keyword graph into sub- graphs with optimized modularity. There are limitations in smaller datasets which lead to imperfect partitioning. Work on improving the disambiguation of stories: - Merge based on overlap of keywords in story clusters with weighting - Split through keyword co- occurrence matrices Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 5. 5 1. News story detection Same stories are partitioned into different clusters whenever the connections between the nodes are weak or there are no connections at all when those components are located in different time range (e.g. New report about a same event across several days). Those components need to be merged. Cluster labels are sufficient for a first merge. We consider the keywords as a bag of words and merge iff 2 cluster labels overlap by at least half. Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 6. 6 1. News story detection Within partitions, there may be keywords which associate together to represent different stories, requiring an approach to split partitions. First, we remove redundant keywords, i.e. repetitions or keywords which are sub-sets of other keywords (e.g. “keys” together with “florida keys”) Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 7. 7 1. News story detection First experiments with keywords of the clusters where stories are merged indicated that there is a weaker co-occurrence between the keywords from different stories than between the keywords from the same story. In the matrix, “finsbury park” is only associated with “attack” while “raqqa” and “coalition” are only associated with “isis” (apart from one another) which is also associated with “attack” and “cruise missile”. Two stories are in this cluster: ATTACK + FINSBURY PARK, ATTACK + ISIS + CRUISE MISSILE + RAQQA + COALITION Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 8. 8 1. News story detection We formalized the idea of identifying keywords which co-occur significantly as triads. These are pair-wise associations between keywords. Keywords not part of a triad are leaf nodes. Triads are grouped when each triad has at least one common node with another triad in the group. Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 9. 9 1. News story detection An example of a story with the top triads shown. Note how here two triad groups have formed: POLLUTION + ACTION + CAMPAIGN STUDY + AUSTRALIAN + AUSPOL + LESS These match well the two stories within this cluster, as reflected in the documents shown. Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 10. 10 1. News story detection After the triad groups are detected, they are split, keeping the leaf nodes directly connected to the triads when they belong to a single group. Those leaf nodes which are connected to triads in multiple groups are removed, because it is hard to determine which group they might correctly belong to. Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 11. 11 1. News story detection After splitting, a content-based merge is performed to correct splits within a single news story (if the keywords used for the story differ sufficiently within the documents, the triad approach will likely split those partitions). For each cluster the top n keywords (n is configurable) with the highest frequency are chosen to represent the cluster. The remaining low frequency keywords, which may contain noise, are not taken into consideration. Overlapping rate is computed between each cluster pair. When one story overlaps significantly with another story (threshold of overlapping rate is configurable), these are merged. Examples of merges: “north korea,nuclear,president donald trump” is merged with “north korea,donald trump,big deal” “security council,north,nikki,haley” is merged with “north korea,security council,sanctions,today,latest,moment,greatest,china,donald trump” Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 12. 12 Evaluating the story detection Story evaluation (a framework for evaluating the accuracy of the topic detection algorithm) • Ground truth evaluation. Comparison of stories occurring in dashboard with stories on aggregation sites, e.g. Current News Portal of Wikipedia (similar to SNOW2014 Challenge) • Story correctness. Boolean test for whether a cluster unambiguously identifies a newsworthy story • Story distinctiveness. Boolean test for whether a cluster uniquely refers to a distinct news story • Story completeness. Measure of the extent to which documents for a single news story are found within a single cluster. • Story homogeneity. Measure of the extent to which each cluster contains only documents relevant to the same news story. Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 13. 13 Evaluating the story detection Story evaluation set-up: Twitter Accounts in EN during 19-23 June, daily stories from all tweets + 5 most popular topics Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 14. 14 Evaluating the story detection Correctness Distinctiveness Completeness Homogeneity Aggregated 5- day 0.90 0.60 0.93 0.93 Correctness could have been 100% if not for one Twitter account which we have now removed. Also the figures were similar across the individual days and the individual topics. We also tested the Twitter News source (which is UGC for “breaking news”. While correctness was unsurprisingly lower it was still good (0.76) and some stories were detected which did not appear in the professional sources, reflecting “public interest” stories like ‘Two elephants work together to save a baby elephant from drowning’. Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 15. 15 2. News video retrieval from social media How to surface the appropriate video content for each story? - Using the story detection, we generate a set of new queries every 6 hours for the social network APIs (Twitter Video, YouTube, Vimeo, DailyMotion) - We create for the top max n results per query a document per response with the video link and extracted metadata - We also cluster the collected video documents around stories Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 16. 16 Evaluating the news video retrieval Findings suggest querying by topic and using stories labels (instead of keyword associations) will significantly improve the Video document collection (breadth of stories) Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu
  • 17. 17 3. How this supports news video verification Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu Public demo at https://invid.weblyzard.com
  • 18. 18 Thank you for your attention! Find our dashboard for news story detection and video retrieval at http://invid.weblyzard.com Any questions? Project consortium and funding agency InVID is funded by the EU Horizon 2020 Programme under the project number ICT-2015-687786 Video Retrieval for Multimedia Verification of Breaking News on Social Networks Lyndon Nixon nixon@modultech.eu

Editor's Notes

  1. Data sampled from the time range 06:00 - 09:00, 13 sep 2017. Data source invid videos.
  2. Data sampled from the time range 06:00 - 09:00, 13 sep 2017. Data source invid videos.
  3. Data sampled from the time range 06:00 - 09:00, 13 sep 2017. Data source invid videos.
  4. Data sampled from the time range 06:00 - 09:00, 13 sep 2017. Data source invid videos.
  5. Triads are marked as red links.
  6. Triads are marked as red links.
  7. Triads are marked as red links.
  8. Triads are marked as red links.
  9. Note to self: first introduce the dashboard now to „demonstrate“ the story detection and video collection we have (theoretically) discussed?