our work on noise-robust training for NER has been accepted to #eacl2027. great work w/ @EMerdjanovska @alan_akbik
Latest posts 4
We’re releasing v1.3 of our decider models today with improved performance and a new locate feature enabling you to extract information from your input. Check it out, it’s free on OpenRouter for the next week.
We're rolling out pplx-decider-v1.3 and pplx-decider-flash-v1.3, the latest version of our open-weight decision models.View quoted post on X
v1.3 adds locate, a new feature that extracts spans based on type or instructions.
Now available in our Decisions API, @huggingface, and @OpenRouter.
perplexity.ai/hub/blog/perpl…
I'll be in Hamburg this week at KONVENS where Pascal Stolzenburg will present our work on news-crawler-LM (arxiv.org/abs/2607.21284)!
At our chair, we develop Fundus (github.com/flairnlp/fundus), a fast and high-quality news crawler using tailored rules for each publisher.
At our chair, we develop Fundus (github.com/flairnlp/fundus), a fast and high-quality news crawler using tailored rules for each publisher.
Happy to share that my last paper of my PhD has been accepted to #EMNLP!
In this work, we empirically investigated the design choices behind multilingual NER models (architectures, loss functions, synthetic data, and backbones) and measure their impact on zero-shot performance.
In this work, we empirically investigated the design choices behind multilingual NER models (architectures, loss functions, synthetic data, and backbones) and measure their impact on zero-shot performance.