@branislavpecher.bsky.social
PhD student at Kempelen Institute of Intelligent Technologies. Working on low-resource and limited labelled data settings. All about LLM fine-tuning, prompting, in-context learning and their sensitivity to non-deterministic choices (initialisation, prompt)
I started doing this for every one of my papers. But I always have a bit of hesitation about whether it is something I am "allowed" to do and whether it will help/achieve something -- it is just more text for the AC to read and by default biased
December 2, 2024 at 8:31 PM
I started doing this for every one of my papers. But I always have a bit of hesitation about whether it is something I am "allowed" to do and whether it will help/achieve something -- it is just more text for the AC to read and by default biased
Similar here. Zero discussion on the paper.
At least the reviewer that submitted LLM generated review answered that they are keeping the score. 🙃 But it was still much better than the NeurIPS reviews
I am really curious how the AC will fare with the "discussion regarding borderline papers"
At least the reviewer that submitted LLM generated review answered that they are keeping the score. 🙃 But it was still much better than the NeurIPS reviews
I am really curious how the AC will fare with the "discussion regarding borderline papers"
November 30, 2024 at 11:51 AM
Similar here. Zero discussion on the paper.
At least the reviewer that submitted LLM generated review answered that they are keeping the score. 🙃 But it was still much better than the NeurIPS reviews
I am really curious how the AC will fare with the "discussion regarding borderline papers"
At least the reviewer that submitted LLM generated review answered that they are keeping the score. 🙃 But it was still much better than the NeurIPS reviews
I am really curious how the AC will fare with the "discussion regarding borderline papers"
But according to reviewers "gradient step in a promising direction" is just an incremental contribution and should be rejected 🙃
On a more serious note, I believe more conferences should adopt the ARR way of publishing a "reviewer guideline", with common review mistakes (e.g., "paper is not novel")
On a more serious note, I believe more conferences should adopt the ARR way of publishing a "reviewer guideline", with common review mistakes (e.g., "paper is not novel")
November 29, 2024 at 4:26 PM
But according to reviewers "gradient step in a promising direction" is just an incremental contribution and should be rejected 🙃
On a more serious note, I believe more conferences should adopt the ARR way of publishing a "reviewer guideline", with common review mistakes (e.g., "paper is not novel")
On a more serious note, I believe more conferences should adopt the ARR way of publishing a "reviewer guideline", with common review mistakes (e.g., "paper is not novel")
Currently struggling with the same thing. One of the papers read like it was written last minute without the authors even reading it afterwards and checking basic things/mistakes.
Now, in the rebuttal phase, the authors have completely rewritten the whole paper. Is this allowed/encouraged? 🤔
Now, in the rebuttal phase, the authors have completely rewritten the whole paper. Is this allowed/encouraged? 🤔
November 25, 2024 at 8:54 PM
Currently struggling with the same thing. One of the papers read like it was written last minute without the authors even reading it afterwards and checking basic things/mistakes.
Now, in the rebuttal phase, the authors have completely rewritten the whole paper. Is this allowed/encouraged? 🤔
Now, in the rebuttal phase, the authors have completely rewritten the whole paper. Is this allowed/encouraged? 🤔
Does complaining about the review process and quality of reviews count as being grumpy? 😅
November 19, 2024 at 8:44 AM
Does complaining about the review process and quality of reviews count as being grumpy? 😅