Associate Professor of Computer Science at Heriot-Watt University & Head of ML and Co-founder at Alana @HeriotWattUni. Researching #NLG #NLProc #NLP
Edinburgh, Scotland
Joined February 2009
- Tweets107
- Following395
- Followers362
- Likes489
There is still time to submit your work on our Information Disorder workshop taking place at LREC 2026!
📢 Deadline extended (24 Feb 2026) – Information Disorder Workshop (InDor)
🧠 InDor aims to advance interdisciplinary NLP research on Information Disorder, with a strong focus on theory, datasets, evaluation, cultural context and real-world impact.
🔗 information-disorder-worksho…
Looking for a cool place to work on NLP, Generative AI, Robotics, ML (among other things)? Join us as an Assistant Professor at HWU; there is still time to apply! Deadline extended to 11/5/2025. linkedin.com/jobs/view/42058…
Yannis Konstas retweeted
So very, very proud to share our new paper “Voices in a Crowd: Searching for Clusters of Unique Perspectives” (arXiv:2407.14259), a novel framework on how to organically find clusters of unique voices (perspectives) in datasets. 🧵 for summary, co-authors @amitkparekh_ @sinantie
Do multimodal models always follow our instructions (in a robotic manipulation environment)? Well... Check out the latest work by @amitkparekh_ @NikVits @ale_suglia !
What if your multimodal model succeeds when given nonsensical instructions? What if it even succeeds without being told explicitly what to do? Have we finally achieved AGI?
We answer at least two of those questions w/ @NikVits @ale_suglia @sinantie (🧵for more)
Want to find out more about egocentric video understanding? Check out AlanaVLM!
Excited to announce that the Embodied AI team at @helloalana developed AlanaVLM, a new foundation model for egocentric video understanding. Best open-weight model which rivals GPT-4V on spatial reasoning questions in OpenEQA.
Thread below!
arxiv.org/abs/2406.13807
#AI #NLProc
Yannis Konstas retweeted
📢WE ARE BACK!💥
It's official! The Athens Natural Language Processing Summer School @AthensNlp is taking place in 2024 in Athens Greece 🇬🇷!
📅Save the dates 19 - 25 September 2024.
Use #AthNLP2024 for updates!
I am really excited to be giving a TEDx Youth talk on online Gender-Based Violence and AI/NLP. Thanks a lot, #TEDxYouthStGeorgesEdinburgh for hosting. Special thanks to @GerrardPoppy and @NikVits for the discussions to prepare the talk! #VoicesOfTheFuture #tedxstge.
🌟📣 Announcing our #TEDxYouthStGeorgesEdinburgh #VoicesOfTheFuture 2024 speaker lineup! We're excited to welcome Yannis Konstas to the #TEDxYouth stage.💻Follow TEDx Youth St Georges Edinburgh on Facebook for more updates 👉👉 ow.ly/xRqC50QOrfI #IdeasWorthSpreading #TEDx
Yannis Konstas retweeted
Just under 24 hours to go for the @SafeConvAI deadline! 🚨 Collocated wth @LRECCOLING
@t_dinkar
@peppeatta
@sinantie
@verena_rieser
@dirk_hovy
Full details here:
New postdoc position! Are you interested in online Gender-Based Violence identification and counterspeech generation? If so, come and work with us at @iLab_hwu @HeriotWattUni! The deadline for applications is 26 February. #NLProc enzj.fa.em3.oraclecloud.com/…
This quoted post is unavailable.
Yannis Konstas retweeted
Our survey paper is now available on @JAIR_Editor: jair.org/index.php/jair/arti…
Many thanks to my co-authors @oliverlemon and @sinantie for supporting me in this great effort!
Looking forward to hearing your feedback, especially on the future research directions section!
Another great news today: our survey paper "Visually Grounded Language Learning: a review of language games, datasets, tasks and models" was accepted by JAIR. This was part of my PhD thesis completed under the supervision of @oliverlemon and @sinantie!
Yannis Konstas retweeted
New Paper 🎉: arxiv.org/pdf/2401.01814.pdf
Can language models relearn removed concepts?
Model editing aims to eliminate unwanted concepts through neuron pruning. LLMs demonstrate a remarkable capacity to adapt and regain conceptual representations which have been removed
🧵1/8
Yannis Konstas retweeted
Welcome to iLab Wrapped!!🎉🎉
It's the end of 2023, so let's take a dive into some of the Interaction Lab's papers.
Please follow your favourite authors so you can keep up with their work beyond their time at @iLab_hwu
(apologies if the thread updates slowly - @Addlesee_AI)
Yannis Konstas retweeted
[1/N] Happy to share that our paper “Multitask Multimodal Prompted Training for Interactive Embodied Task Completion” has been accepted at #EMNLP2023! 🎉 @MNikandrou @amitkparekh_ @hsbhathiya @ale_suglia
Yannis Konstas retweeted
🎉 New paper to appear at ACL 2023 : arxiv.org/abs/2305.17553
Large Language Models (LLMs) are powerful tools, but they can memorize false or outdated associations. Model editing techniques promise to solve this, but do they really work?
1/
Yannis Konstas retweeted
New paper: arxiv.org/abs/2305.15507
LLMs are good at programing tasks, and they are now being widely used to assist code generation.
But do they actually understand the semantics of programming languages, or do they rely on superficial, "shortcut" correlations?
Let's find out!
1/
"The Dangers of trusting Stochastic Parrots: Faithfulness and Trust in Open-domain Conversational Question Answering" was accepted at findings of #ACL2023!
Joint work with Sabrina Chiesurin,@dmkpls,Marco Antonio Sobrevilla Cabezudo,@arash_eshghi, @g_papaioannou, @verena_rieser
We use open-domain question answering systems as our test-bed for task based dialog generation and compare several models, with a focus on highlighting the danger of systems that appear to be trustworthy by "parroting" user input while providing an unfaithful response.
2/3