imuhammad @imuhammad@mastodon.utwente.nl
The boy who drank stars.
Joined Oct 2018
imuhammad
boosted
Say hi to RobBERT, a Dutch language model based on RoBERTa with some tasks specific to Dutch.
Blog & paper:
https://people.cs.kuleuven.be/~pieter.delobelle/robbert/
imuhammad
boosted
Slides for all lectures at the #Siks #IR course:
https://www.informagus.nl/events/siks-ir-2019.html
Here, you also find the information for the group experiment of this afternoon.
imuhammad
boosted
Looking for a search engine that does not know everything about you...
imuhammad
boosted
imuhammad
boosted
An AI Wizard of Words
by @wftl
A look at using OpenAI's Generative Pretrained Transformer 2 (GPT-2) to generate text.
https://www.linuxjournal.com/content/ai-wizard-words
#python #AI #HowTo
A database of Computer Science arXiv data collected by Matthew Kenney
imuhammad
boosted
now has an option for a single column interface! Go to "Preferences" -> "Appearance" and untick "Enable advanced web interface".
This will be the default interface for new users.
**Fair is Better than Sensational:
Man is to Doctor as Woman is to Doctor**
https://arxiv.org/pdf/1905.09866.pdf
A recent paper by some researchers at University of Groningen that shows some analogies like "a man is to computer programmer as woman is to homemaker" are not the most accurate way to show the bias that exists in the data.
imuhammad
boosted
[ICYMI on twitter] The Deep Learning track at #TREC2019 will feature two tasks—a DOCUMENT RANKING task and a PASSAGE RANKING task—both with large training datasets.
Approximate dataset stats: 367K training queries + 3.2M documents for the document ranking task; and 500K training queries + 8.8M passages for the passage ranking task.
Official guidelines: https://microsoft.github.io/TREC-2019-Deep-Learning/
Coordinators: Nick Craswell, Bhaskar Mitra, Emine Yilmaz, Daniel Campos.
Please share. #TRECDL2019
I used "trendyy", an interesting R package for querying Google Trends, to obtain and visualize the search popularity of Dutch universities over the past few years.
The results were not exactly what I expected to be!🤔
imuhammad
boosted
This flawed decision has to be overturned.
Science is not politics.
IEEE, a major science publisher, bans Huawei scientists from reviewing papers | Science | AAAS
https://www.sciencemag.org/news/2019/05/ieee-major-science-publisher-bans-huawei-scientists-reviewing-papers
Scientists worldwide should work together.
About the politics:
Note that we HAVE serious doubts about so many things in CS. CISCO and Intel "bugs" could be by design, or be widely known in Secret Services, who perhaps co-fund Google, etc etc. NIST and NSA story on encryption. Research co-funded by DARPA.
Full Stack Deep Learning bootcamp 2019 lectures and other resources :
https://fullstackdeeplearning.com/march2019
imuhammad
boosted
Visualization of the world of search engines:
imuhammad
boosted
Double standards:
Huawei - flaws indicate spying.
Intel - flaws indicate, well, flaws.
Who will be awarded as best Dutch Data Science & AI expert of 2019?
How to use https://mybinder.org to make the jupyter notebooks in your github reproducible .
imuhammad
boosted
Academics hide humans from #AI surveillance cameras with 2D prints
https://www.zdnet.com/article/academics-hide-humans-from-surveillance-cameras-with-2d-prints/
The boy who drank stars.
Joined Oct 2018
