Hypogenic

Overview

Hypogenic provides automated hypothesis generation and testing using large language models to accelerate scientific discovery. The framework supports three approaches: HypoGeniC (data-driven hypothesis generation), HypoRefine (synergistic literature and data integration), and Union methods (mechanistic combination of literature and data-driven hypotheses).

Quick Start

Get started with Hypogenic in minutes:

# Install into its own environment: hypogenic 0.3.5 pins numpy~=1.26, pandas~=2.1,
# torch~=2.4, transformers~=4.45, openai~=1.40 and anthropic~=0.32 (Python 3.10–3.12)
uv venv .venv-hypogenic --python 3.10 && source .venv-hypogenic/bin/activate
uv pip install hypogenic

# Clone example datasets
git clone https://github.com/ChicagoHAI/HypoGeniC-datasets.git ./data

# Run data-driven hypothesis generation (model_type: gpt | claude | vllm | huggingface)
hypogenic_generation --task_config_path ./data/your_task/config.yaml \
    --model_type gpt --model_name <provider-model-id> \
    --max_num_hypotheses 20 --output_folder ./outputs/your_task

# Run inference with a saved hypothesis bank
hypogenic_inference --task_config_path ./data/your_task/config.yaml \
    --hypothesis_file ./outputs/your_task/<hypotheses_*.json> \
    --model_type gpt --model_name <provider-model-id>

Flags verified against the hypogenic 0.3.5 console scripts (hypogenic_cmd). The CLI covers data-driven HypoGeniC generation and inference; HypoRefine and Union runs are driven by the example scripts (below). Run --help for the full option list.

Or use the example scripts (the library ships runnable scripts under examples/; there is no one-line fluent task.generate_hypotheses(...) API — see "Python API Usage" below for the real classes):

python ./examples/generation.py   --help   # HypoGeniC data-driven generation
python ./examples/inference.py     --help   # single-hypothesis inference

When to Use This Skill

Use this skill when working on:

• Generating scientific hypotheses from observational datasets
• Testing multiple competing hypotheses systematically
• Combining literature insights with empirical patterns
• Accelerating research discovery through automated hypothesis ideation
• Domains requiring hypothesis-driven analysis: deception detection, AI-generated content identification, mental health indicators, predictive modeling, or other empirical research

Does NOT Trigger

Key Features

Automated Hypothesis Generation

• Generate 10-20+ testable hypotheses from data in minutes
• Iterative refinement based on validation performance
• Support for both API-based (OpenAI, Anthropic) and local LLMs

Literature Integration

• Extract insights from research papers via PDF processing
• Combine theoretical foundations with empirical patterns
• Systematic literature-to-hypothesis pipeline with GROBID

Performance Optimization

• Redis caching reduces API costs for repeated experiments
• Parallel processing for large-scale hypothesis testing
• Adaptive refinement focuses on challenging examples

Flexible Configuration

• Template-based prompt engineering with variable injection
• Custom label extraction for domain-specific tasks
• Modular architecture for easy extension

Reported Results (from arXiv:2410.17309, Literature Meets Data)

• +8.97% over few-shot, +15.75% over literature-only, +3.37% over data-driven-only baselines
• Human accuracy improved +7.44% (deception detection) and +14.19% (AI-generated content detection)
• A redundancy checker prunes near-duplicate hypotheses to keep the final bank diverse

Core Capabilities

1. HypoGeniC: Data-Driven Hypothesis Generation

Generate hypotheses solely from observational data through iterative refinement.

Process:

1. Initialize with a small data subset to generate candidate hypotheses
2. Iteratively refine hypotheses based on performance
3. Replace poorly-performing hypotheses with new ones from challenging examples

Best for: Exploratory research without existing literature, pattern discovery in novel datasets

2. HypoRefine: Literature and Data Integration

Synergistically combine existing literature with empirical data through an agentic framework.

Process:

1. Extract insights from relevant research papers (typically 10 papers)
2. Generate theory-grounded hypotheses from literature
3. Generate data-driven hypotheses from observational patterns
4. Refine both hypothesis banks through iterative improvement

Best for: Research with established theoretical foundations, validating or extending existing theories

3. Union Methods

Mechanistically combine literature-only hypotheses with framework outputs.

