Joachim Wagner
@Wjrgo
Followers
493
Following
23K
Media
15
Statuses
2K
Natural Language Processing #NLPProc, #MachineLearning, High Performance Computing #HPC @ADAPTCentre @DublinCityUni
Dublin, Ireland
Joined October 2014
ADAPT is delighted to announce that Dublin has been selected to host ECAI 2029, one of Europe’s premier artificial intelligence conferences. This puts Ireland firmly on the map as a global AI innovation hub. The successful bid was a collaboration between the Research Ireland
0
1
1
This is a crazy story. People are using non visible embedded prompts within the text of articles with instructions to AI “reviewers.” A real pyramid of fraud.
asia.nikkei.com
Instructions in preprints from 14 universities highlight controversy on AI in peer review
1
22
48
People often claim they know when ChatGPT wrote something, but are they as accurate as they think? Turns out that while general population is unreliable, those who frequently use ChatGPT for writing tasks can spot even "humanized" AI-generated text with near-perfect accuracy 🎯
24
164
1K
I'm shocked to see racism happening in academia again, at the best AI conference @NeurIPSConf. Targeting specific ethnic groups to describe misconduct is inappropriate and unacceptable. @NeurIPSConf must take a stand. We call on Rosalind Picard @MIT @medialab to retract and
133
276
2K
Newly disabled 20-somethings never properly informed by their governments about #LongCovid are confused and angry. "All I was warned about is that I could die if I had comorbidities & [was] old. Had I known it could destroy my life I would have taken more precautions." #COVID19
78
2K
4K
#Hiring: #PhD Position in Children’s online safety and protection of #LGBTQ+ groups with @DCU @AntiBullyingCen & @dcu_lawandtech Apply: https://t.co/E0QGKW6vin Deadline: 31st May 2024
0
7
10
Postdoctoral Researcher in Applied Natural Language Processing - Cilter: Dublin City University - Research - Other https://t.co/aHyyysN0pZ
#universityvacanciesireland
0
1
1
LLMs are now trained >1000x as much language data as a child, so what happens when you train a "BabyLM" on just 100M words? The proceedings of the BabyLM Challenge are now out along with our summary of key findings from 31 submissions: https://t.co/zli0jzA1XP Some highlights 🧵
13
189
1K
Generative AI news! We detect #AI generated text, including #chatgpt, by leveraging emotion. https://t.co/iXTQeRoM9j Research funded by @scienceirel CRT @ml_labs_irl at @dcucomputing @DCU #ScienceWeek With @fosterjendublin and Yvette Graham.
arxiv.org
Recent developments in generative AI have shone a spotlight on high-performance synthetic text generation technologies. The now wide availability and ease of use of such models highlights the...
1
6
8
A new online platform has been launched today by Cilter and its various partners which requests anonymous submissions from parents/guardians of examples of harmful online language experience by children. Further details: https://t.co/yVPFfAZiwS
#kidsonlinesafetyresearch
0
5
5
One of our key sources of human data is no longer fully “human"! We estimate that 33-46% of crowd workers on MTurk used large language models (LLMs) in a text production task - which may increase as ChatGPT and the like become more popular and powerful. https://t.co/SJfKjDM6gX
42
527
2K
Dr Joachim Wagner @Wjrgo @DCU joins the #ADAPTSciCon2023 knowledge sharing session including a discussion on gaBERT, an Irish Language Model #LLMs #ASC2023 #neurallanguagemodels #NLP
0
1
1
Twitter has announced that it is cutting off free access to its API. This is a dramatic restriction that will disrupt critical public interest research projects on everything from disasters and child safety to health and disinformation.
independenttechresearch.org
Twitter's API restrictions threaten essential research, innovation, & knowledge. Stand together to protect the public goods from data access.
1
11
19
For a new blog article, I investigated 20 years of publications about BLEU and identified 12 critical flaws that are often discussed in the scientific literature. https://t.co/IuseL6xPS8
medium.com
Why you shouldn’t trust BLEU according to 37 studies published over 20 years
3
9
60
Interesting thread about good old sota NLP technologies in 2009 that was also hallucinating stuff thought that Jews were causing aids. Look at the comment section in the blog post! Nice sharing this @jasonbaldridge !
With the discussion about LLMs, it’s perhaps instructive to recall the healthBase system of 2009, and how it produced many bizarre response when used outside of its domain (and problems in domain too). This Language Log post is a good entry to the story:
0
1
2
Be very careful whom you spend time with without mitigations this winter, because a lot of people now think they need to get sick to be well, and they need to get you sick to help you.
15
248
1K
Congratulations on successfully defending your PhD thesis today @jamesarbarry
1
1
14
I refuse to be delegated to the "skeptics box" in someone else's framing of a debate. Here is my response to @stevenbjohnson 's NYT Magazine article about LLMs and OpenAI. On NYT Magazine on AI: Resist the Urge to be Impressed
link.medium.com
Context
36
394
1K