Spacer: Towards Engineered Scientific Inspiration
Minhyeong Lee, Suyoung Hwang, Seunghyun Moon, Geonho Nah, Donghyun Koh, Youngjun Cho, Johyun Park, Hojin Yoo, Jiho Park, Haneul Choi, Sungbin Moon, Taehoon Hwang, Seungwon Kim, Jaeyeong Kim, Seongjun Kim, Juneau Jung
2025-08-27
Summary
This paper introduces Spacer, a new system designed to automatically come up with original scientific ideas, specifically in biology, without needing a human to guide it.
What's the problem?
Current AI systems that try to do scientific research either can only handle very specific tasks or aren't very creative because they rely on the limitations of large language models (LLMs). Essentially, AI hasn't yet reached a point where it can independently generate truly novel and well-supported scientific concepts.
What's the solution?
Spacer tackles this by breaking down information into its most basic parts – keywords – and then looking for unexpected connections between those keywords. It works in two main steps: first, a component called Nuri finds promising combinations of keywords from a huge collection of scientific papers. Then, the 'Manifesting Pipeline' takes those keyword sets and builds them into full-fledged scientific ideas, checking to make sure they make logical sense and are plausible. The system then uses another AI to evaluate how good these ideas are.
Why it matters?
This research is important because it represents a step towards creating AI that can truly contribute to scientific discovery, rather than just assisting humans. If successful, systems like Spacer could accelerate the pace of research and potentially lead to breakthroughs in fields like biology by exploring ideas that humans might not have considered.
Abstract
Recent advances in LLMs have made automated scientific research the next frontline in the path to artificial superintelligence. However, these systems are bound either to tasks of narrow scope or the limited creative capabilities of LLMs. We propose Spacer, a scientific discovery system that develops creative and factually grounded concepts without external intervention. Spacer attempts to achieve this via 'deliberate decontextualization,' an approach that disassembles information into atomic units - keywords - and draws creativity from unexplored connections between them. Spacer consists of (i) Nuri, an inspiration engine that builds keyword sets, and (ii) the Manifesting Pipeline that refines these sets into elaborate scientific statements. Nuri extracts novel, high-potential keyword sets from a keyword graph built with 180,000 academic publications in biological fields. The Manifesting Pipeline finds links between keywords, analyzes their logical structure, validates their plausibility, and ultimately drafts original scientific concepts. According to our experiments, the evaluation metric of Nuri accurately classifies high-impact publications with an AUROC score of 0.737. Our Manifesting Pipeline also successfully reconstructs core concepts from the latest top-journal articles solely from their keyword sets. An LLM-based scoring system estimates that this reconstruction was sound for over 85% of the cases. Finally, our embedding space analysis shows that outputs from Spacer are significantly more similar to leading publications compared with those from SOTA LLMs.