Here's the third in a series of webinars, hosted by the European Broadcasting Union (EBU). In this session, members of the University of Sheffield present vera.ai work and outcomes, namely work and results in synthetic text detection, credibility signals and Object Character Recognition (OCR). The recording is just a few clicks away. 

University of Sheffield presents work on synthetic text detection, credibility signals and OCR in EBU webinar series no. 3

Bildschirmfoto 2024-12-20 um 13.38.09.png

Here's the third in a series of webinars, hosted by the European Broadcasting Union (EBU). In this session, members of the University of Sheffield present vera.ai work and outcomes, namely work and results in synthetic text detection, credibility signals and Object Character Recognition (OCR). The recording is just a few clicks away.&nbsp;

It's that time of the year again.... In  fact: the end of the year is near. This means: we will sign off until early 2025. We very much hope you can do the same, namely relax, take it easy, lean back (or lie down for some time) and get up again, full of energy, in due time. Looking forward to starting again in 2025 and bringing our topics and endeavors further. Take care and all the very best!

Happy holiday period - relax, enjoy and have a good time!

IMG_2344.jpeg

It's that time of the year again.... In &nbsp;fact: the end of the year is near. This means: we will sign off until early 2025. We very much hope you can do the same, namely relax, take it easy, lean back (or lie down for some time) and get up again, full of energy, in due time. Looking forward to starting again in 2025 and bringing our topics and endeavors further. Take care and all the very best!

We invite you to become a beta tester for the Coordinated Sharing Detection Service, powered by CooRTweet. Developed within the Vera AI Horizon EU project, this cutting-edge tool is designed to detect and analyze coordinated sharing behaviors on social media platforms, playing a vital role in the fight against online disinformation.

Join Us as a Beta Tester for the Coordinated Sharing Detection Service!

coord_b_de2019b.png

We invite you to become a beta tester for the Coordinated Sharing Detection Service, powered by CooRTweet. Developed within the Vera AI Horizon EU project, this cutting-edge tool is designed to detect and analyze coordinated sharing behaviors on social media platforms, playing a vital role in the fight against online disinformation.

vera.ai's project partners CERTH, USFD and Ontotext presented advances in synthetic content detection, textual analysis and narrative exploration at the Sofia Information Integrity Forum (SIIF), held from 7–9 November 2024 in Sofia, Bulgaria. Here's a summary of their contributions, plus insights into work done, including results to date as well as approaches and challenges.

Work of vera.ai that contributes to the domain of information integrity

_VN_3463.jpg

vera.ai's project partners CERTH, USFD and Ontotext presented advances in synthetic content detection, textual analysis and narrative exploration at the Sofia Information Integrity Forum (SIIF), held from 7–9 November 2024 in Sofia, Bulgaria. Here's a summary of their contributions, plus insights into work done, including results to date as well as approaches and challenges.

Following vera.ai tools being featured in a report of the European External Action Services (EEAS), vera.ai project partner EU DisinfoLab is inviting anyone interested to a webinar on 12 December from 14:30-15:30 CET with Maria Giovanna Sessa and Raquel Miguel Serrano in which some of the report findings will be presented. More about the event, including a link for (compulsory) registration below.

 EUDL webinar alert: join us on 12 Dec 2024 for a talk on verification tools and guidelines

Bildschirmfoto 2024-11-29 um 13.10.02.png

Following vera.ai tools being featured in a report of the European External Action Services (EEAS), vera.ai project partner EU DisinfoLab is inviting anyone interested to a webinar on 12 December from 14:30-15:30 CET with Maria Giovanna Sessa and Raquel Miguel Serrano in which some of the report findings will be presented. More about the event, including a link for (compulsory) registration below.

On Thursday 14 November 2024 we launched our presence on BlueSky, a federated social network that in many ways resembles old Twitter (not so much X as it is called now, and after it had been purchased by Elon Musk and was turned into a rather toxic platform). We’ll be posting more and more on BlueSky and see how things develop there - also keeping a very close eye on what happens on X.

vera.ai has joined BlueSky: hope to see you there! 

Bluesky-logo xx.webp

On Thursday 14 November 2024 we launched our presence on BlueSky, a federated social network that in many ways resembles old Twitter (not so much X as it is called now, and after it had been purchased by Elon Musk and was turned into a rather toxic platform). We’ll be posting more and more on BlueSky and see how things develop there - also keeping a very close eye on what happens on X.

