InternAgent-1.5: A Unified Agentic Framework for Long-Horizon Autonomous Scientific Discovery
This is the official repository for HypoGeniC (Hypothesis Generation in Context) and HypoRefine, which are automated, data-driven tools that leverage large language models to generate hypothesis for open-domain research. For more details, please see the original paper using the link below.
Research feedback grounded in the papers that define your field
Systematically learn and evaluate the latent geometry of high-dimensional data, with a focus on scRNAseq analysis
Official implementation of the ACL 2024: Scientific Inspiration Machines Optimized for Novelty
AGATHA: Automatic Graph-mining And Transformer based Hypothesis generation Approach
A platform for Interactive AI-assisted Hypothesis Generation [ACL 2025]
Deep Critical Research Agent : The first AI-driven Critical Analysis to turn the huge amount of preclinical biology data and information into knowledge and hence disease cures.
Locally hosted AI Agent Python Tool To Generate Novel Research Hypothesis + Titles + Abstracts
Symbolic XAI explainable machine learning & non-linear regression platform for MATLAB
AstroAgents: Multi-Agent AI for Hypothesis Generation from Mass Spectrometry Data
Workflow to generate interactive feature selection report for cross-sectional or longitudinal studies
Official code for the paper: "Simple LLM Baselines are Competitive for Model Diffing"
A open source Literature Based Discovery system, using Groovy & Grails, including a re-implementation of the Arrowsmith algorithm(s)
How do you train retrievers to find inspirations? [ACL 2025]
A Python library for creating and exploring structured review articles.
Autonomous scientific discovery engine - 82 discovery modules, 16-phase pipeline with refinement and reflexion. GFlowNet hypothesis tournaments, adversarial skeptic review, multi-agent debate, causal reasoning, knowledge graphs, 17 domain ontologies. Terminal-native.
Cognitive Discovery System - pure Python scientific computing: quantum, ML (PCA/k-means/kNN), stats, stiff ODE solvers. 22 modules, zero runtime deps.
Scientific agent skills for Claude Code and Codex that turn research goals into auditable hypothesis generation, review, ranking, evolution, and synthesis.
IRIS Gate Protocol: Cross-architecture phenomenological convergence research (RFC v0.2)
To associate your repository with the hypothesis-generation topic, visit your repo's landing page and select "manage topics."