Sandeep Kumar
@Sandeepiitp
NLP Researcher, Prime Minister Research Fellow, LLM in Science. interested in HCL +NLP
Thanks for sharing our paper. We further conducted a human evaluation on computer science and found that LLMs have the potential to generate relevant, distinct, feasible, and novel ideas to some extent.
This is the 3rd big paper in the span of a couple weeks that investigates whether LLM can generate *novel* scientific research ideas. This paper reports that Claude and GPT models tend to align more with the author's perspectives on future research ideas. This is measured across…
🌅 Finally, a real challenger to Google’s dominance — OpenAI launches Atlas. Will Google become the next Yahoo? 🤔. We knew this was coming when Sam Altman said “I legitimately cannot tell you the last time I did a Google search.” #OpenAI #Atlas #AI
🚀 JUST-NLP 2025 – Shared Tasks on Legal NLP🚀 We invite you to participate in the JUST NLP 2025 Shared Tasks on Legal NLP, co-located with IJCNLP–AACL 2025, at the Victor Menezes Convention Centre, Indian Institute of Technology, Bombay, India, on December 24, 2025. #NLP #AACL
🎉 Excited to share that our paper have been accepted to EMNLP 2025 (Main Track)! Title: Can Large Language Models Unlock Novel Scientific Research Ideas? Paper link and Code: arxiv.org/abs/2409.06185 Looking forward to presenting in Suzhou, China 🇨🇳! #EMNLP2025 @emnlpmeeting
✨ Our paper "CourtEval: A Courtroom-Based Multi-Agent Evaluation Framework" got accepted to ACL (Findings)! 🎉 Huge thanks to my collaborators See you in Vienna! 📷#ACL2025 #NLProc @aclmeeting #NLP #LLMs
Thrilled to share that I attended 𝗡𝗔𝗔𝗖𝗟 𝟮𝟬𝟮𝟱 𝗶𝗻 𝗔𝗹𝗯𝘂𝗾𝘂𝗲𝗿𝗾𝘂𝗲, 𝗡𝗲𝘄 𝗠𝗲𝘅𝗶𝗰𝗼, where I presented my paper “MixRevDetect: Towards Detecting AI-Generated Content in Hybrid Peer Reviews” Paper: lnkd.in/d3-V9Yn2 Albuquerque was equally memorable.
🤖Is your peer reviews generated by ChatGPT? Can they be detected? #EMNLP2024 Attend my talk, Nov 13 (Wed), session 12, 16:00-16:15 - Room: Flagler
🌊Reached #EMNLP2024 , presenting paper on Wed Eve! Aside from this, I'd love to chat about: • AI text detection/Watermarking • AI in Science • LLM as Judge • Hypothesis Generation Also, I'm on the lookout for a summer 2025 internship/PostDoc!
🎉EMNLP 2024🎉 🎉Thrilled to announce that our paper is accepted at EMNLP 2024 Main! 'Quis custodiet ipsos custodes?' Who will watch the watchmen? On Detecting AI-generated peer-reviews" We propose an algorithm for detecting AI generated Fake Reviews. #EMNLP2024.
Emnlp results are out early on Open Reviews. #Emnlp @emnlpmeeting
📚 Paper "Can Large Language Models Unlock Novel Scientific Research Ideas?" - Gives insights into the evolving role of LLMs in idea generation, both its capability and limitations. - Talks about some common matrics in this regard. - Contributes to the ongoing efforts in…
New research reveals that Claude-2 and GPT-4 outperform other models in generating aligned and diverse research ideas, contributing valuable insights into the evolving role of LLMs in academic creativity. : emergentmind.com/papers/2409.06…
A path towards AGI: LLMs that generate novel ideas What if we could enhance LLMs so that they can generate novel ideas independently and go further in their own decision-making? Several studies analyzed LLMs' creativity in generating new research ideas, showing the tendency of…
10. Can Large Language Models Unlock Novel Scientific Research Ideas? 🔑 Keywords: Large Language Models, idea generation, research domains, Claude-2, GPT-4 💡 Category: Natural Language Processing 🌟 Research Objective: To explore the capability of Large Language Models…
The paper explores the capability of Large Language Models (LLMs) in generating novel research ideas based on information from research papers. The authors conduct a thorough examination of 4 LLMs - Claude-2, Gemini-1.0, GPT-3.5, and GPT-4 - across five domains: Computer Science,…
LLMには”優れた新アイディアを生み出す能力がある”という事実が続々と報告されている。 科学研究のテーマで、LLMのアイディア創出能力を専門家が評価した結果、 - GPT-4のアイディアの7%は極めて斬新で、93%は実現可能 と判断された。 さらに、RAGなどで背景情報を加えると、実現性がより高まる。
This is the 3rd big paper in the span of a couple weeks that investigates whether LLM can generate *novel* scientific research ideas. This paper reports that Claude and GPT models tend to align more with the author's perspectives on future research ideas. This is measured across…
Sounds like a promising step forward for AI in scientific research
This was already proven on a narrow basis in a paper published by Nature in 2019. Looks like narrow to general took 5 years.
This is the 3rd big paper in the span of a couple weeks that investigates whether LLM can generate *novel* scientific research ideas. This paper reports that Claude and GPT models tend to align more with the author's perspectives on future research ideas. This is measured across…
LLM 跨学科研究创意生成能力评估 本篇论文主要是在探究 LLMs 是否能够根据研究论文生成新颖的研究创意，也是近期讨论很多的话题，感谢 @omarsar0 分享论文见解。 研究者对 4 个 LLM (Claude-2、GPT-3.5、GPT-4 和 Gemini) 在 5 个学科领域 (化学、计算机科学、经济学、医学和物理学)…
This is the 3rd big paper in the span of a couple weeks that investigates whether LLM can generate *novel* scientific research ideas. This paper reports that Claude and GPT models tend to align more with the author's perspectives on future research ideas. This is measured across…
United States Trends
- 1. $UNI 1,705 posts
- 2. Rempe 1,186 posts
- 3. Vesia N/A
- 4. Bannon 13.5K posts
- 5. Chuck 93K posts
- 6. Logan Wilson 1,392 posts
- 7. Dragon 91.2K posts
- 8. #LightningStrikes N/A
- 9. Pacers 10.3K posts
- 10. #USWNT N/A
- 11. Emil Heineman N/A
- 12. Jim Marshall N/A
- 13. Aaron Jones 2,649 posts
- 14. Macklin N/A
- 15. Lilleberg N/A
- 16. #BoltUp 1,566 posts
- 17. Chargers -3 27.6K posts
- 18. #LOVERGIRL 6,139 posts
- 19. Candace 41.9K posts
- 20. Jordan Mason 1,019 posts
Something went wrong.
Something went wrong.