This is the second edition in our webinar series organised by the European Broadcasting Union (EBU). It features Luca Cuccovillo and Milica Gerhardt of Fraunhofer IDMT who talk about and showcase their vera.ai work which focusses on tools, techniques and their research work related to the analysis of audio content - another activity that is to support in countering disinformation.

Fraunhofer IDMT talk on audio analysis in EBU webinar series no. 2

Bildschirmfoto 2024-11-11 um 18.17.45.png

This is the second edition in our webinar series organised by the European Broadcasting Union (EBU). It features Luca Cuccovillo and Milica Gerhardt of Fraunhofer IDMT who talk about and showcase their vera.ai work which focusses on tools, techniques and their research work related to the analysis of audio content - another activity that is to support in countering disinformation.

On 29 October 2024 the University of Sheffield hosted a one-day event that brought together some of the world's leading experts in the field of multimodal and cross-platform coordinated behavior detection. They showcased their work and discussed how to advance the state of the art  in the. domain. Here's a detailed summary of the event.

Coordinated Sharing Behavior Detection (Conference): Report and Insights

CIB Conf 1.jpg

On 29 October 2024 the University of Sheffield hosted a one-day satellite event that brought together some of the world's leading experts in the field of multimodal and cross-platform coordinated behavior detection. They showcased their work and discussed how to advance the state of the art &nbsp;in the. domain. Here's a detailed summary of the event.

Here's a webinar recording of a session that took place on 11 September 2024, organised by the EBU. It features Symeon Papadopoulos (CERTH) and Luisa Verdoliva (University Federico II of Naples), who explain the advances, tools and challenges of synthetic image detection – a technology that tackles one of the biggest challenges in fact-checking, namely the proliferation of AI-generated or AI-modified images. Follow on to view the recording.

EBU webinar featuring CERTH & UNINA colleagues outlining vera.ai work on synthetic image detection

Bildschirmfoto 2024-10-16 um 13.58.41.png

Here's a webinar&nbsp;recording of a session that took place on 11 September 2024, organised by the EBU. It features Symeon Papadopoulos&nbsp;(CERTH) and Luisa Verdoliva&nbsp;(University Federico II of Naples), who explain the advances, tools and challenges of synthetic image detection&nbsp;– a technology that tackles one of the biggest challenges in fact-checking, namely the proliferation of AI-generated or AI-modified images. Follow on to view the recording.

A lot in vera.ai centers around data and code. In other words: the dealing with data, its analysis, experimentation, and more. In October 2024 we have started to bring together and make available all code that a) was created in and b) is used by vera.ai members in one central destination on the project website. This includes direct links to respective repositories - in most cases GitHub.  

Code made available - you can get your hands on it

Bildschirmfoto 2024-10-14 um 13.32.52.png

A lot in vera.ai centers around data and code. In other words: the dealing with data, its analysis, experimentation, and more. In October 2024 we have started to bring together and make available all code that a) was created in and b) is used by vera.ai members in one central destination on the project website. This includes direct links to respective repositories - in most cases GitHub. &nbsp;

On 14 September 2024, vera.ai completed its second project year. Reason enough for Ontotext's Nedelina Mitankina to look back on a year of activities, as well as look ahead at what is to come. In her contribution, Nedelina portrays in particular the work carried out in relation to the so-called Database of Known Fakes that can support factcheckers and verification experts alike. In focus in particular: entity linking and event extraction.

A year in review: Ontotext's Nedelina Mitankina looks back - and ahead 

logo_ontotext.jpg

On 14 September 2024, vera.ai completed its second project year. Reason enough for Ontotext's Nedelina Mitankina to look back on a year of activities, as well as look ahead at what is to come. In her contribution, Nedelina portrays in particular the work carried out in relation to the so-called Database of Known Fakes that can support factcheckers and verification experts alike. In focus in particular: entity linking and event extraction.

The MeVer team of CERTH-ITI was at the Thessaloniki International Fair. Team members presented and showcased their work on synthetic media detection that is being developed in the vera.ai project. 

MeVer @ Thessaloniki International Fair - presenting work on synthetic media detection

Bildschirmfoto 2024-09-19 um 20.24.42.png

The MeVer team of CERTH-ITI was at the Thessaloniki International Fair. Team members presented and showcased their work on synthetic media detection that is being developed in the vera.ai project.

