wmayner/pyphi

A toolbox for integrated information theory.

What it solves

PyPhi is a research platform designed to compute integrated information (Φ), a metric used in Integrated Information Theory (IIT) to determine if a system exists as an integrated whole rather than a collection of independent parts. It allows researchers to analyze the cause-effect structure of a system to identify irreducible mechanisms and the relations that bind them.

How it works

The software takes a substrate—a network of interacting units defined by transition probabilities—and a specific state. It then calculates $\phi_s$ (system integrated information) by finding the partition that makes the least difference to the system's behavior. It also computes the $\Phi$-structure, which identifies the distinctions and relations within the system.

Who it’s for

It is primarily intended for researchers in neuroscience, consciousness studies, and complex systems who are applying Integrated Information Theory (IIT 3.0 and 4.0) to model physical substrates of consciousness.

Highlights

  • Multi-version Support: Implements both the current IIT 4.0 formalism and the earlier IIT 3.0 version.
  • Causal Analysis: Tools for determining actual causation and quantifying how a system's causal structure matches its environment.
  • Scale Analysis: Ability to coarse-grain or black-box substrates to analyze integrated information across different spatial and temporal scales.
  • Substrate Tools: Includes utilities for generating substrates from weight matrices or Ising models and estimating substrates from observed data.
  • Extensibility: Supports exporting results to pandas DataFrames, xarray, and NetworkX, with optional Redis-backed caching and Dask-based cluster execution.

Related

  • Project
    pgmpy/pgmpy

    A Python library for causal and probabilistic reasoning using graphical models, providing tools for structure learning, parameter estimation, and causal inference.

  • Project
    castorini/pyserini

    Pyserini is a Python toolkit that wraps the Anserini Lucene search engine and FAISS to provide easy, reproducible first‑stage retrieval (sparse, dense, or hybrid) for information‑retrieval research and RAG pipelines. It includes pre‑built indexes, evaluation scripts, and a REST server, and can be installed via `pip`.

  • Project
    pymc-labs/CausalPy

    A Python library for research-grade causal inference in quasi-experimental designs, offering Bayesian and OLS estimation with full uncertainty quantification.

  • Project
    Lightning-AI/LitServe

    LitServe is a Python-based framework for building custom inference servers, allowing developers to define their own logic for batching, routing, and orchestration for models, agents, and RAG pipelines.

  • Project
    fastai/fastai

    A deep learning library built on PyTorch that provides high-level components for rapid productivity and low-level components for research flexibility.