@skdh@dfg_public This is a shame to hear, and I hear this all too often with academia :( Will definitely grab a copy of your new book to support! Also a huge fan of your neutral and non-hyped educational videos on Youtube. Really a great service to the community
@Plinz@livcomp This is so refreshing to hear. I've long wondered why we've no ritual in the tech world to declare a software product *complete*, say when it reaches an inflection point around dev investment vs output usefulness - and subsequently allow devs to move on to create smth new
@livcomp When I was young, I thought that we would develop all software to asymptotically approach their optimum, with updates getting less and less frequent. When did we lose our way
A whole bunch of people are gonna get rich with this LLM gold rush and leave the mess for everyone else to clean up. So much proliferation, so many random startups, zero guardrails, regulation, basic engineering principles of extensive testing or evaluation.
I thought it might be fun to do a quick distraction project on the mathematically optimal way to play wordle.
Wrote a little program yesterday which uses a bank of ~13,000 words along with their frequencies in English, tried it today. Not bad!
🟨🟨⬜🟨⬜
⬜⬜🟨🟨🟩
🟩🟩🟩🟩🟩
As of today, 2 out of 3 political institutions of the EU which hold executive and legislative power are led by women.
This is a good day for Europe, for equal representation and a powerful sign to the world.
#Metsola#vonderLeyen
Just as with Alexa, the issue of context-loss when serving QA content from webpages as answers to questions is real. I treat QA responses from such ML systems as simply guidance to a potential source doc, and not as the truth itself.
@IAugenstein What are your thoughts on the reasons for this oversupply? Do you think it is driven by hyper-interest in certain fields (esp. AI) or perhaps by an increasing industry demand for research scientists? I.e. people going into PhD programmes whose end goal is industry not academia.
@Zemploid Honestly no matter how badly I sometimes miss Malta, and contemplate moving back, visiting home always destroys those day dreams fast. The Malta I love and remember is no longer recognisable to me. I have been gone for 10 years though (based in Germany)
R E M I N D E R!
1st Social ComQuant Summer School will take place online from 26-31 July 2021 &is open to anyone interested in learning about methods of Computational Social Science. The summer school comes with no costs. Don't forget to apply before May 14!
#CSS#socialcomquant
@JosephineLukito I once tutored a student who would take one look at a math problem and say “I have no idea what to do here”. So I’d tell him “Ok, so just take the wildest, dumbest guess you can and take it from there”. His self determined „dumb“ guesses were usually 100% correct.
Helping fact checkers to fight misinformation: our COSMOS work automatically detects out-of-context image use!
@shivangi2201 just released the dataset of 200K images & 450K textual captions!
Video https://t.co/JnZRC2zxyT
Data https://t.co/SxiO0IZsDp
@shivangi2201@chrisbregler
Check out our workshop on Misinformation at #thewebconf and #www2021 tomorrow! (15 April)
Great line up of invited speakers and contributed talks
https://t.co/o4RgamUzNr
New preprint in collab. w. @mvitevit and @Fede_Botta:
https://t.co/bdPS8A5dVn
We investigate language, emotions and pics of popular tweets about #COVID vaccines in English or Italian between 12/2020 and 3/2021.
Social discourse was mostly relative to vaccine distribution but..
GESIS Fall Seminar in Computational Social Science 2021 – Registration is open!
We are excited to announce the program of the #GESISfallseminar held virtually from 13 September to 01 October 2021.
More info and registration https://t.co/8R0hGricyi