The KInIT research team thoroughly investigated capabilities of large language models (LLMs) to automatically generate disinformation. In the following summary, they describe novel findings that their research revealed.

Exploring Disinformation Capabilities of Large Language Models

llm-capabilities-for-disinfo-headerII.png

The <a href="https://kinit.sk/publication/disinformation-capabilities-of-large-language-models/">KInIT research team</a> thoroughly investigated capabilities of large language models (LLMs) to automatically generate disinformation. In the following summary, they describe novel findings that their research revealed.

Meet vera.ai at the upcoming International Broadcasting Convention (IBC), 13-16 September 2024, Amsterdam. You can find us in Hall 14 where the all-new IBC AI Tech Zone is introduced at this year’s event. The AI Tech Zone will be bustling with industry leaders and visionaries eager to explore how AI is transforming media and entertainment. Among them, vera.ai will showcase the cutting edge tools developed to enhance verification and combat disinformation.

vera.ai at the International Broadcasting Convention 2024 - meet us in Amsterdam!

veraaiatibc24.png

Meet vera.ai at the upcoming International Broadcasting Convention (IBC), 13-16 September 2024, in Amsterdam. You can find us in Hall 14 where the all-new IBC AI Tech Zone is introduced at this year’s event. The AI Tech Zone will be bustling with industry leaders and visionaries eager&nbsp;to explore how AI is transforming media and entertainment. Among them, vera.ai will showcase the cutting edge tools developed to enhance verification and combat disinformation.

The EU DisinfoLab has authored a report under the EU funded project veraA,I entitled ‘Revisit the Coordinated Inauthentic Behaviour detection tree' (as part of the project deliverable D4.2). Read it here. 

Report on “Coordinated inauthentic behaviour detection tree”

cib-detection-tree-report.jpg

The present report revisits the Coordinated Inauthentic Behaviour (CIB) detection tree developed by EU DisinfoLab in 2021. It streamlines four branches of CIB into one comprehensive document and thoroughly examines their current status alongside the CIB definition and AI technologies' impact.

This article outlines how automated fact-checking (AFC) is being developed and progresses in order to support with verification tasks that, manually, take up a lot of time. Furthermore, manual fact-checking usually can not keep up with the quick spread of false information. That is why researchers and journalists / fact-checkers need to collaborate.

Assisting automated fact-checking via external evidences

Featured image.png

This article outlines how automated fact-checking (AFC) is being developed and progresses in order to support with verification tasks that, manually, take up a lot of time. Furthermore, manual fact-checking usually can not keep up with the quick spread of false information. That is why researchers and journalists / fact-checkers need to collaborate.

What are recent trends, technological developments and countermeasures when it comes to Generative AI and disinformation? What can be done about it? Who's leading the race? Read all about it in this extensive report put together jointly by consortium partners of vera.ai.

Generative AI and disinformation: taking a more detailed look. (Report)

disinfo-report_title.png

What are recent trends, technological developments and countermeasures when it comes to Generative AI and disinformation? What can be done about it? Who's leading the race? Read all about it in this extensive report put together jointly by consortium partners of vera.ai.

For the second consecutive year, vera.ai was invited to present some of what we got up to at the world's largest fact-checking summit, Global Fact, now in its 11th edition - hence Global Fact 11. It took place from 26-28 June 2024, in Sarajevo, Bosnia and Herzegovina. Our Denis Teyssou had the opportunity to present latest improvements incorporated in the verification plugin and share best practices to verify synthetic content and deepfakes. 

vera.ai at IFCN's Global Fact 11 in Sarajevo 

OnStage2.jpg

For the second consecutive year, vera.ai was invited to present some of what we got up to at the world's largest fact-checking summit, Global Fact, now in its 11th edition - hence Global Fact 11. It took place from 26-28 June 2024, in Sarajevo, Bosnia and Herzegovina. Our Denis Teyssou had the opportunity to present latest improvements incorporated in the verification plugin and share best practices to verify synthetic content and deepfakes.&nbsp;

"Meet the Future of AI – Generative AI and Democracy" was an event hosted by six European-funded projects, together with the European Commission. It took place on 19 June 2024 in Brussels. Symeon Papadopoulos, veraAI's project coordinator, now points out which conclusions can be drawn from a full day of exchange and expertise!

