How to do transfer learning with limited data: we show which techniques and embeddings work well for the case of predicting urgency status of crisis tweets for low resource languages:
Paper: https://t.co/JHPFNxvONr
Repo: https://t.co/z8eXRcUt0D
#coling2020
Check out how to improve production parsers with cross-lingual synthetic data: Scalable Cross-lingual Treebank Synthesis for Improved Production Dependency Parsers
https://t.co/BjGSMBSWje Work done during my AI Residency at IBM Research #coling2020#ibm
In "Multi-Stage Pre-training for Low-Resource Domain Adaptation" paper, we use inherent structure in text to formulate synthetic pre-training tasks that transfer to downstream tasks: QA, document ranking & duplicate question detection #EMNLP2020 https://t.co/7imVt9TAtC
Findings of EMNLP paper "Answer Span Correction in Machine Reading Comprehension", we demonstrate an approach to correct partial match answers into exact match by training an explicit corrector model with improvements in mono&multilingual eval https://t.co/KbRPtrVYKa #EMNLP2020
Iyi ki Sehir Plancısı Olmuşum diyen Doç. Dr. Pelin Sarıoğlu Erdoğdu bu haftaki konuğumuz.
Kendisi ile konut araştırmaları ve temel tasarım üzerine keyifli bir söyleşi gerceklestirecegiz.
#konut#gayrimenkul#sehirplancisi#ogretimgorevlisi
Katilim Linki:
https://t.co/E3cSzsIY6y
The final installment of field notes on pandemic teaching reflects on the possibilities of building communities online, working across differences, and how to adapt when sufficient bandwidth — or even internet access at all — isn't readily available. https://t.co/1Lu6xeIaA2
This month, the US stock market lost two trillion dollars in market-cap due to a problem that science R&D could have preemptively solved.
That figure exceeds the sum of all funds ever received over all years since 1950 by the National Science Foundation, the NIH, and the CDC.
with @zoom_us 's new video background feature, i can stay engaged and nodding thoughtfully, even when i'm not even in the room -- this is the kind of innovation we need!
IBM Research AI launched a new question answering leaderboard for domain adaptation for enterprise QA systems.
https://t.co/5i1VcCPxYS #ml#questionanswering#TechQA
In case you'd like to take a look at the work from @geoffreyhinton while he was working on #backpropagation while at #scsatcmu
Plaut, D., Nowlan, S. and Hinton, G. E.
Experiments on learning by back-propagation
Technical Report CMU-CS-86-126 - https://t.co/SU97Neg0vC #ml#ai