Can LLMs do research?
summary
Data-to-paper claims an LLM-driven pipeline that converts annotated CSVs into executable Python analysis, figures, and full manuscripts with programmatic traceability and review. It runs code (pandas scipy sklearn), calls Semantic Scholar/Crossref, enforces rule-based guardrails, back-traces every reported number into code and outputs as data-chained manuscripts, and reaches 80–90% autonomous correctness on simple datasets while relying on human co-piloting for complex tasks.
AI-written summary · about
chapters (16)
- 0:00LLM research capabilities
- 1:25Naive LLM paper approach
- 2:32AI research frameworks
- 3:36Hallmarks of scientific research
- 5:44Forced reasoning frameworks
- 7:31Research goal setting
- 9:20Rule-based agents validation
- 11:46Data to Paper system
- 12:28GUI app user experience
- 13:31Evaluating research papers
- 15:41Testing limits and co-piloting
- 17:45Verifiability with data chaining
- 20:59Trade-offs in AI science
- 21:20Current applications and adoption
- 22:03Conclusion and open-source
- 22:43Q&A on journal policies
related talks
33mDesigning LLM Applications for Massive Datasets
30mWhat Can LLMs Do with Messy Data?
15mLLMs for Data-Driven Decision Making in Product Development
10mThe Impact of LLMs on the Circular Economy and Beyond
47mBuilding Python based AI Systems with LLMs
23mSynthetic data that meets enterprise compliance needs
49mNext-Gen Documentation with AI
01h 19mBuilding an LLM on your Laptop to...