Generative AI and Democracy: Drawing Conclusions 

Meet the future panel 1 (5).jpg

The previous article gave insights and summarized the overall event "Meet the Future of AI – Generative AI and Democracy", which was hosted by six European-funded projects, together with the European Commission. It took place on 19 June 2024 in Brussels.&nbsp; In this follow-up, veraAI's project coordinator, Symeon Papadopoulos points out which conclusions can be drawn from a full day of exchange and expertise.

"As a newbie, I was unprepared for the sheer scale of this event." Read about Milica Gerhardt's overwhelming experience at the five-day conference on Computer Vision and Pattern Recognition (CVPR). The conference was held in Seattle from 17 to 21 June 2024. Over 12,000 participants attended, including several vera.ai partners. The conference kicked off with two days of workshops, followed by three days of the main track. 

Impressions from a Computer Vision and Pattern Recognition (CVPR) Newbie

vera.ai_representatives.jpg

"As a newbie, I was unprepared for the sheer scale of this event." Read about Milica Gerhardt's overwhelming experience at the five-day conference on Computer Vision and Pattern Recognition (CVPR). The conference was held in Seattle from 17 to 21 June 2024. Over 12,000 participants attended, including several vera.ai partners. The conference kicked off with two days of workshops, followed by three days of the main track.&nbsp;

Over the past months the vera.ai project had the privilege of participating in a number of significant webinars and events aimed at training fact-checkers and raising awareness of civil society on deepfakes and AI manipulations. This included disseminating our latest tools, sharing best practices, and raising awareness about potential threats. The overall aim: to help safeguarding the integrity of elections. Here's insights into what this involved in detail.

Our contribution to the integrity of elections

EDMO-training.png

Over the past months the vera.ai project had the privilege of participating in a number of significant webinars and events aimed at training fact-checkers and raising awareness of civil society on deepfakes and AI manipulations. This included disseminating our latest tools, sharing best practices, and raising awareness about potential threats. The overall aim: to help safeguarding the integrity of elections. Here's insights into what this involved in detail.&nbsp;

In this contribution, Raquel Miguel, Senior Researcher at EU DisinfoLab and a partner in vera.ai, takes a look at AI policy updates and investigates how labelling can (or won't) support when it comes to countering disinformation.

Platforms’ AI policy updates in 2024: labelling as the silver bullet? 

Midjourney image by EUDL.png

In this contribution, Raquel Miguel, Senior Researcher at EU DisinfoLab and a partner in vera.ai, takes a look at AI ploicy updates and investigates how labelling can (or won't) support when it comes to countering disinformation.

On 19 June 2024, six European-funded projects, together with the European Commission, hosted the event "Meet the Future of AI - Generative AI and Democracy" in Brussels. Participants explored the critical interplay between AI and democratic processes, presenting innovative AI-powered solutions to counter disinformation and fostering insightful discussions. Here are the highlights and key takeaways from the sessions.

Generative AI and Democracy: Event Summary and Insights

Krisztina Stump during her opening address

Meet the future 24 (1).jpg

On 19 June 2024 six European-funded projects, together with the European Commission, hosted the event "Meet the Future of AI - Generative AI and Democracy" in Brussels. Participants explored the critical interplay between AI and democratic processes, presenting innovative AI-powered solutions to counter disinformation and fostering insightful discussions. Here are the highlights and key takeaways from the sessions.

On 10 June 2024, EU projects AI4Media, veraAI and AI4Debunk co-organized a joint workshop on Multimedia AI against Disinformation (MAD’24). The event is now in its 3rd edition and was held in Phuket, Thailand. Here, one of the organisers, Luca Cuccovillo of Fraunhofer IDMT, provides a summary of the event and gives an outlook into the future of this successful event series. 

MAD‘24 Workshop on Multimedia AI against Disinformation: Summary & Aftermath

phuket-ai-generated.jpg

On 10 June 2024, EU projects AI4Media, veraAI&nbsp;and AI4Debunk co-organized a joint workshop on Multimedia AI against Disinformation (MAD’24). The event is now in its 3rd edition and was held in Phuket, Thailand. Here, one of the organisers, Luca Cuccovillo of Fraunhofer IDMT, provides a summary of the event and gives an outlook into the future of this successful event series.&nbsp;

In this short conversation with Deutsche Welle's Anna Schild, Andrey Tagarev of ONTOTEXT outlines the Bulgarian company's approach behind the so-called Database of Known Fakes - a repository that collects existing debunks and makes them searchable. Find out more about what's behind it and the approach taken.

