Philipp Christmann

The preprint for our WWW 2024 paper (joint work with @ZhenJia9 and Gerhard Weikum) on "Faithful Temporal Question Answering over Heterogeneous Sources" is now available at: arxiv.org/abs/2402.15400. Code, new TIQ benchmark and data at faith.mpi-inf.mpg.de.

Delighted to share that our full paper on „Faithful Temporal Question Answering over Heterogeneous Sources“ got accepted at The Web Conference 2024! Joint work with @ZhenJia9 (first author) and Gerhard Weikum. Looking forward to traveling to Singapore in May! #thewebconf2024

I will present our work on „Explainable Conversational QA over Heterogeneous Sources via Iterative Graph Neural Networks“ tomorrow at 9 AM in the „Conversational search and recommendation“ session at #SIGIR2023 in Taipei. Joint work with @RishirajSahaRoy and Gerhard Weikum.

@d_aumiller @RishirajSahaRoy CLOCQ can provide disambiguations and relevant KB-facts. If the KB-facts are not required, they can be left untouched. The 370 GB memory requirement is only for the server, which loads a whole Wikidata dump in memory (~2TB originally). The client doesn't require (much) memory.

@RishirajSahaRoy @PhilippChr97 But can I just do the NEL annotations over text? It seems overly complex for the task that I want to do, which is simply to annotate entities in text. Not the whole QA bit associated with that. 370 GB memory requirement seems oddly high, given that spacy-nel works with <2GB.

During our lecture "Text Analytics", we were talking about Named Entity Linking (NEL) and it turns out that there is *no* system that works "out of the box" for this?? Like there are packages, but none of them would let me actually run this with a single pip install + import.

@ZhenJia9 @RishirajSahaRoy @SIGIRConf Thanks @ZhenJia9!

@PhilippChr97 @RishirajSahaRoy @SIGIRConf Congratulations!!

Delighted to share that our full paper on "Explainable Conversational Question Answering over Heterogeneous Sources via Iterative Graph Neural Networks" is accepted at #SIGIR2023. Joint work with @RishirajSahaRoy and Gerhard Weikum. Further details coming soon! #convqa @SIGIRConf

@RishirajSahaRoy A dedicated website is now available at qa.mpi-inf.mpg.de/compmix/. Anecdotal questions that exhibit complex phenomena are shown as examples. More details coming soon!

Just gave a talk on how our #WSDM22 work CLOCQ (with @RishirajSahaRoy and Gerhard Weikum) can be used for entity and relation linking in the SMART Task 3.0. Thanks for the opportunity @nand_ai @MohnishDubey @debayan et al. Links: clocq.mpi-inf.mpg.de, smart-task.github.io

@RishirajSahaRoy Thanks a lot Rishi! Definitely a result of the many research discussions we had in the last four years, and your great work as a supervisor! :)

Finally we have the #mpireloaded pic. Tried photoshopping @lauradietz99 in but it’s best to have two pics .. #mpii #sigir2022 .. had a great time catching up with old friends and meeting new colleagues. Adios Madrid.

I will present our #SIGIR2022 paper on „Conversational Question Answering on Heterogeneous Sources“ tomorrow in the Conversational IR session in Madrid. Joint work with @RishirajSahaRoy and Gerhard Weikum. Link to paper: dl.acm.org/doi/10.1145/34…. #SIGIR

The code and data for our #SIGIR2022 paper "Conversational Question Answering on Heterogeneous Sources" (with @RishirajSahaRoy and Gerhard Weikum), including the new ConvMix benchmark, are now available on the project website: convinse.mpi-inf.mpg.de. #SIGIR #questionanswering

@DrCh0le Thank you!

@PhilippChr97 Congrats!

Very happy that my doctoral consortium paper on „Explainable Conversational Question Answering over Heterogeneous Sources“ has been accepted at #SIGIR2022 as well. Looking forward to the discussions in Madrid! #qa #convqa #SIGIR

@svakulenk0 @RishirajSahaRoy @scai_workshop @svakulenk0 Sure, I will be happy to do so!

@RishirajSahaRoy @PhilippChr97 @PhilippChr97 pls bring your poster to @scai_workshop afterwards, looking forward to see you there!

Very happy to feature in the list of 21 outstanding PC members at WSDM 2022! :) #wsdm2022

A preprint for our WSDM 2022 long paper "Beyond NED: Fast and Effective Search Space Reduction for Complex Question Answering over Knowledge Bases" is now online at arxiv.org/abs/2108.08597 (joint work with @RishirajSahaRoy and Gerhard Weikum). #ned #questionanswering #wsdm2022