hypogenic

Name: hypogenic
Author: K-Dense-AI

31kstars3.1kforks

How to install hypogenic?

k-dense-ai/scientific-agent-skills/hypogenic

$npx -y skills add K-Dense-AI/scientific-agent-skills --skill hypogenic

Installs into the current project.

Summary

Automated LLM-driven hypothesis generation and testing on tabular datasets. Use when you want to systematically explore hypotheses about patterns in empirical data (e.g., deception detection, content analysis). Combines literature insights with data-driven hypothesis testing. For

Files · 1

View on GitHub

SKILL.md

1	# Hypogenic
2
3	## Overview
4
5	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).
6
7	## Quick Start
8
9	Get started with Hypogenic in minutes:
10
11	```bash
12	# Install the package
13	uv pip install hypogenic
14
15	# Clone example datasets
16	git clone https://github.com/ChicagoHAI/HypoGeniC-datasets.git ./data
17
18	# Run basic hypothesis generation
19	hypogenic_generation --config ./data/your_task/config.yaml --method hypogenic --num_hypotheses 20
20
21	# Run inference on generated hypotheses
22	hypogenic_inference --config ./data/your_task/config.yaml --hypotheses output/hypotheses.json
23	```
24
25	Or use Python API:
26
27	```python
28	from hypogenic import BaseTask
29
30	# Create task with your configuration
31	task = BaseTask(config_path="./data/your_task/config.yaml")
32
33	# Generate hypotheses
34	task.generate_hypotheses(method="hypogenic", num_hypotheses=20)
35
36	# Run inference
37	results = task.inference(hypothesis_bank="./output/hypotheses.json")
38	```
39
40	## When to Use This Skill
41
42	Use this skill when working on:
43	- Generating scientific hypotheses from observational datasets
44	- Testing multiple competing hypotheses systematically
45	- Combining literature insights with empirical patterns
46	- Accelerating research discovery through automated hypothesis ideation
47	- Domains requiring hypothesis-driven analysis: deception detection, AI-generated content identification, mental health indicators, predictive modeling, or other empirical research
48
49	## Key Features
50
51	Automated Hypothesis Generation
52	- Generate 10-20+ testable hypotheses from data in minutes
53	- Iterative refinement based on validation performance
54	- Support for both API-based (OpenAI, Anthropic) and local LLMs
55
56	Literature Integration
57	- Extract insights from research papers via PDF processing
58	- Combine theoretical foundations with empirical patterns
59	- Systematic literature-to-hypothesis pipeline with GROBID
60
61	Performance Optimization
62	- Redis caching reduces API costs for repeated experiments
63	- Parallel processing for large-scale hypothesis testing
64	- Adaptive refinement focuses on challenging examples
65
66	Flexible Configuration
67	- Template-based prompt engineering with variable injection
68	- Custom label extraction for domain-specific tasks
69	- Modular architecture for easy extension
70
71	Proven Results
72	- 8.97% improvement over few-shot baselines
73	- 15.75% improvement over literature-only approaches
74	- 80-84% hypothesis diversity (non-redundant insights)
75	- Human evaluators report significant decision-making improvements
76
77	## Core Capabilities
78
79	### 1. HypoGeniC: Data-Driven Hypothesis Generation
80
81	Generate hypotheses solely from observational data through iterative refinement.
82
83	Process:
84	1. Initialize with a small data subset to generate candidate hypotheses
85	2. Iteratively refine hypotheses based on performance
86	3. Replace poorly-performing hypotheses with new ones from challenging examples
87
88	Best for: Exploratory research without existing literature, pattern discovery in novel datasets
89
90	### 2. HypoRefine: Literature and Data Integration
91
92	Synergistically combine existing literature with empirical data through an agentic framework.
93
94	Process:
95	1. Extract insights from relevant research papers (typically 10 papers)
96	2. Generate theory-grounded hypotheses from literature
97	3. Generate data-driven hypotheses from observational patterns
98	4. Refine both hypothesis banks through iterative improvement
99
100	Best for: Research with established theoretical foundations, validating or extending existing theories
101
102	### 3. Union Methods
103
104	Mechanistically combine literature-only hypotheses with framework outputs.
105
106	Variants:
107	- Literature ∪ HypoGeniC: Combines literature hypotheses with data-driven generation
108	- Literature ∪ HypoRefine: Combines literature hypotheses with integrated approach
109
110	Best for: Comprehensive hypothesis coverage, eliminating redundancy while maintaining diverse perspectives
111
112	## Installation
113
114	Install via pip:
115	```bash
116	uv pip install hypogenic
117	```
118
119	Optional dependencies:
120	- Redis server (port 6832): Enables caching of LLM responses to significantly reduce API costs during iterative hypothesis generation
121	- s2orc-doc2json: Required for processing literature PDFs in HypoRefine workflows
122	- GROBID: R

