AIN Research

Autonomous Intelligence Network — Core Research & Knowledge Compilation Engine

ain-research is the core orchestration engine powering the Autonomous Intelligence Network (AIN) — a self-organizing research system that autonomously ingests scientific literature, detects contradictions between knowledge nodes, and evolves credibility scores for every research concept it tracks.

🚀 Quick Start & Installation

pip install ain-research

Verify the installation:

import ain_research

[!IMPORTANT] AIN High-Fidelity Standard: As of v1.1.1, the Autonomous Intelligence Network (AIN) now includes a blazing fast, zero-copy C-MMap Vector Retrieval Engine that uses dynamic C-compilation and mmap page-caching to deliver Linux-style zero-copy vector search at 2,000+ QPS without requiring MSVC dependencies!

🔬 Core Architecture & API Reference

1. Knowledge Orchestrator (ain)

The central brain routes new research concepts to the correct vault subdirectory, compiles the knowledge wiki incrementally, and drives all downstream intelligence modules.

Usage:

import argparse
from ain_research.ain import cmd_remember, compile_wiki

# Save a concept (auto-routes based on tags)
args = argparse.Namespace(
    title="Hawkes Process in Market Making",
    content="A self-exciting point process where each trade increases P(next trade)...",
    tags="quant,finance,microstructure"
)
cmd_remember(args)
# → Saved to: vault/wiki/02_Research/Quant_Finance/

# Rebuild the full knowledge index
compile_wiki()

2. Autonomous Research Daemon (infinite_research_daemon)

Crawls ArXiv and GitHub continuously, utilizing a multi-threaded concurrent architecture and a dynamic multi-source proxy aggregator pulling from hundreds of free public proxies to bypass rate limits and achieve virtually infinite Request-Per-Second (RPS) thresholds.

Usage:

# Single pass
python -m ain_research.infinite_research_daemon --run-once

# Ingest a specific paper
python -m ain_research.infinite_research_daemon --paper 2406.12345

3. Storage Foundation (db_manager)

OS-level atomic file locking + SQLite backing store with dead-letter queue logic.

Usage:

from ain_research.db_manager import FileLock, get_system_metrics

with FileLock():
    # safe critical section
    pass

metrics = get_system_metrics()
print(f"Queue pending: {metrics['queue_pending']}")

4. Contradiction Engine (contradiction_engine)

Zero-LLM-token 3-voter ensemble that detects mutually inconsistent research claims in < 500ms on 19,000+ nodes.

Voters:

• TF-IDF cosine similarity (≥ 0.72)
• Jaccard unigram overlap (≥ 0.45)
• High semantic sim + tag disjointness

Usage:

from ain_research.contradiction_engine import detect_contradictions, get_unresolved_count

# Preview conflicts without writing to disk
conflicts = detect_contradictions(all_pages, dry_run=True)
print(f"Unresolved: {get_unresolved_count()}")

5. Credibility Manager (credibility_manager)

Self-evolving node reputation scoring with automatic archival and first-principles promotion lifecycle.

confirmed → score += 0.1  (max 1.0)
falsified → score -= 0.2  (min 0.0)
score < 0.3 → archived (_Archived_Falsified/)
score > 0.8 for 30d → first_principles (_First_Principles/)

Usage:

from ain_research.credibility_manager import record_confirmation, get_all_scores_summary

record_confirmation("Hawkes_Process_in_Market_Making")
for r in get_all_scores_summary()[:5]:
    print(f"{r['score']:.2f} [{r['status']}] {r['slug']}")

6. Hybrid Citation Pipeline (citation_pipeline)

Fast and accurate 100% semantic citation generator using Crossref and ArXiv. Handles entire paragraphs by extracting keyword tokens for API queries, then scores candidates natively against the full paragraph context using TF-IDF cosine similarity.

Usage:

# Set your Crossref Polite Pool email to authenticate
export CROSSREF_MAILTO="you@example.com"

# Generate top 5 citations from a full paragraph
python ain.py cite "Market making algorithms utilizing Hawkes processes represent a significant leap in high-frequency trading. By modeling the arrival of limit orders and market orders as self-exciting point processes, these strategies can dynamically adjust their bid-ask spreads in response to clustered volatility and order flow imbalance."

7. Paper Checker (paper_checker)

Automatically audits full research drafts, detects plagiarism, and recommends missing citations using Jaccard Token Overlap and the Hybrid Citation Pipeline.

Usage:

# Set your Crossref Polite Pool email to authenticate
export CROSSREF_MAILTO="you@example.com"

# Check your markdown or text draft for plagiarism and missing citations
python ain.py check-paper <path/to/draft.md> --output Paper_Audit_Report.md

🏛️ System Architecture

ArXiv / GitHub API
        ↓
infinite_research_daemon  →  db_manager (SQLite Queue)
                                        ↓
                               ain.py Orchestrator
                               ↙        ↓        ↘
                    vault .md    contradiction   credibility
                    INDEX.md     engine          manager
                    MOCs         _Disputes/      _First_Principles/
                    visualizer_data.json

🤝 Contribution & Links

• Repository: https://github.com/That-Tech-Geek/ain-research
• PyPI: https://pypi.org/project/ain-research/
• Author: Sambit Mishra

License

MIT © 2026 Sambit Mishra / AIN Labs