Oana-Maria Camburu
@oanacamb
Followers
2K
Following
2K
Statuses
1K
Leverhulme Early Career Fellow, UCL. XAI, NLP. Previously: PhD and Postdoc at University of Oxford, Intern at Google, Undergrad+MSc at École Polytechnique Paris
Joined December 2016
Such a wonderful surprise to see this tweet, thank you @anmarasovic! Maybe explain-then-predict wasn't such a catchy name but it is the same as Chain-of-Thought (CoT). Your work has also done so much for this line of research, @anmarasovic! Keep it coming!
Rebranding is an effective way to overlook those who drove the original ideas into the mainstream. Oana-Maria Camburu @oanacamb who popularized free-text explanations with e-SNLI should have her work cited in all your prompting papers that generate their reasoning + prediction.
0
2
30
🚨Check out our ✨iLLuMinate✨ framework accepted at #AAAI2025, where we use LLMs to communicate explanations in education 🧑🏫
How should LLMs integrate with eXplainable AI models? We aim to answer this question with our new @RealAAAI 2025 paper, iLLuMinaTE! 🌟 We position LLMs as a communication layer for XAI, encoding social science theories of explanation towards actionable student feedback. 1/
1
3
15
RT @PMinervini: Check out this piece on Strawberry 🍓/o1 we just authored on TheConversation! With the amazing @PontiEdoardo and @FelineAuto…
0
5
0
RT @PedramHosseini: 🧵(7/N) Adding to the thread, one of the outstanding papers from #EMNLP2024: Fool Me Once? Contrasting Textual and Visua…
0
1
0
RT @max_nlp: Incredibly honoured to see our work recognised as an outstanding paper. @magikarp_tokens dove deep into the dark depths of tok…
0
15
0
🥳🎉Incredibly happy that our paper received an ✨Outstanding Paper Award✨ at #EMNLP2024! Huge congrats particularly to @maximek3, who did an incredible job! So lucky to have you as my student! Check out our paper below ⬇️
🚨🌟Very excited to share our #EMNLP2024 work on XAI for healthcare 🩺 We did a large user study with 💫85 medics💫 to find out how natural language explanations and saliency maps influence clinical decision-making Check out the details below ⬇️ #XAI #Medical #UserStudy
2
6
86
RT @atanasovapepa: 🚨 @IAugenstein and I are offering a fully funded PhD Fellowship in Interpretable NLP at @UCPH_Research (@DIKU_Institut)…
0
39
0
RT @_joestacey_: Here's a little blog post describing our EMNLP 2022 and 2024 papers about atomic inference, sharing some thoughts about th…
0
10
0
RT @_kire_kara_: 👋Psst! Want more faithful, verifiable and robust #LLM reasoning than with CoT, but using external solvers is meh? Our FLAR…
0
10
0
🚨🌟Very excited to share our #EMNLP2024 work on XAI for healthcare 🩺 We did a large user study with 💫85 medics💫 to find out how natural language explanations and saliency maps influence clinical decision-making Check out the details below ⬇️ #XAI #Medical #UserStudy
Are self-rationalizing VLMs helpful in a real-world, safety-critical setting? In our new #EMNLP2024 paper 🩺 we recruited 85 medics 🩻 and made them each collaborate with an AI on 80 X-rays under 4 different XAI scenarios It turns out that: 🧵 (1/n)
0
2
25
🚨The camera-ready of one of our #EMNLP2024 papers is out ⬇️
Here is a little 🧵 about our new #EMNLP2024 paper! In this paper we use LLM-generated facts as atoms in atomic inference systems (for NLI). So first of all, what is atomic inference, and why should you care about it?
1
5
39
✨2/2 papers accepted at #EMNLP2024✨ Happy to have such great collaborators! 1) Do Explanations Help or Hurt? Saliency Maps vs NLEs in a Clinical Decision-Support Setting: Arxiv soon! 2) Atomic Inference for NLI with Generated Facts as Atoms:
2
8
56
RT @PMinervini: My amazing collaborators will be presenting 3/3 papers at EMNLP 2024!🚀🚀🚀 @emnlpmeeting Crazy simple KV Cache Compression: h…
0
12
0
Check out our work at #ACL2024 on improving the robustness of SOTA LLMs using natural language explanations ⬇️
Excited about using synthetic natural language explanations to boost ICL's robustness against adversarial samples? Join us at our poster session at 10:30 AM (14/08)! #ACL2024 #ICL #AdversarialRobustness
1
3
19
3) SparseFit: Few-shot Prompting with Sparse Fine-tuning for Jointly Generating Predictions and Natural Language Explanations
0
0
2
RT @PMinervini: Jesus Solano (now a PhD student at @ETH working on vision and security) presenting SparseFit, a joint project with @oanacam…
0
6
0