Dear all,
I'm happy to inform you that we are resuming the NLP Reading Group in the Autumn semester. The meeting will take place on Tuesday (18.10) at 10:00 in room N207 in IMPAKT.
Zdeněk Kasner will present the paper:
Standing on the Shoulders of Giant Frozen Language Models by Zoav Levine et al.
The authors introduce three new ways to how to apply very large pretrained models (>1B parameters) in downstream tasks without finetuning. The technique include: (1) using the model with input-dependent prompts, (2) using the model to generate an answer from retrieved documents, (3) using the model to re-rank its own outputs. These techniques allow to reach or exceed state-of-the-art performance on the downstream tasks while being computationally efficient.
Please at least skim the paper before the meeting. There will be a possibility to join remotely through zoom. I'll send a link in a separate mail.
See you,
Tomasz