Andrey Tagarev of ONTOTEXT outlines approach behind their Database of Known Fakes

Bildschirmfoto 2024-06-04 um 17.12.58.png

In this short conversation with Deutsche Welle's Anna Schild, Andrey Tagarev of ONTOTEXT outlines the Bulgarian company's approach behind the so-called Database of Known Fakes - a repository that collects existing debunks and makes them searchable. Find out more about what's behind it and the approach taken.

On 19 June 2024 from 2.30-7 pm we - together with sister projects AI4Media, Titan, AI4Trust, AI4Debunk and AI-CODE - are hosting another "Meet the Future of AI event" in Brussels, following the success of the 2023 edition. It takes place at the Belgian Institute for Postal Services and Telecommunications (BIPT), Bld du Roi Albert II 35, in the Belgian capital. Reserve your slot now, as spaces are limited. (Last updated: 5 June 2024) 

Save the date: Meet the Future of AI event. UPDATE

Bildschirmfoto 2024-05-17 um 14.17.32.png

On 19 June 2024 from 2.30-7 pm we - together with sister projects AI4Media, Titan, AI4Trust, AI4Debunk and AI-CODE - are hosting another "Meet the Future of AI event" in Brussels, following the success of the 2023 edition. It takes place at the Belgian Institute for Postal Services and Telecommunications (BIPT), Bld du Roi Albert II 35, in the Belgian capital. Reserve your slot now, as spaces are limited. (Last updated: 5 June 2024)&nbsp;

Here's another interview that was conducted at the February 2024 consortium meeting in Thessaloniki. Anna Schild of Deutsche Welle talked to Nicola Righetti of the University Urbino about their work and what outcomes are expected from it. A focus of the team's work: detecting coordinated inauthentic behaviour. 

Anna Schild interviews Nicola Righetti of University Urbino

Bildschirmfoto 2024-06-04 um 16.35.07.png

Here's another interview that was conducted at the February 2024 consortium meeting in Thessaloniki. Anna Schild of Deutsche Welle talked to Nicola Righetti of the University Urbino about their work and what outcomes are expected from it. A focus of the team's work: detecting coordinated inauthentic behaviour.&nbsp;

At our project consortium meeting in Thessaloniki in February 2024 we conducted a number of interviews with project partners and asked them to summarize and highlight some of their work. Here's DW's Jochen Spangenberg talking to project coordinator Symeon Papadopoulos of the Mever team at CERTH-ITI. 

Jochen Spangenberg talks to project coordinator Symeon Papadopoulos of CERTH-ITI. 

spapadopoulos_interview-thessaloniki.png

At our project consortium meeting in Thessaloniki in February 2024 we conducted a number of interviews with project partners and asked them to summarize and highlight some of their work. Here's DW's Jochen Spangenberg talking to project coordinator Symeon Papadopoulos of the Mever team at CERTH-ITI.&nbsp;

The EU Horizon Europe co-funded projects TITAN, AI4Media, AI4Trust and vera.ai have joined forces - several project participants co-wrote a White Paper entitled "Generative AI and Disinformation: Recent Advances, Challenges, and Opportunities." It takes a look at past and present developments and raises issues for the future in this highly relevant field.

White Paper on Generative AI and Disinformation: Recent Advances, Challenges, and Opportunities 

Bildschirmfoto 2024-02-15 um 19.20.30.png

The EU Horizon Europe co-funded projects TITAN, AI4Media, AI4Trust and vera.ai have joined forces - several project participants co-wrote a White Paper entitled "Generative AI and Disinformation: Recent Advances, Challenges, and Opportunities." It takes a look at past and present developments and raises issues for the future in this highly relevant field.

At our project consortium meeting in Thessaloniki in February 2024 we conducted a number of interviews with project partners and asked them to summarise and highlight some of their work. Here's DW's Jochen Spangenberg talking to Luca Cuccovillo of Fraunhofer IDMT. 

Jochen Spangenberg talks to Luca Cuccovillo about their work at Fraunhofer IDMT

Jochen_Luca_Thessaloniki.jpg

At our project consortium meeting in Thessaloniki in February 2024 we conducted a number of interviews with project partners and asked them to summarise and highlight some of their work. Here's DW's Jochen Spangenberg talking to Luca Cuccovillo of Fraunhofer IDMT.&nbsp;

Following a call to action by the European Commission requesting feedback on the systemic risks related to electoral processes, individuals from a number of vera.ai consortium partners sketched out their views on the matter. In this article, some of the contributions made are recapped and shared. The full submission to the EC is attached, too.

