Project HOBBIT retweeted
#Job: Open PhD position on #MachineLearning with multiple representations on knowledge graphs. Apply here: linkedin.com/jobs/view/34501…. Please retweet. #openscience #knowledgegraph @unipb
Project HOBBIT retweeted
We are live! Investigating and deploying human centred explanations for AI systems combining machine learning and knowledge graphs @unipb @UvA_Amsterdam @NCSR_Demokritos @DATEV @eusatcen @webLyzard
Funded by @HorizonEU. More:
uni-paderborn.de/en/news-ite…
Project HOBBIT retweeted
The 21st International Semantic Web Conference is over 🥺 But the presentation videos are still online! 😃
iswc2022.semanticweb.org/ind…
The talks of our group members are also available on YouTube:
youtube.com/playlist?list=PL…
#iswc2022 #iswc_conf #semanticWeb @iswc_conf (thread ...)
Project HOBBIT retweeted
Happy to announce that our paper "Hashing the Hypertrie: Space- and Time-Efficient Indexing for SPARQL in Tensors" was accepted at @iswc_conf
22. Thx to the co-authors Charlotte Behning, @LixiCorrados, @saleem_muhamad, @NgongaAxel #iswc2022 #iswc_conf (T-*=#tentris=our system)
Project HOBBIT retweeted
I-AID: Identifying Actionable Information From Disaster-Related Tweets
Authors: @hamadazahera @RrichaJalota @MAhmedSherif @NgongaAxel
Published in IEEE Access @IEEEAccess
xitter.cf/hamadazahera/status/14… [4/4]
I am delighted to share that our paper “I-AID: Identifying Actionable Information from Disaster-related Tweets” has been accepted in the IEEE Access Journal @DiceResearch @RrichaJalota @MAhmedSherif @NgongaAxel preprint here papers.dice-research.org/202… (1/2)
Project HOBBIT retweeted
We are looking for top talents for our new project OpenGPT-X in which we aim to improve large neural language models and their usability for dialogue systems & question answering. Very large supercomputing facilities are available. You can apply here: iais.fraunhofer.de/en/career…
Project HOBBIT retweeted
Recently, I got many questions about the role of neural language models in #AI (due to our new OpenGPT-X project) and decided to write a blog post on the topic: jens-lehmann.org/blog/neural… #languagemodels #gpt3 #ArtificialIntelligence #OpenGPTX
Project HOBBIT retweeted
Open position (full professor in #machinelearning) @unipb . Check out the description at uni-paderborn.de/fileadmin/z…
Project HOBBIT retweeted
Ante-Hoc Explainable #machinelearning on #KnowledeGraphs. We combine reinforcement learning with a novel refinement operator to accelerate #XAI. Check it out at arxiv.org/abs/2106.15373 @DiceResearch
Project HOBBIT retweeted
Two exciting fully funded PhD positions on explainable AI and multi-view learning @DiceResearch. Check out details here: dice-research.org/jobs and let me know if you have any questions #AI #KnowledgeGraphs #ML #OpenScience
Project HOBBIT retweeted
One of those papers one is rather well pleased with ... Go @_CaglarDemir
Looking forward to presenting our work on Convolutional Complex KGE (bit.ly/3xe0Ayj).We provide an answer to Trouillon et al's final question on complex embeddings @eswc_conf Session #11 June 10th #eswc2021 #AI #KnowledgeGraph #MachineLearning
@DiceResearch @NgongaAxel
Project HOBBIT retweeted
Looking forward to presenting our work on Convolutional Complex KGE (bit.ly/3xe0Ayj).We provide an answer to Trouillon et al's final question on complex embeddings @eswc_conf Session #11 June 10th #eswc2021 #AI #KnowledgeGraph #MachineLearning
@DiceResearch @NgongaAxel
Project HOBBIT retweeted
In arxiv.org/abs/2105.12524, we (1) show that common benchmarks for knowledge graph embeddings and link prediction contain out-of-vocabulary entities and (2) provide improved datasets free of this limitation. Enjoy! #AI #KnowledgeGraph #MachineLearning @DiceResearch
Project HOBBIT retweeted
Looks like complex spaces enjoy a convolution to achieve better embeddings. Our approach ConEx achieves nice results on benchmarks with a comparatively small number of parameters. More at github.com/dice-group/Convol… @_CaglarDemir #OpenScience #KnowledgeGraph #machinelearning
Project HOBBIT retweeted
Another pleasant news: Our paper on the verbalization of link specifications for the integration of #KnowledgeGraphs is now available at sciencedirect.com/science/ar…. Code: github.com/dice-group/limes. This is a #OpenSource contribution to explainable #AI. More coming up. Enjoy.
Project HOBBIT retweeted
Need to link #KnowledgeGraphs? Our article on the #OpenSource link discovery framework LIMES (github.com/dice-group/limes) is now available at link.springer.com/content/pd…. With LIMES, we support #OpenScience on the integration of knowlege graphs @MAhmedSherif @kvndrsslr @mommi84
Project HOBBIT retweeted
Our paper “5* Knowledge Graph Embeddings with Projective Transformations” by Mojtaba Nayyeri, @SaVahdati, Can Aykul, @JLehmann82 is just presented at #AAAI2021! Join us at the conference or find all information in our blog post: sda-research.medium.com/5-kn…
Project HOBBIT retweeted
We are very proud that our group leader @JLehmann82 was selected as Fellow for @ELLISforEurope in recognition of his achievements in AI research. He is the first researcher at @UniBonn and the second researcher at @Fraunhofer to receive this honour. #AI @FraunhoferIAIS
Project HOBBIT retweeted
Interesting in cardinality estimations (e.g., for query planning)? Check out our latest paper on the subject matter at semantic-web-journal.net/con…. Congrats to @UQudus @saleem_muhamad and thanks to the reviewers and editor. #LinkedData #knowledgegraphs @knowgraphs @DiceResearch
Project HOBBIT retweeted
In case you missed my presentation at @iswc_conf #ISWC2020, you can find my slides at bit.ly/2IaHepg and my presentation at youtu.be/-IMERb_1y1U. We are running for the best student paper, check out our paper. @DiceResearch @DwaraknathG @thiagocasfer @NgongaAxel