by
  • Idea: Could journal rejection feedback be anonymised and pooled as public tr...

    Research Idea

    Rejection letters contain valuable feedback, but they are private. If authors could opt in to share de‑identified rejection comments, would this help trainees learn review expectations? What risks?

    Fang Fang, Ethan Cook and 28 others
    2 Comments
  • Will domain‑specific fine‑tuned small models compete with large general‑purp...

    Research Idea

    Large foundation models dominate many discussions, yet domain‑specific smaller fine‑tuned models have advantages on specialised tasks. What roles will small domain‑adapted models occupy in future academic research workflows?

    Gregory Adams, George Perez and 38 others
    4 Comments
    View more comments
    • Small domain‑specific models cost far less compute to run and reproduce.

      2
      • They may reach competitive performance on narrow well‑defined domain tasks.

        2
        • View more comments