AI Science

title-bg-left-2

Data-to-paper

Collapse

icon-collapse

Expand

data-to-paper is an automation framework that systematically navigates interacting AI agents through a complete end-to-end scientific research, starting from raw data alone and concluding with transparent, backward-traceable, human-verifiable scientific papers.

This repository is the code implementation for the paper "Autonomous LLM-Driven Research — from Data to Human-Verifiable Research Papers".

Key features

  • End-to-end field-agnostic research. The process navigates through the entire scientific path, from data exploration, literature search and ideation, through data analysis and interpretation, to the step-by-step writing of a complete research papers.
  • Traceable "data-chained" manuscripts. Tracing information flow, data-to-paper creates backward-traceable and verifiable manuscripts, where any numeric values can be click-traced all the way up to the specific code lines that created them (data-chaining DEMO).
  • Autopilot or Copilot. The platform can run fully autonomously, or can be human-guided through the Copilot App, allowing users to:
    • Oversee, Inspect and Guide the research
    • Set research goals, or let the AI autonomously raise and test hypotheses
    • Provide review, or invoke on-demand AI-reviews
    • Rewind the process to prior steps
    • Record and replay runs
    • Track API costs
  • Coding guardrails. Standard statistical packages are overridden with multiple guardrails to minimize common LLM coding errors.

We invite people to try out data-to-paper with their own data and are eager for feedback and suggestions.

title-bg-left-2

Quibbler

Collapse

icon-collapse

Expand

Interactive, reproducible and efficient data analytics.

What is Quibbler?

Quibbler is a toolset for building highly interactive, yet reproducible, transparent and efficient data analysis pipelines. Quibbler allows using standard Python syntax to process data through any series of analysis steps, while automatically maintaining connectivity between downstream results and upstream raw data sources.

quibicon

Quibbler facilitates and embraces human interventions as an inherent part of the analysis pipeline: input parameters, as well as exceptions and overrides, can be specified and adjusted either programmatically, or by interacting with “live” graphics, and all such interventions are automatically recorded in well-documented human-machine readable files.

Changes to such parameters propagate downstream, pinpointing which specific data items, or even specific elements thereof, are affected, thereby vastly saving unnecessary recalculations.

Quibbler, therefore, facilitates hands-on interactions with data in ways that are not only flexible, fun and interactive, but also traceable, well-documented, and highly efficient.

Main Features

Here are a few of the things that Quibbler does:

  • Easily build powerful GUI-like interaction with data, without a need for callbacks and event listeners.
  • Interactive specification of inputs and overrides of parameter values.
  • Automatically create human-readable records of user interventions and parameter specifications.
  • Independently calculate, cache and validate/invalidate individual slices of heavy-to-calculate arrays.
  • Present a dependency graph between raw data and downstream results.
  • Provide inherent undo/redo functionalities.

All-of-the-above using completely standard functions and programming syntax – there is very little to learn to get started!

Best “Quibble” award