#cikm2023 — Public Fediverse posts

Building a #knowledgeGraph? At #CIKM2023, Lucas Jarnac presented how to bootstrap it by selecting subparts of interest from @wikidata with zero-shot analogical inference.
Congrats Lucas! #proudSupervisor
Paper: https://arxiv.org/pdf/2306.16296
#ArtificialIntelligence #machineLearning

Welcome to Day #3 of #CIKM2023! @ingo gives a run-down of the final day of the conference with a keynote by Steffen Staab, the CIKM business meeting for reflection and the future of the conf, more sessions, and a conference closing.

RT by @wikiresearch: Don't miss today our poster "A Comparative Study of Reference Reliability in Multiple Language Editions of Wikipedia" w/ @toshaaa_n @e__migrante @mad_astronaut @nekozzang

#CIKM2023 poster board #39
📝 https://dl.acm.org/doi/10.1145/3583780.3615254 https://twitter.com/elaragon/status/1716721246845472851

RT by @wikiresearch: Building a #knowledgeGraph? At #CIKM2023, @lucas_jarnac presented how to bootstrap it by selecting subparts of interest from @Wikidata with zero-shot analogical inference.
Congrats Lucas! #proudSupervisor
Paper: https://arxiv.org/pdf/2306.16296
#ArtificialIntelligence #machineLearning https://twitter.com/piermonn/status/1716835079480119699

Yulan He discusses calculating epistemic uncertainty in her #CIKM2023 keynote and describes the PLM Classifier Uncertainty Explanation (CUE) framework.

The HINT model can split sentences with mixed sentiment. Yulan He discusses her model that is fine-tuned for classification. #CIKM2023

Yulan He overviews what they will cover in today's #CIKM2023 keynote: a classifier w/ interpretability, pre-training/fine-tuning, in-context learning, and leveraging the models for explainable recommendation.

@ingo opens #CIKM2023 Day Two with a preview of the conference banquet tonight at Aston Villa Football Club.

It was a great pleasure to welcome everybody to #CIKM2023! Day 1 is done, now looking forward to the next day!

#Birmingham

That wraps up the presentations for #CIKM2023 Day 1. Join us in Affinity 1 for the posters and demos!

Time for another #CIKM2023 coffee break! ☕ Join us back at 4:30 pm for the final sessions of the day on Knowledge Discovery and Retrieval, Users & Behaviour, Transfer Learning, Computer Vision, User Modelling, Knowledge Representation, Sequence Modelling, & Recommendation.

#CIKM2023 attendees will now split into 8 concurrent sessions for the first (10:30-12:30) presentation slot. See https://uobevents.eventsair.com/cikm2023/programme---monday-23rd-october to choose your session and a summary of each respective set of presentations.

And now for a coffee break following the completion of Chirag Shah's #CIKM2023 keynote.

In his #CIKM2023 keynote Chirag Shah (UWashington) shows when using ChatGPT, it lack of confidence in answering when the system is questioned. "If I don't have the literacy to question the answer, I may not realize this is all wrong."

Brisk (41°F) morning for the second day of the #CIKM2023 pre-conference (workshops & industry day) in Birmingham, UK but it's finally not raining, so I'll take it!

With the completion of the #CIKM2023 PhD Symposium and the tutorials, the first day of the pre-conference has come to the end.

Join us tomorrow for the second day of CIKM with Workshops* and Industry Day†

* https://uobevents.eventsair.com/cikm2023/pre-conference-programme#:~:text=Sunday%2022nd%20October%20%2D-,Workshops,-Arrival%20Refreshments%20%2D%208

† https://uobevents.eventsair.com/cikm2023/industry-day-programme

Wow, thanks @cikm2023 for live-tooting about the tutorials and doctoral consortium at #cikm2023.

In using large tabular data as a source, data with decision trees (using GAM-based Explanation of Forests) were initially used with one/two features per tree to produce a uni-variate or bi-variate functions, which are easy to plot. #CIKM2023

Next at the #CIKM2023 PhD Symposium, Alberto Veneri presents their research Explaining Learning-to-Rank Methods to Improve Them. Their explanations are system-centric — make the models simpler to improve the models themselves.