Variants:

• Literature ∪ HypoGeniC: Combines literature hypotheses with data-driven generation
• Literature ∪ HypoRefine: Combines literature hypotheses with integrated approach

Best for: Comprehensive hypothesis coverage, eliminating redundancy while maintaining diverse perspectives

Installation

Install via pip:

uv pip install hypogenic

Because of the pinned dependency versions above, keep hypogenic out of your main analysis environment.

Optional dependencies:

• Redis server (port 6832): Enables caching of LLM responses to significantly reduce API costs during iterative hypothesis generation
• s2orc-doc2json: Required for processing literature PDFs in HypoRefine workflows
• GROBID: Required for PDF preprocessing (see Literature Processing section)

Clone example datasets:

# For HypoGeniC examples
git clone https://github.com/ChicagoHAI/HypoGeniC-datasets.git ./data

# For HypoRefine/Union examples
git clone https://github.com/ChicagoHAI/Hypothesis-agent-datasets.git ./data

Dataset Format

Datasets must follow HuggingFace datasets format with specific naming conventions:

Required files:

• <TASK>_train.json: Training data
• <TASK>_val.json: Validation data
• <TASK>_test.json: Test data

Required keys in JSON:

• text_features_1 through text_features_n: Lists of strings containing feature values
• label: List of strings containing ground truth labels

Example (headline click prediction):

{
  "headline_1": [
    "What Up, Comet? You Just Got *PROBED*",
    "Scientists Made a Breakthrough in Quantum Computing"
  ],
  "headline_2": [
    "Scientists Everywhere Were Holding Their Breath Today. Here's Why.",
    "New Quantum Computer Achieves Milestone"
  ],
  "label": [
    "Headline 2 has more clicks than Headline 1",
    "Headline 1 has more clicks than Headline 2"
  ]
}

Important notes:

• All lists must have the same length
• Label format must match your extract_label() function output format
• Feature keys can be customized to match your domain (e.g., review_text, post_content, etc.)

Configuration

Each task requires a config.yaml file specifying:

Required elements:

• Dataset paths (train/val/test)
• Prompt templates for:
• Observations generation
• Batched hypothesis generation
• Hypothesis inference
• Relevance checking
• Adaptive methods (for HypoRefine)

Template capabilities:

• Dataset placeholders for dynamic variable injection (e.g., ${text_features_1}, ${num_hypotheses})
• Custom label extraction functions for domain-specific parsing
• Role-based prompt structure (system, user, assistant roles)

Configuration structure:

task_name: your_task_name

train_data_path: ./your_task_train.json
val_data_path: ./your_task_val.json
test_data_path: ./your_task_test.json

prompt_templates:
  # Extra keys for reusable prompt components
  observations: |
    Feature 1: ${text_features_1}
    Feature 2: ${text_features_2}
    Observation: ${label}
  
  # Required templates
  batched_generation:
    system: "Your system prompt here"
    user: "Your user prompt with ${num_hypotheses} placeholder"
  
  inference:
    system: "Your inference system prompt"
    user: "Your inference user prompt"
  
  # Optional templates for advanced features
  few_shot_baseline: {...}
  is_relevant: {...}
  adaptive_inference: {...}
  adaptive_selection: {...}

Refer to references/config_template.yaml for a complete example configuration.

Literature Processing (HypoRefine/Union Methods)

To use literature-based hypothesis generation, you must preprocess PDF papers:

Step 1: Setup GROBID (first time only)

bash ./modules/setup_grobid.sh

Step 2: Add PDF files Place research papers in literature/YOUR_TASK_NAME/raw/

Step 3: Process PDFs

# Start GROBID service
bash ./modules/run_grobid.sh

# Process PDFs for your task
cd examples
python pdf_preprocess.py --task_name YOUR_TASK_NAME

This converts PDFs to structured format for hypothesis extraction. Automated literature search will be supported in future releases.

