@mixlamalice Ha ha, j’ai acheté des tickets pour Paris au départ d’une gare zone 5 et je ne peux pas les valider au départ de la gare suivante sur la ligne qui est zone 4...
Français
Aurélie N.
377 posts
Aurélie N.
@AlmostAisling
Researcher in biomedical natural language processing, tea lover and occasional potter
France Katılım Şubat 2014
117 Takip Edilen100 Takipçiler
When I first attended AMIA and learned about ACMI, it seemed like a distant, near-impossible goal. Now, I'm just reflective on how much great mentoring I've received that got me to this place. This is an amazing community.
amia.org/news-publicati…
English
@squig @leeflang_m "If you try to get adequate permission to use your data, then the project/PhD student contract is over before you can even access the data. "
English
Looking for some "bad" but relatively common anti-open science takes for a class.
Things like:
If you preregister you are never allowed to explore your data / deviate from it.
Requiring appropriate power is unfair to people with lower research budgets
etc.
English
@gg42554 @UndefBehavior @ReviewAcl Experience from the biomedical domain shows that this effectively reduces the reviewer pool by about 30%. Do we need less reviewers? Do we need this type of reviewer self-selection?
English
Writing the EMNLP rebuttals. I'm now convinced (also after having served for a year as EiC for @ReviewAcl) that nothing short of publicly releasing reviews *with reviewer identities* will substantially improve the (currently appaling) average review quality in #NLProc.
English
@UndefBehavior Well arguably, it should be part of the PhD training. A first reviewing experience should be a co-review (with credit) with a more experienced colleague who can take the time to discuss both the paper and the reviewing process.
English
This. No PhD student should review.
Eval4NLP@eval4nlp
@gg42554 @ReviewAcl The reviewing quality is bad in NLP, agreed. But it's also often because there are so many junior people (who sometimes need to step in because there aren't enough reviewers). Exposing them publicly may also be problematic for various reasons.
English
Aurélie N. retweetledi
Tweets are great material for #NLProc #research; however it is essential to engage tweet authors and seek consent for use in studies as explained by Dr. @cfiesler
journals.sagepub.com/doi/full/10.11…
English
@keepha_research Great! I hope we get started with the project soon.
English
@AlmostAisling Me! Here it is, our new account for the KEEPHA study!
English
@keepha_research congrats on your new account, I'm so glad you joined twitter today! 😀
English
I also loved the idea that any research talk in NLP should be prefaced by a few words in the speakers’ native language - or language of choice. Thanks @cigilt !
English
That was a wonderful keynote on so many levels: great work on language resource development and NLP for low resource languages but also insightful analysis and comments on ethics and policy.
Dr. Teresa Lynn@cigilt
Enjoying a great welcome at #TALN2022 in Avignon as a keynote speaker. Especially great to meet so many French NLP researchers working on low-resourced languages 🙌🏻 #nlproc
English
@Gwen_psd @gblanc05 Le linguiste Christophe Benzitoun explique que la langue évoluant, il y a mécaniquement une plus grande distance entre français oral et écrit aujourd'hui qu'hier. Apprendre l'orthographe est donc plus difficile pour les enfants d'aujourd'hui youtu.be/CZiJqA8foss?t=…
YouTube
Français
Aurélie N. retweetledi
Our new study shows that data availability statements are not very useful; 1670 (93%) authors who indicated that data are available on request either did not respond or declined to share their data with us. Journal of Clinical Epidemiology: doi.org/10.1016/j.jcli…
English
Aurélie N. retweetledi
What is it with organizers and presenters not coming to their workshops and their regular/invited talks, i've Never seen anything like that.
Do people not want to discuss their work or is it not worth enough for the handful in the audience?
English
Aurélie N. retweetledi
I am happy to announce that our position paper "You Reap What You Sow: On the Challenges of Bias Evaluation Under Multi-Lingual Settings" has been accepted for presentation at the @BigscienceW #acl2022 ☘️ workshop! 🧵⬇️ aclanthology.org/2022.bigscienc…
English
Aurélie N. retweetledi
@MeganPrictor The Practice Fusion case shines a spotlight on the potential for clinical decision support systems to be vehicles for fraud, leading to widespread harm. The tools for uncovering such fraud and harm in Australia are less robust than in the US link.springer.com/article/10.100…
English
Editors: please let reviewers know this is not acceptable peer review content!
Aurélie N.@AlmostAisling
Reviewers: please provide actual constructive advice on how to improve the manuscript. Do not speculate on authors' seniority or qualification and the alleged merit of your criticism.
English