Kislay Raj is presenting their PhD research, "A neuro-symbolic approach to enhance Interpretability of Graph neural network through the integration of external knowledge" at the #CIKM2023 PhD Symposium.

And the fourth tutorial, continued from the morning session, "Reasoning beyond Triples: Recent Advances in Knowledge Graph Embeddings" is continuing into the afternoon at #CIKM2023 pre-conference.

Again, their site: https://kg-beyond-triple.github.io/
Paper: https://dl.acm.org/doi/10.1145/3583780.3615294

Another tutorial on User Simulation for Evaluating Information Access Systems is being presented at #CIKM2023

Tutorial web site: https://usersim.ai/cikm2023-tutorial/
Paper: https://dl.acm.org/doi/10.1145/3583780.3615296

@femoraesg, @hmproenca, @fire_fox, Javier Albert, and Dima Goldenberg are presenting their tutorial, "Uplift Modeling: from Causal Inference to Personalization" this afternoon at #CIKM2023 pre-conference.

Slides: https://docs.google.com/presentation/d/e/2PACX-1vQUGwe2xxHnv3j_ti8oHcwFuVNqAnNWK73LmEOw8uNkc869174kaT3qsGHaItdoFQ/pub?start=false&loop=false&delayms=3000#slide=id.p14
Paper: https://dl.acm.org/doi/10.1145/3583780.3615298

The #CIKM2023 pre-conference tutorials are now continuing after lunch with 3 new tutorials and one from the morning session continuing into the afternoon.

Ready for the first day of #cikm2023 pre-conference at University of Birmingham with a plethora of tutorials ranging from IR+NLP to AI and Knowledge Graphs and much more!

https://uobevents.eventsair.com/cikm2023/pre-conference-programme

Now en route to CIKM 2023: Philly (PHL) ✈︎ (BHX) Birmingham UK #cikm2023 @cikm2023

My #CIKM2023 pre-conference run - from (almost) the conference location in Birmingham to (almost) my office in Wolverhampton.

https://www.instagram.com/p/CyUT64dqw0S/?igshid=MzRlODBiNWFlZA==

We’re still accepting online registrations for #CIKM2023. You can also still register for our industry day. More info at https://cikm2023.org/registration-rates.

We are very happy to have Yulan He (Interpretable Natural Language Understanding), Chirag Shah (Generative AI and the Future of Information Access ) and Steffen Staab (Knowledge Graphs for Knowing More and Knowing for Sure ) as our keynote speakers at #CIKM2023!

More info and registration (online and limited in-person places) are https://cikm2023.org/

Proud to announce our new paper, "Relevant Entity Selection: Knowledge Graph Bootstrapping via Zero-Shot Analogical Pruning" with Lucas Jarnac and Miguel Couceiro, accepted in #cikm2023

#knowledgeGraph #machineLearning #graphEmbedding #transferLearning #zeroShot #openScience

Proud to announce our new paper, "Relevant Entity Selection: Knowledge Graph Bootstrapping via Zero-Shot Analogical Pruning" with Lucas Jarnac and Miguel Couceiro, accepted in #cikm2023

#knowledgeGraph #machineLearning #graphEmbedding #transferLearning #zeroShot #openScience

Proud to announce our new paper, "Relevant Entity Selection: Knowledge Graph Bootstrapping via Zero-Shot Analogical Pruning" with Lucas Jarnac and Miguel Couceiro, accepted in #cikm2023

#knowledgeGraph #machineLearning #graphEmbedding #transferLearning #zeroShot #openScience

Proud to announce our new paper, "Relevant Entity Selection: Knowledge Graph Bootstrapping via Zero-Shot Analogical Pruning" with Lucas Jarnac and Miguel Couceiro, accepted in #cikm2023

#knowledgeGraph #machineLearning #graphEmbedding #transferLearning #zeroShot #openScience

Proud to announce our new paper, "Relevant Entity Selection: Knowledge Graph Bootstrapping via Zero-Shot Analogical Pruning" with Lucas Jarnac and Miguel Couceiro, accepted in #cikm2023

#knowledgeGraph #machineLearning #graphEmbedding #transferLearning #zeroShot #openScience