CLI Usage

Hypothesis Generation

hypogenic_generation --help

Key parameters:

• Task configuration file path
• Model selection (API-based or local)
• Generation method (HypoGeniC, HypoRefine, or Union)
• Number of hypotheses to generate
• Output directory for hypothesis banks

Hypothesis Inference

hypogenic_inference --help

Key parameters:

• Task configuration file path
• Hypothesis bank file path
• Test dataset path
• Inference method (default or multi-hypothesis)
• Output file for results

Python API Usage

The library is not a one-call fluent API — generation runs as an explicit init/update loop over the algorithm classes (DefaultGeneration, DefaultInference, DefaultUpdate, DefaultReplace), and inference runs through the inference_register. Copy and adapt the scripts under examples/; for HypoRefine/Union adapt examples/union_generation.py.

Full import list, the generation/inference loop, inference_type strategy options, and the critical extract_label() parsing contract: see references/python_api.md.

Workflow Examples

Three end-to-end scenarios — data-driven (HypoGeniC, AI-content detection), literature-informed (HypoRefine, deception in hotel reviews), and comprehensive coverage (Union, mental-stress detection) — with dataset prep, generation, and inference commands for each: see references/workflow_examples.md.

Performance Optimization

Caching: Enable Redis caching to reduce API costs and computation time for repeated LLM calls

Parallel Processing: Leverage multiple workers for large-scale hypothesis generation and testing

Adaptive Refinement: Use challenging examples to iteratively improve hypothesis quality

Troubleshooting

Issue: Generated hypotheses are too generic Solution: Refine prompt templates in config.yaml to request more specific, testable hypotheses

Issue: Poor inference performance Solution: Ensure dataset has sufficient training examples, adjust hypothesis generation parameters, or increase number of hypotheses

Issue: Label extraction failures Solution: Implement custom extract_label() function for domain-specific output parsing

Issue: GROBID PDF processing fails Solution: Ensure GROBID service is running (bash ./modules/run_grobid.sh) and PDFs are valid research papers

Creating Custom Tasks

Adding a new task follows five steps: (1) prepare train/val/test JSON with text_features_* + label keys, (2) author config.yaml, (3) implement a custom extract_label(), (4) optionally process literature PDFs for HypoRefine/Union, and (5) run generation + inference.

Full step-by-step guide with the custom extract_label implementation and BaseTask wiring: see references/custom_tasks.md.

Repository Structure

Core code lives in hypogenic/, CLI entry points in hypogenic_cmd/, the HypoRefine agent in hypothesis_agent/, PDF/literature tools in literature/ + modules/, and runnable scripts in examples/. Full annotated directory tree: see references/repository_structure.md.

Related Publications

The framework rests on three papers from ChicagoHAI: HypoBench (2025, arXiv:2504.11524), Literature Meets Data (2024, arXiv:2410.17309, introduces HypoRefine), and the original Hypothesis Generation with Large Language Models (2024, EMNLP NLP4Science). Full citations, descriptions, and BibTeX entries: see references/publications.md.

Additional Resources

Official Links

• GitHub Repository: https://github.com/ChicagoHAI/hypothesis-generation
• PyPI Package: https://pypi.org/project/hypogenic/
• License: MIT License
• Issues & Support: https://github.com/ChicagoHAI/hypothesis-generation/issues

Example Datasets

Clone these repositories for ready-to-use examples:

# HypoGeniC examples (data-driven only)
git clone https://github.com/ChicagoHAI/HypoGeniC-datasets.git ./data

# HypoRefine/Union examples (literature + data)
git clone https://github.com/ChicagoHAI/Hypothesis-agent-datasets.git ./data

For contributions or questions, visit the GitHub repository and check the issues page.

Local Resources

references/

• config_template.yaml — a runnable-shape config.yaml matching the real hypogenic

schema (task_name, train/val/test_data_path, prompt_templates with role-based system/user sub-keys and ${...} placeholders). Read it before authoring a config: it documents the ${...} substitution, the reusable "extra key" mechanism, and which settings belong in CLI flags rather than the YAML.

• python_api.md — the init/update generation loop, inference registry, inference_type

strategy options, and the extract_label() parsing contract.

• workflow_examples.md — three end-to-end scenarios (HypoGeniC, HypoRefine, Union).
• custom_tasks.md — five-step guide for adding a new task or dataset.
• repository_structure.md — annotated layout of the upstream repository.
• publications.md — full citations and BibTeX for the HypoGeniC/HypoRefine/HypoBench papers.

Part of the AlterLab Academic Skills suite.