4 min read
Research
Turn a research paper into a runnable project: DLWAY reads the PDF, lets you review what it found, and generates code, runs and figures.
From paper to project
Research takes one or more papers (PDFs, or an arXiv identifier) and works towards a project that reproduces the experiment they describe.
The stages
- Sources: add the paper and any supplements
- Review: check what was extracted, and settle places where the sources disagree
- Configuration: every setting is listed with where its value came from, and can be overridden
- Generate & run: concrete project files are written, and a quick smoke run checks they work
- Results: metrics, training history and figures for each run
What you need
Analysis uses the AI provider configured in Settings. Importing a finished Research export needs no AI connection.
What it produces
Editable PyTorch, TensorFlow.js or scikit-learn code, the configuration behind it, and a record of each run that can be exported as a single archive.
Go deeper
The Research Guide covers every stage, supported models, limits and recovery. To try it without a paper or an AI key, follow Turn a Paper into a Running Experiment.