I'm a linguist, interested in corpus-computational studies, NLG (mostly REG) and Reference (not the bibliographical one of course! :-D)

Cologne, Germany
Joined November 2020
Super happy to share that my dissertation "Referring expression generation in context: combining linguistic and computational approaches" (2022) has now been published as a book by @LangSciPress 😃🥳 It's open access, downloadable from: langsci-press.org/catalog/bo… (1/2)
Just published "Referring expression generation in context: Combining linguistic and computational approaches" by Fahime Same #openaccess #tgdi langsci-press.org/catalog/bo…
5
5
1
28
2,793
It offers an in-depth analysis of the production/generation of referring expressions in context and explores various aspects of the task including the choice of corpus, method (rule-based, feature-based, neural) and evaluation approach (automatic and human). (2/2)
4
99
#CallForPapers for #INLG2024 (23-27 Sept) If you work on #NaturalLanguageGeneration, #TextGeneration (with or w/o #LLMs) Deadlines are UTC-12 • regular papers: May 31 • ARR commitment papers: June 24 • demo papers: June 24 • Notification: July 15 inlg2024.github.io/calls.htm…
1
9
15
1,721
📢📰Interested in naturalistic language? Check out our corpora from German novels for referring expressions in natural language: "Multi-layered Annotation of Conversation-like Narratives in German" ➡️rb.gy/9gcia This is joint work with @pb_schumacher @fafa_same
2
14
1,090
Fafa Same retweeted
were these the offices of the scott paper company? #ACL2023NLP
3
6
1,288
Attending #ACL2023NLP virtually while having my dinner. Who needs #Netflix when you've got inspiring keynote talks to watch?
12
302
Yes! Totally! Absolutely! And my FOMO's getting stronger & stronger. #ACL2023NLP
This quoted post is unavailable.
5
279
Many congratulations to Maria @maria_bardaji on winning the Wilhelm von Humboldt best dissertation prize! You 'totoli' deserved it! 😍😍 #DGfS2023
1
20
2,877
Happy to share that Max (Hörl) and I hold an online workshop on "Data Transformation and Wrangling in R" on Tuesday, Feb 14, 14-16, as part of the International @LoveDataWeek. Registration is free, and open to everyone. #LoveData23 #FDM Link to register: uni-koeln.zoom.us/meeting/re…
1
9
13
1,445
The #DAAD supports #refugee researchers from #Iran in starting and continuing their #research projects with the programme "Bridge Scholarship Iran". Applications are open until 28 February. More Information 👉 lnk.tu-bs.de/semLQu @DAAD_Germany
29
5
52
23,681
Job notification💡Project INF of the #SFB1252 is looking for a WMA who is interested in statistical methods in linguistics. Find more information (in German) here: sfb1252.uni-koeln.de/sites/s… and apply by December 20. 🗓️ @dch_cologne @LinguistikKoeln @IdhTweets
16
10
Excited to share that I successfully defended my PhD yesterday & as the picture shows, couldn't be happier! 😁😁
And yet another dissertation defense in the #sfb1252: Congrats @fafa_same and supervisor Nikolaus Himmelmann.
12
1
47
Can we vote to have a musical interlude at every future LREC!! Thank you @StevenBird for your classic guitar performance 👏🏻👏🏻 #lrec2022 @lrec2022
2
7
2
61
Add-on: Für den Jahresbericht 2021 haben wir mit @fafa_same und Maximilian Hörl vom INF-Projekt @SFB1252 @Philfak_koeln @UniCologne über unsere Zusammenarbeit, #FDM im Projekt, Perfektionismus und pragmatische Lösungen gesprochen – nachlesbar ab S. 26: dch.phil-fak.uni-koeln.de/ue….
3
8
Briefly, we did a systematic evaluation of different approaches (rule-based, ML feature-based, and E2E), using the task of generating referring expressions in context as an example. We examined 2 different datasets (WEBNLG & WSJ) and performed both automatic & human evaluations.
1
1
Interestingly, our classic non-neural models achieved on-par or better performance than our E2E REG systems. We hope that our findings can encourage future researchers to also consider well-designed non-neural model architectures. 😃
2