rtroncy Profile Banner
Raphaël Troncy Profile
Raphaël Troncy

@rtroncy

Followers
3K
Following
1K
Media
409
Statuses
8K

Prof and Researcher at @EURECOM. AI, Knowledge Graph, Open Data, LLM, Information Extraction, Recommender Systems, https://t.co/CgUPPwuBlP

Sophia Antipolis (FR)
Joined May 2009
Don't wanna be here? Send us removal request.
@yoan_chabot
Yoan Chabot
7 days
100+ citations! 🎊 Our survey on Semantic Table Interpretation-a @Orange_Future & @EURECOM collaboration-reviews how to extract meaningful information from tabular data. Proud to co-author with @yansera1 @rtroncy @vp_huynh. 🔗 https://t.co/cO5l99Nm6f #SemanticWeb #KnowledgeGraph
0
2
3
@rtroncy
Raphaël Troncy
8 days
@claudeai @GeminiApp @ChatGPTapp @MistralAI @Alibaba_Qwen We released our best fine-tuned model, named PERSEUS. Benchmark results detailed at https://t.co/kfzG9dhfCz Joint work between @EURECOM and @lettria_fr as part of the #LettRAGraph project funded by @Bpifrance #iswc2025
0
0
0
@rtroncy
Raphaël Troncy
8 days
We have then compared numerous closed LLMs (@claudeai @GeminiApp @ChatGPTapp) with fine-tuned open weights LLMs (@MistralAI @Alibaba_Qwen Gemma3, Phi4). There are clearly 2 leagues: fine-tuned specialized models OUTPERFORM larger closed models and are more efficient. #iswc2025
1
0
0
@rtroncy
Raphaël Troncy
8 days
Excited to have presented the Text2KGBench-LettrIA benchmark at the KB-LLM workshop at @iswc_conf. ~5k sentences annotated with respect to 19 ontologies to train #LLMs to build Knowledge Graphs. #iswc2025 Paper: https://t.co/HNnR6jNgDa Slides: https://t.co/agcWABHw2B
1
1
7
@rtroncy
Raphaël Troncy
30 days
Very proud to have presented with @IngerLeemans @odeuropa selected among the 4 winners in the Research Category! We developed digital tools to recognize smell experiences in images and texts and we reconstructed scents from the past! @EURECOM
@europanostra
Europa Nostra
30 days
👋 Hi from Brussels! 🇧🇪 We’re thrilled to be hosting the #EuropeanCulturalHeritageSummit2025, starting with the Excellence Day showcase of our 30 #EuropeanHeritageAwards laureates. Our high-level Awards Ceremony is tomorrow! 🎥 Bookmark the livestream: https://t.co/fOGuPWMY1H
0
6
7
@PasqLisena
Pasquale Lisena
2 months
If you are at @ACMRecSys, pass by spot #21 and try our demo for generating your playlist by choosing its title in Natural Language. demo online https://t.co/F8vZggJSPb @rtroncy @RebboudY #musicReccomendation #ai #recsys
0
1
1
@PasqLisena
Pasquale Lisena
2 months
Exploiting playlist titles can push good recommendation and feed a cold-start strategy for #playlist generation. Check our presentation of this morning at @ACMRecSys + code and links at last slide https://t.co/sVEYsCvVNt // cc @rtroncy #recsys #ai
0
1
0
@PasqLisena
Pasquale Lisena
2 months
Here at #CLEF2025 @CLEF_Conference presenting the strategy for subjectivity and scientific discourse analysis adopted by the #ClimateSense team for the #CheckThat! challenge. More at https://t.co/ktNDujzpoM // @rtroncy @evhart @halani
0
4
4
@GwenComte
EURECOM - Library
2 months
The 2025 Most Influential Researcher Award Honorable Mention in recognition of outstanding & dynamic contributions between 2015 and 2024 was awarded to Prof. @paolopapotti in the field of databases (ranked #35) & Prof. @rtroncy in the field of knowledge engineering (ranked #80)
0
1
5
@ELRAnews
ELRA
3 months
#LREC2026 2nd Calls for Papers & Proposals are out Deadline: October 17, 2025 Main Conference Papers: https://t.co/sLybFmZ6cZ Workshop Proposals: https://t.co/p3EC1hBHif Tutorial Proposals: https://t.co/OguHdUs61Q Palma de Mallorca, May 11-16, 2025 #nlproc
0
9
5
@chavalarias
David Chavalarias
4 months
Le message publié par X (ex-Twitter) face à l’enquête française est une manœuvre de désinformation classique doublée d'un aveu (X ne respecte pas la loi) et d'un avertissement (votre droit n'est pas son problème) : Le message de X ➡️ https://t.co/ZoQlZUaTVq Décryptage.🧵👇 1/46
@GlobalAffairs
Global Government Affairs
4 months
French authorities have launched a politically-motivated criminal investigation into X over the alleged manipulation of its algorithm and alleged “fraudulent data extraction.” X categorically denies these allegations. This investigation, instigated by French politician Eric
2
4
12
@rtroncy
Raphaël Troncy
4 months
Can We Trust the Judges? This is the question we asked in validating factuality evaluation methods via answer perturbation. Check out the results at the #EvalLLM2025 workshop at #TALN2025. Blog: https://t.co/jumKpMqMMS Watch: https://t.co/IKOrRRiPrD Play: https://t.co/Xg6ZDXbVgx
0
0
3
@GwenComte
EURECOM - Library
5 months
Congratulations to Youssra REBBOUD who successfully defended her PhD thesis today, supervised by @rtroncy @PasqLisena @marazul https://t.co/VQ1FsdKAxu @EURECOM
0
2
4
@rtroncy
Raphaël Troncy
5 months
Thanks to the jury members for having evaluated this work: @elen2016 Enrico Motta @serena_villata @paolopapotti / co supervised with @PasqLisena
0
0
1
@rtroncy
Raphaël Troncy
5 months
The final contribution is a rule-based system that again makes use of fine-grained causal relationships between events for explaining fact-checking results. Evaluated on the #FEVEROUS and #AVeriTeC datasets.
1
0
1
@rtroncy
Raphaël Troncy
5 months
A third contribution is a method that relies on the JoinGT system but incorporates fine-grained causal relationships between events from FARO for generating narratives and stories of events. We demonstrated that qualitatively, the generated summaries are more complete.
1
0
2
@rtroncy
Raphaël Troncy
5 months
This dataset can then be used to fine-tune language models (RoBERTa and REBEL) to be more effective at detecting fine-grained causal relationships between events in texts. A demonstrator making use of these models is available at https://t.co/zMEcKAZ3O0
1
0
1
@rtroncy
Raphaël Troncy
5 months
A second major contribution is the largest dataset containing fine-grained annotations of causal relationships between events (enable, prevent, intend, directly cause). This dataset comes from common sense (ATOMIC), news corpus (CNC) and synthetic #LLM generated data.
1
0
1
@rtroncy
Raphaël Troncy
5 months
Youssra also investigated how capable #LLMs are for reverse-engineering competency questions and deriving user stories out of formal ontologies. Results demonstrate that recent large models such as GPT4 are effective for general domain.
1
0
2
@rtroncy
Raphaël Troncy
5 months
A first contribution is the Facts and Events Relationships Ontology (FARO) that harmonizes different models representing event relations. This is available at https://t.co/qqvYLBimIA
1
0
2