📢Now available the much anticipated and greatly needed: 🆕Meltemi (/meltémi/), the first open-source Greek LLM trained by #ILSP of @athenaRICinfo . Read more: https://t.co/kr4qFKXTUh
If you are an enthusiast of AI don't miss the opportunity to learn all the details about our AI Living Lab and tools in its final event!😃
When❓ Next November 27th
Registrations and more info 👉 https://t.co/Y4UiAaZ1y2
#BepartofIntelComp#H2020#policymaking#data
AUEB NLP Group meeting, *Fri* 16 Sept, 17:00-19:00 (remotely): Dimitris Pappas presenting the work of his PhD thesis, titled "Biomedical Question Answering Systems". Pre-defence presentation to his 3-member committee, all welcome. @dvpappas
https://t.co/MlyoKFv89B
"A Neural Model for Joint Document and Snippet Ranking in Question Answering for Large Document Collections", D. Pappas & I. Androutsopoulos. ACL 2021, main conference, session 11D (Tuesday). https://t.co/gFswYDfKCT #ACL2021NLP@dvpappas@ionandrou
Humans learn by doing - Language Models such as BERT are pretrained to predict masked words and then fine-tuned on the target task. We propose another step where we train the language model on a multiple diverse set of tasks before fine-tuning. Paper: https://t.co/ISN0JMuzuL #NLP
Last week, EleutherAI released two checkpoints for GPT Neo, an *Open Source* replication of OpenAI's GPT-3
These checkpoints, of sizes 1.3B and 2.7B are now available in🤗Transformers!
The generation capabilities are truly🤯, try it now on the Hub: https://t.co/LbLXuCNYip
An Experimental Document & Snippet Retrieval Search Engine for the CORD-19 (COVID-19) dataset, based on our best BioASQ7 system. #COVID19@allen_ai@BioASQ
Online demo: https://t.co/gZymDgF0Ob
Papers: https://t.co/DZjwe8PlYb https://t.co/5bzLcP2ZqG
Code: https://t.co/hEP8XWNDI1
With @AUEBNLPGroup, we released a Greek edition of BERT that you can literally use with 2 lines of code via @huggingface transformers for both @PyTorch / @TensorFlow. Thanks TensorFlow Research Cloud (TFRC) for supporting our research. 🤗🇬🇷See details at https://t.co/VHE0chqsLv
The slides of my @eurnlp talk on node embeddings are now at https://t.co/Qp1v412tcj. The video of the talk can be found at https://t.co/sd9wYaRLRR (Session 2, after J. Nivre, B. Webber).
Facebook AI researchers are sharing an all-attention layer to simplify the Transformer model and an adaptive attention span method to make it more efficient. Even with a much simpler architecture, these methods match or improve state-of-the-art results. https://t.co/wCbK4P7Bgv
Dimitris Pappas will present his PhD thesis proposal on Biomedical Question Answering on Friday 22/2, 15:30-17:00, Troias building, room T103. All welcome.