Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers

Chenglei Si, Diyi Yang, Tatsunori Hashimoto, 2024. arXiv.

A study in which more than 100 NLP researchers wrote and reviewed research ideas, comparing ideas generated by language models with those written by experts.

Voynich ManuscriptPreprintChecked against the original0

What it's used for here

The inquiry-framework survey cites its result that LLM ideas were rated more novel but slightly less feasible, as evidence that novelty is cheap.

Cited by on this site 1 record

1 methods surveys

Frameworks of inquiry

SRC

Cite it

Si, C., Yang, D., & Hashimoto, T. (2024). Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers. arXiv. https://arxiv.org/abs/2409.04109

@article{si2024can,
  author = {Si, Chenglei and Yang, Diyi and Hashimoto, Tatsunori},
  title = {Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers},
  journal = {arXiv},
  year = {2024},
  url = {https://arxiv.org/abs/2409.04109}
}

TY  - GEN
TI  - Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers
AU  - Si, Chenglei
AU  - Yang, Diyi
AU  - Hashimoto, Tatsunori
PY  - 2024
JO  - arXiv
UR  - https://arxiv.org/abs/2409.04109
ER  -

Si, Chenglei et al.. “Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers.” arXiv (2024).