Discussion Solved

Experiences with CSIR-UGC NET.preparation strategies that worked (66)

Back to Forum
Growing 60 pts 0 followers
National Institute of Technology Karnataka · Posted
Operations & Supply Chain Educational Assessment Higher Education Policy

I have been on both sides.as a reviewer for IEEE and ACM journals, and as an author who has received peer review. The experience has made me wonder whether the current closed peer review model is serving science well.

As a reviewer: I put significant effort into reviews (usually 1500–2500 words). But I occasionally see reviews from other reviewers that are 3 sentences and completely uninformative.

As an author: I have received reviews that clearly misunderstood the paper's contribution. I have also received brilliant reviews that genuinely improved the paper.

Discussion question: What changes to peer review would actually improve quality? Open peer review? Paid reviewing? Double-blind everywhere?

Interested to hear from people across different fields and career stages.

Sign in to interact

Sign in to join the discussion.

1 Reply

Oldest Top
0
Rohan Desai Distinguished 950 pts · Accepted answer

Great question.I went through something very similar in my second year.

The key insight for me was that LoRA is actually quite well-suited for NER tasks, especially in low-resource settings. I would recommend:

  1. Use LoRA with r=8 or r=16.don't go higher for 8K samples
  2. Apply LoRA to attention layers only, not the feed-forward layers
  3. Use a cosine learning rate schedule with warm-up (10% of steps)

For Telugu-English code-mixed NER specifically, you might also look at MuRIL.it is pretrained on Indian language data and often outperforms XLM-R on Indic tasks even with less fine-tuning data.