Lucie Flek
@lucie_nlp
#NLProc researcher, computer science prof @UniBonn Nothing new to see here in 2025. You will find my news via https://www.linkedin.com/in/flekova/ #eXit
Excited about our new Nature @NatureHumBehav paper, exploring societal impacts of #LLMs, where we argue that we should be thinking about those impacts through the lens of collective intelligence. nature.com/articles/s4156…
If you're not sure what to do with your #ARR #EMNLP or #NeurIPS reviews, how about submitting your revised version directly to #INLG2025? Deadline this Friday, July 18th. #NLProc
The first papers are already submitted, but for anyone hoping for an extra week to work on their paper, I have very good news! Deadline for #INLG2025 has been extended until July 18th (midnight, AoE). Submit your most exciting work on generation! 2025.inlgmeeting.org
Thank you, @hyperindexed ! For those who are physically or virtually enjoying #Vancouver: Vahid will be presenting this on the #NeurIPS "System 2 Reasoning at Scale" this week, pass by to say hi! Blog post: lamarr-institute.org/blog/sot-llm-r… Joint ongoing work with @JTrippas & @lal_yash
Can narrative techniques like analogical reasoning and progressive disclosure help LLMs produce better answers? Cool paper by @lucie_nlp and team, detailing "Story of Thought"
Teuken-7B: OpenGPT-X veröffentlicht Open-Source-KI-Sprachmodell. Das KI-Sprachmodell Teuken-7B wurde mit allen 24 europäischen Amtssprachen trainiert. glm.io/191150?t
OpenAI’s biggest competitor just gave AI the ability to control computers We gave 250+ hackers 24 hours to see what it’s capable of Here’s what we saw at the Nexgen Computer Use Agents Hackathon w/ @AnthropicAI + @AgentOpsAI @notablecap at @AGIHouseSF (🧵):
“A female professor on home office in a flu season” must be what Fritz Zuber-Buhler had in mind as a title. #AcademicChatter
I’m not sure anymore if it‘s GPT or if LLM researchers simply added the delving into an intricate tapestry to their daily vocabulary by now. #ICLR
RLHF aligns LMs to a fixed weighted combinations of rewards, but cannot retrain a new model for each user's preferred weightings. To achieve customization, prior work merges the parameters of single-objective models. Can we do better? We show that merging logits is better!
At the AI All Hands Meeting, @CaisaLab 's Vahid presented the perspective taking and identity guessing demo game FanToMAS . Scan the QR code with your phone and try if you can beat the leaderboard! aivalify.com/fantomas
For the next two days, the All Hands Meeting of the German AI community takes place in Dresden. This morning, the AI centers @bifoldberlin, @DFKI, @LamarrInstitute, @MunichCenterML, @Sca_DS and Tübingen AI center presented their research highlights.
With @CaisaLab and the @LamarrInstitute , we're at the All Hands Meeting of the AI competence centers in Dresden, looking forward to discussing our research! (More on this paper here: caisa-lab.github.io/2024/08/15/lam…) #NLProc #LLM
For the next two days, the All Hands Meeting of the German AI community takes place in Dresden. This morning, the AI centers @bifoldberlin, @DFKI, @LamarrInstitute, @MunichCenterML, @Sca_DS and Tübingen AI center presented their research highlights.
Spanende Diskussion zu Forschungsdatenstrukturen an der #BMBF . “Wir müssen die Sicherheit schaffen, dass es weiter geht. Und zwar nicht mit kleinen Insellösungen.“
So happy and proud of the first two #PhD birds of @CaisaLab that left our nest this summer with flying colors! @JoanPlepi on Dynamic User Representations and Personalization in #NLP, @AllieLahnala on Operationalizing Empathic and Supportive Communication in #NLP. Well #PhDone !!
