iiinova — where your digital self keeps evolving

Reposted by Patrick Haller

Digital Linguistics Lab @dili-lab.bsky.social · 26/03/2025

Excited to share that our group will present 9 papers at this year's ACM Symposium on Eye Tracking Research & Applications (ETRA) in Tokyo! We will post summaries of each paper in the coming weeks, but here's a quick sneak peek 👀

173

Reposted by Patrick Haller

Digital Linguistics Lab @dili-lab.bsky.social · 25/03/2025

At this year's ACL in Vienna, @lenajaeger.bsky.social and David Reich from our group, together with @whylikethis.bsky.social and Omer Shubi, will be hosting a tutorial on EyeTracking and NLP 👀 🖥️ Be there to join us! More information can be found here: acl2025-eyetracking-and-nlp.github.io

acl2025-eyetracking-and-nlp.github.io

ACL 2025 Tutorial: Eyetracking and NLP

ACL 2025 Tutorial on Eyetracking and NLP

073

Reposted by Patrick Haller

Naomi Saphra @nsaphra.bsky.social · 20/12/2024

Transformer LMs get pretty far by acting like ngram models, so why do they learn syntax? A new paper by sunnytqin.bsky.social, me, and @dmelis.bsky.social illuminates grammar learning in a whirlwind tour of generalization, grokking, training dynamics, memorization, and random variation. #mlsky #nlp

arxiv.org

Sometimes I am a Tree: Data Drives Unstable Hierarchical Generalization

Language models (LMs), like other neural networks, often favor shortcut heuristics based on surface-level patterns. Although LMs behave like n-gram models early in training, they must eventually learn...

514230

Reposted by Patrick Haller

Rico Sennrich @ricosennrich.bsky.social · 06/12/2024

Congratulations to Dr. @tannonk.bsky.social, who just successfully defended his thesis on "Leveraging Data, Decoding, and Context for Controlling Text Generation from Pretrained Language Models". Special thanks to the external examiner @feralvam.bsky.social!

0195