Happy to share our paper accepted by #EMNLP2024!🎉
We introduce a framework for dialogue constructiveness assessment, to train interpretable LLM feature-based models that outcompete (or on par with) neural models and standard feature-based models, with better robustness.
🧵1/9
Thrilled to share my 1st PhD preprint 🥳. We uncovered a few surprises about alternative splicing and common diseases and found an unexpected role of splice junctions used in a minority of transcripts! 🧵 and preprint👇 (1/15)
https://t.co/vwxgDyfeGV
It's great to see our colleagues @vlachos_nlp and Youmna Farag involved in this @EPSRC project on @OpeningUpMinds : Engaging Dialogue Generated from Argument Maps' with @OUComputing@sheffielduni and the Cambridge Lab of Toshiba Research Europe.
#NAACL2018 paper "Neural Automated Essay Scoring and Coherence Modeling for Adversarially Crafted Input" by Youmna Farag, @HYannakoudakis and Ted Briscoe. Neural essay scoring model, able to detect incoherent input. #nlproc https://t.co/F1fQwhCcyx