Berlin
Joined September 2019
- Tweets144
- Following214
- Followers195
- Likes969
Pinned Tweet
Our work (w/ @HinrichSchuetze and @4digitaldignity) collecting #HateSpeech data in collaboration with affected communities around the globe has been accepted to #ACL2022 Findings. [1/3]
⚙️ bit.ly/xtremecode
📝 bit.ly/xtremepaper
Antonis Maronikolakis retweeted
Good news :) We are honored to have been awarded the International Communication Association outstanding article award. Many thanks everyone, esp partnering factcheckers! w/ @antmarakis Axel Wisiorek @icahdq @BigDataSoc journals.sagepub.com/doi/ful…
Greece is selling off more of our largest airport and calling it a success, business as usual. A bunch of numbers on short-sighted metrics will go up and everybody will think Greece is now a healthy economy, history repeating itself.
reuters.com/markets/deals/gr…
Antonis Maronikolakis retweeted
What are the limits of AI in content moderation?
Check out this new article from Sahana Udupa, @antmarakis, and Axel Wisiorek where they propose the concept of "ethical scaling" to "highlight highlight moderation process as political praxis"
buff.ly/3O52Rr8
Antonis Maronikolakis retweeted
New article! 🚨 ⚡ "Ethical scaling for content moderation: Extreme speech and the (in)significance of #ArtificialIntelligence" by Sahana Udupa, @antmarakis, and Axel Wisiorek.
#AI, #extremespeech, #ethicalscaling, #decoloniality. Check it out! ▶️ buff.ly/3O17uCJ
Antonis Maronikolakis retweeted
I have submitted the text of my PhD thesis and now it is available online🎉
Also, if you want to listen about fake news and texts detoxification, you are very welcome to join translation🤗
Official web-page with all info:
skoltech.ru/en/education/phd…
Antonis Maronikolakis retweeted
🎉 Prof. Hinrich Schütze has been awarded an ERC Proof of Concept Grant for the Respond2Hate project!
Respond2Hate allows users the agency to remove hate speech from their social media feeds, with a focus on currently underrepresented countries and groups. Details 👇
Congratulations to medical researcher Sebastian Kobold and computational linguist Hinrich Schütze! The LMU researchers have obtained two #ProofOfConcept Grants from @erc_research.
lmu.de/en/newsroom/news-over…
Antonis Maronikolakis retweeted
Congratulations to medical researcher Sebastian Kobold and computational linguist Hinrich Schütze! The LMU researchers have obtained two #ProofOfConcept Grants from @erc_research.
lmu.de/en/newsroom/news-over…
(Very) Late to the paper announcement party, but super proud of my students Shuzhou Yuan and Philip Baader for publishing two #HateSpeech #NLProc papers @WOAHWorkshop and @genderbiasnlp. 🥳🥳
Shuzhou worked on reducing the rate of hatespeech false negatives across multiple languages: aclanthology.org/2022.woah-1…
Philip analysed bias along intersectional (gender and race) axes in hate speech datasets and models: aclanthology.org/2022.gebnlp…
Antonis Maronikolakis retweeted
I will be hiring two PhD students in a couple of months. 1⃣ Any tips on how to do it right? 🤔🤷♂️ 2⃣ If you'd like to work with me, let me know. 👋👩🎓👨🎓
Antonis Maronikolakis retweeted
How do we address the shortcomings of AI in tackling the ambiguities and complexities of annotating hateful language online? Fall '21 @ShorensteinCtr Fellow @4digitaldignity & her co-authors emphasize the need for "ethical scaling" in new paper: shorensteincenter.org/ethica…
Worse thing than @DublinAirport is the @lufthansa app. My GOD what an abomination. I started ranting about its myriads of issues, but it would take multiple tweets and I can't be bothered. Maybe it's high time Lufthansa joined the rest of the tech industry and hired programmers?
Antonis Maronikolakis retweeted
Analyzing Hate Speech Data along Racial, Gender and Intersectional Axes
deepai.org/publication/analy…
by @antmarakis et al.
#Computation #Language
Antonis Maronikolakis retweeted
Excited to announce our new paper "CaMEL: Case Marker Extraction without Labels 🐫", with @vjhofmann, @_masoudjalili, and @HinrichSchuetze was accepted to #ACL2022!
arxiv.org/abs/2203.10010 (1/5)
Our multidisciplinary team curated a hate speech dataset from Brazil, Germany, India and Kenya, collaborating with members from these countries with vested interested in social justice and familiarity with the affected communities for collection of more representative data. [2/3]
I played around a bit with @MilaNLProc's Twitter Demographer, and I just wish I knew about it earlier. Not only can we get geolocation of tweets (sooo many headaches in the past from this), but also other variables such as age.
Pretty neat, check it out! github.com/MilaNLProc/twitte…
Antonis Maronikolakis retweeted
Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments
deepai.org/publication/liste…
by @antmarakis et al.
#Computation #Language