Mitigation of systemic risks in the disinformation space - vera.ai perspectives

EU flag-2313980_1280.png

Following a call to action by the European Commission requesting feedback on the systemic risks related to electoral processes, individuals from a number of vera.ai consortium partners sketched out their views on the matter. In this article, some of the contributions made are recapped and shared. The full submission to the EC is attached, too.

For over 18 months now, the vera.ai consortium has been doing research, developed new tools and spread its results. In mid-March 2024 it was time to publish a set of three deliverables that present and summarize the research and development work that has been done in the respective work packages so far. Here, you can find more about them.

Part of vera.ai's R&D work in a nutshell: publicly available deliverables

3_Deliverables.jpg

For over 18 months now, the vera.ai consortium has been doing research, developed new tools and spread its results. In mid-March 2024 it was time to publish a set of three deliverables that present and summarize the research and development work that has been done in the respective work packages so far. Here, you can find more about them.

At our project consortium meeting in Thessaloniki in February 2024 we conducted a number of interviews with project partners and asked them to summarise and highlight some of their work. Here's DW's Anna Schild talking to Richard Rogers of the University of Amsterdam.. 

Anna Schild talks to Richard Rogers about their work at UvA

Richard Rogers interview.png

At our project consortium meeting in Thessaloniki in February 2024 we conducted a number of interviews with project partners and asked them to summarise and highlight some of their work. Here's DW's Anna Schild talking to Richard Rogers of the University of Amsterdam.&nbsp;

Project partner ONTOTEXT is working on a variety of multilingual matters, using AI in this context. In this article, the Bulgaria-based company outlines how state-of-the-art AI models for multilingual entity linking enable working with concepts in 100+ languages, and how this helps to connect and analyze linguistically diverse content.

Gaining Global Insights with Multilingual Entity Linking

2 OT Search.png

Project partner ONTOTEXT is working on a variety of multilingual matters, using AI in this context. In this article, the Bulgaria-based company outlines how state-of-the-art AI models for multilingual entity linking enable working with concepts in 100+ languages, and how this helps to connect and analyze linguistically diverse content.

A team of vera:ai researchers from the Fraunhofer Institute for Digital Media Technology (IDMT), Germany and the Centre for Research and Technology Hellas (CERTH-ITI), Greece  have developed a new synthetic speech detection dataset. Find out more why they did it and what this contains.

ODSS: An Open Dataset of Synthetic Speech. Call to use and cooperate

Synthetic Utterances Dataset.png

A team of vera:ai researchers from the Fraunhofer Institute for Digital Media Technology (IDMT), Germany and the Centre for Research and Technology Hellas (CERTH-ITI), Greece &nbsp;have developed a new synthetic speech detection dataset. Find out more why they did it and what this contains.

Co-hosted with the ACM International Conference on Multimedia Retrieval (ICMR’24) that takes place in Phuket, Thailand, from 10-13 June 2024, vera:ai project partners are co-organising the 3rd ACM International Workshop on Multimedia AI against Disinformation (MAD’24) in Phuket, Thailand. Here's more, including the respective Call for Papers.

You can go MAD again! Join us at the International Workshop on Multimedia AI against Disinformation

IMG_5220.jpg

Co-hosted with the ACM International Conference on Multimedia Retrieval (ICMR’24) that takes place in Phuket, Thailand, from 10-13 June 2024, vera:ai project partners are co-organising the 3rd ACM International Workshop on Multimedia AI against Disinformation (MAD’24) in Phuket, Thailand. Here's more, including the respective Call for Papers.

This post explains the basics behind a paper entitled “VERITE: a robust benchmark for multimodal misinformation detection accounting for unimodal bias”, published in the International Journal of Multimedia Information Retrieval (IJMIR). It has been authored by researchers of the mever group at project partner CERTH-ITI.

Overcoming Unimodal Bias in Multimodal Misinformation Detection

mever photo 2.png

This post explains the basics behind a paper entitled “VERITE: a robust benchmark for multimodal misinformation detection accounting for unimodal bias”, published in the International Journal of Multimedia Information Retrieval (IJMIR). It has been authored by researchers of the mever group at project partner CERTH-ITI.