People who liked this also viewed

6 picks

Type

hypogenic

31kstars3.1kforks

View on GitHub

How to install hypogenic?

k-dense-ai/scientific-agent-skills/hypogenic

$npx -y skills add K-Dense-AI/scientific-agent-skills --skill hypogenic

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Hypogenic
2
3	## Overview
4
5	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).
6
7	## Quick Start
8
9	Get started with Hypogenic in minutes:
10
11	```bash
12	# Install the package
13	uv pip install hypogenic
14
15	# Clone example datasets
16	git clone https://github.com/ChicagoHAI/HypoGeniC-datasets.git ./data
17
18	# Run basic hypothesis generation
19	hypogenic_generation --config ./data/your_task/config.yaml --method hypogenic --num_hypotheses 20
20
21	# Run inference on generated hypotheses
22	hypogenic_inference --config ./data/your_task/config.yaml --hypotheses output/hypotheses.json
23	```
24
25	Or use Python API:
26
27	```python
28	from hypogenic import BaseTask
29
30	# Create task with your configuration
31	task = BaseTask(config_path="./data/your_task/config.yaml")
32
33	# Generate hypotheses
34	task.generate_hypotheses(method="hypogenic", num_hypotheses=20)
35
36	# Run inference
37	results = task.inference(hypothesis_bank="./output/hypotheses.json")
38	```
39
40	## When to Use This Skill
41
42	Use this skill when working on:
43	- Generating scientific hypotheses from observational datasets
44	- Testing multiple competing hypotheses systematically
45	- Combining literature insights with empirical patterns
46	- Accelerating research discovery through automated hypothesis ideation
47	- Domains requiring hypothesis-driven analysis: deception detection, AI-generated content identification, mental health indicators, predictive modeling, or other empirical research
48
49	## Key Features
50
51	Automated Hypothesis Generation
52	- Generate 10-20+ testable hypotheses from data in minutes
53	- Iterative refinement based on validation performance
54	- Support for both API-based (OpenAI, Anthropic) and local LLMs
55
56	Literature Integration
57	- Extract insights from research papers via PDF processing
58	- Combine theoretical foundations with empirical patterns
59	- Systematic literature-to-hypothesis pipeline with GROBID
60
61	Performance Optimization
62	- Redis caching reduces API costs for repeated experiments
63	- Parallel processing for large-scale hypothesis testing
64	- Adaptive refinement focuses on challenging examples
65
66	Flexible Configuration
67	- Template-based prompt engineering with variable injection
68	- Custom label extraction for domain-specific tasks
69	- Modular architecture for easy extension
70
71	Proven Results
72	- 8.97% improvement over few-shot baselines
73	- 15.75% improvement over literature-only approaches
74	- 80-84% hypothesis diversity (non-redundant insights)
75	- Human evaluators report significant decision-making improvements
76
77	## Core Capabilities
78
79	### 1. HypoGeniC: Data-Driven Hypothesis Generation
80
81	Generate hypotheses solely from observational data through iterative refinement.
82
83	Process:
84	1. Initialize with a small data subset to generate candidate hypotheses
85	2. Iteratively refine hypotheses based on performance
86	3. Replace poorly-performing hypotheses with new ones from challenging examples
87
88	Best for: Exploratory research without existing literature, pattern discovery in novel datasets
89
90	### 2. HypoRefine: Literature and Data Integration
91
92	Synergistically combine existing literature with empirical data through an agentic framework.
93
94	Process:
95	1. Extract insights from relevant research papers (typically 10 papers)
96	2. Generate theory-grounded hypotheses from literature
97	3. Generate data-driven hypotheses from observational patterns
98	4. Refine both hypothesis banks through iterative improvement
99
100	Best for: Research with established theoretical foundations, validating or extending existing theories
101
102	### 3. Union Methods
103
104	Mechanistically combine literature-only hypotheses with framework outputs.
105
106	Variants:
107	- Literature ∪ HypoGeniC: Combines literature hypotheses with data-driven generation
108	- Literature ∪ HypoRefine: Combines literature hypotheses with integrated approach
109
110	Best for: Comprehensive hypothesis coverage, eliminating redundancy while maintaining diverse perspectives
111
112	## Installation
113
114	Install via pip:
115	```bash
116	uv pip install hypogenic
117	```
118
119	Optional dependencies:
120	- Redis server (port 6832): Enables caching of LLM responses to significantly reduce API costs during iterative hypothesis generation
121	- s2orc-doc2json: Required for processing literature PDFs in HypoRefine workflows
122	- GROBID: R

People who liked this also viewed

6 picks

Type