I've had enough.
Colleagues posting their successes here, discussing cool findings, blogging events: thank you so much!
I'll miss your thoughts and reading suggestions dearly.
But, it's been too long that I supported this platform. There was no real alternative, but...
1/3
If not: see you in real life, somewhere, some time.
Staying to participate here would legitimise the platform, keep it sustained - I no longer want to do that.
I'll end my presence here with a small experiment.
So long, and thanks for all the fish!!
#riptwitter
3/3
I've had enough.
Colleagues posting their successes here, discussing cool findings, blogging events: thank you so much!
I'll miss your thoughts and reading suggestions dearly.
But, it's been too long that I supported this platform. There was no real alternative, but...
1/3
This has changed!
Maybe I can convince more of you to start social presence on Mastodon, e.g. on my own https://t.co/RmnBlaiGoJ instance (IR), at https://t.co/y3LzwQBkjF (ML/AI) or https://t.co/cjdrCkdQjb (Dutch researchers), or any other instance. If so, drop me a note!
2/3
We finally finished writing the description of ClueWeb22, the corpus with 10 billion web pages + rich visual information: https://t.co/1f6FBI9sOJ. I hope you all find the description in the paper informative.
@XiongChenyan@metzlerd@CharlotteHase Courageous to say that out loud. I agree with you Chenyan!
When you put out results that you later learn that are incorrect, then you ought to retract the paper; so others do not continue to build on the wrong results. Award or no award.
@juansequeda@erikapullum@EmilyRiederer tnx, I read the 70s paper but of course IBM had to build a real product and find some way to squeeze the real world's irregularities into their approach :-)
Job opportunity. Prof Bart Jacobs & I are looking for a Postdoctoral Researcher: Use of Bulk Data by Intelligence Services and Other State Bodies
https://t.co/gm1Xe2EQ1p
I am excited to announce that this year's ICT with Industry workshop features 6 cases from KB, VeiligheidNL, Textmetrics, RTL and Axini.
PhD students, this is your chance to delve into real-world #AI problems and make an impact.
Registration dl: Dec 4
@NWO_Science@lorentzcenter
Overall, itโs clear that Transformers can be trained to encode far-from-trivial learning algorithms, but *how* they encode such algorithms is still a wide open question.
See our paper for additional experiments and many more open questions [https://t.co/3aQ0XWWPV9]!