A project by vera.ai, entitled “The Persistence and Resilience of Misinformation in the Face of Fact Checking,” was facilitated by Fabio Giglietto, Massimo Terenzi (UNIURB) and Richard Rogers (UvA) at the Digital Media Initiative Winter School at the University of Amsterdam. We are pleased to share with you the results.

The persistence and resilience of misinformation in the face of fact checking

newplot (1).png

A project by vera.ai, entitled “The Persistence and Resilience of Misinformation in the Face of Fact Checking,” was facilitated by Fabio Giglietto, Massimo Terenzi ( UNIURB) and Richard Rogers (UvA) at the Digital Media Initiative Winter School at the University of Amsterdam. We are pleased to share with you the results.

The EMNLP (Empirical Methods in Natural Language Processing) conference is one of the top NLP conferences. KInIT researchers Róbert Móro and Ján Čegiň presented three full papers at this prestigious event. Here’s more about what they did, and the related works.

Project partner KInIT at the EMNLP conference in Singapore represented with three papers

conference event-venue.jpg

The EMNLP (Empirical Methods in Natural Language Processing) conference is one of the top NLP conferences. KInIT researchers Róbert Móro and Ján Čegiň presented three full papers at this prestigious event. Here’s more about what they did, and the related works.

A new paper that includes recent findings of disinformation analysis has been published in the Journal of Quantitative Description: Social Media. It is entitled “Cracking Open the European Newsfeed”, co-authored by vera.ai team members based at University of Urbino “Carlo Bo”, Italy. The paper contributes to the ongoing effort to describe and quantify the quality of information shared on large social media platforms.

Unlocking Insights: paper "Cracking Open the European Newsfeed" is finally out in JQD:DM

Bildschirmfoto 2023-12-21 um 13.01.23.png

A new paper that includes recent findings of disinformation analysis has been published in the Journal of Quantitative Description: Social Media. It is entitled “Cracking Open the European Newsfeed”, co-authored by vera.ai team members based at University of Urbino “Carlo Bo”, Italy. The paper contributes to the ongoing effort to describe and quantify the quality of information shared on large social media platforms.

It's that time of the year again .... 2023 is nearing its end. That is why we would like to wish all our friends and followers a time for reflection and especially health and happiness. May the year ahead be a good one - and let's all strive to make it a better place - and do so in whichever way we can.

Season's Greetings: all the best for a happy and healthy 2024 

Winter 2.png

It's that time of the year again .... 2023 is nearing its end. That is why we would like to wish all our friends and followers a time for reflection and especially health and happiness. May the year ahead be a good one - and let's all strive to make it a better place - and do so in whichever way we can.

Would you like to meet some of the VERA.AI team members? If so: don’t miss the chance and look for us in Nuremberg from 4-7 December 2023 at the IEEE Workshop in Information Forensics and Security.

Meet us in Nuremberg at the IEEE WIFS 2023 Conference!

snowflakes-xmas market.jpg

Would you like to meet some of the VERA.AI team members? If so: don’t miss the chance and look for us in Nuremberg from 4-7 December 2023 &nbsp;at the IEEE Workshop in Information Forensics and Security.

Who is behind vera.ai? Meet the people and teams that make up the vera.ai project. #10 Kamila Koronska, University of Amsterdam

Who is behind vera.ai? Meet the people and teams that make up the vera.ai project. #10 Kamila Koronska

whoisbehindveraai_kkoronska.jpg

Who is behind vera.ai? Meet the people and teams that make up the vera.ai project. Check-out interview #10 Kamila Koronska, University of Amsterdam.

Dataset of 9.000 AI-generated images described in the paper "Synthbuster: Towards Detection of Diffusion Model Generated Images" (Quentin Bammey, 2023, Open Journal of Signal Processing).

Synthbuster: Towards Detection of Diffusion Model Generated Images

fabio-oyXis2kALVg-unsplash.jpg

Dataset of 9.000 AI-generated images described in the paper "Synthbuster: Towards Detection of Diffusion Model Generated Images" (Quentin Bammey, 2023, Open Journal of Signal Processing).

In this article, the vera.ai consortium provides a series of policy recommendations aimed at enhancing data access tools, streamlining application processes, and promoting transparency in the use of social media data for academic and investigative purposes. Recommendations are largely based on past experiences and learnings to date.

vera.ai Policy Relevant Evidence

board-1647323_1920.jpg

In this article, the vera.ai consortium provides a series of policy recommendations aimed at enhancing data access tools, streamlining application processes, and promoting transparency in the use of social media data for academic and investigative purposes. Recommendations are largely based on past experiences and learnings to date.

A year in review – vera.ai turned ONE

It was the 15th of September 2022 when the vera.ai project officially started and sent 14 partners from across Europe on the mission to develop and build trustworthy AI (Artificial Intelligence) solutions in the fight against disinformation. A couple of days later, the consortium got together for the first time in Marathon, Greece. That is how the vera.ai-journey started. 

What happened in vera.ai’s first year? What has the project accomplished and where is it heading? Deutsche Welle's Anna Schild looks back at an active year and highlights a selection of achievements.

A year in review - vera.ai turned ONE

brooke-lark-pGM4sjt_BdQ-unsplash.jpg

It was the 15th of September 2022 when the vera.ai project officially started and sent 14 partners from across Europe on the mission to develop and build trustworthy AI solutions in the fight against disinformation. What happened in the first year of this journey? What has been accomplished, and where is it all going? Deutsche Welle's Anna Schild looks back at an active year and highlights a selection of achievements.

Many researchers attempt to develop tools that support journalists and fact-checkers in their verification tasks with a variety of tools and services. One such tool is the Database of Known Fakes (DBKF), which is being developed and further enhanced in vera.ai by Bulgarian company ontotext. Here's more about it.

The Database of Known Fakes: a useful tool in the verification workflow

ONTO_DBKF_rahmen.png.jpg

Many researchers attempt to develop tools that support journalists and fact-checkers in their verification tasks with a variety of tools and services. One such tool is the Database of Known Fakes (DBKF), which is being developed and further enhanced in vera.ai by Bulgarian company ontotext. Here's more about it.

The vera.ai project was showcased on 30 June 2023 during the main worldwide gathering of fact-checkers at Global Fact 10 in Seoul, South-Korea.

vera.ai featured at Global Fact 10 in Seoul, Korea.

Denis Teyssou presenting the vera.ai project at Global Fact 10 in Seoul, on June 30.

33LQ7ZH-Preview_W.jpg

Our innovation manager Denis Teyssou (AFP) conducted a 50 minutes dedicated breakout session named AI vs disinformation: the vera.ai project: enhancing the InVID-WeVerify plugin. Around 40 participants, including well-known fact-checkers like Craig Silverman (ProPublica), attended the session in-person. The talk was recorded for online participants on the <a href="https://www.globalfact10.com" target="_blank" rel="noopener noreferrer">Global Fact 10 website</a>.

Achieving generalisation is one of the fundamental challenges in machine and deep learning methods. In an effort to contribute to the investigation of such an important challenge, the MeVer group performed a study on the setting of detecting synthetic images across different concept classes, i.e. synthetic image detection in a cross-concept scenario.

Improving Synthetically Generated Image Detection in Cross-Concept Settings

CERTH_MAD_2.PNG

Achieving generalisation is one of the fundamental challenges in machine and deep learning methods. In an effort to contribute to the investigation of such an important challenge, the MeVer group performed a study on the setting of detecting synthetic images across different concept classes, i.e. synthetic image detection in a cross-concept scenario.

Here's a video recording of a talk by Fabio Giglietto of the University of Urbino, given in June 2023 at the Queensland University of Technology (QUT) . Topic: using OpenAI models to identify the most salient topics circulated via Facebook links in the run-up to the Italian general elections.

OpenAI Models for Topic Modelling in Social Media Analysis

Bildschirmfoto 2023-07-07 um 14.35.00.png

Here's a video recording of a talk by Fabio Giglietto of the University of Urbino, given in June 2023 at the Queensland University of Technology (QUT). Topic: using OpenAI models to identify the most salient topics circulated via Facebook links in the run-up to the Italian general elections.

Verification Plug-in: Overview

EUDL webinar alert: join us on 12 Dec 2024 for a talk on verification tools and guidelines

vera.ai at IFCN's Global Fact 11 in Sarajevo

Our contribution to the integrity of elections

vera.ai featured at Global Fact 10 in Seoul, Korea.

Using Google Images for image similarity search

The Verification Plugin: State of play and overview