CoolFace
Apppublic

esachdev12/CLINOVA

sourceHugging Facemitupdated 6mo agoView on Hugging Face
0likes
App README

FinomIQ — Autonomous Financial Strategy Intelligence Platform

FinomIQ is a production-grade decision intelligence system for autonomous financial strategy design, simulation, and evolution. It provides a comprehensive Strategy Sandbox where users can construct complex, rule-based strategies that interact with AI-driven market environments.


System Architecture

FinomIQ is built on a modular, multi-layered architecture:

1. Strategy Sandbox Engine (CORE)

  • Strategy DSL: A rule-based Domain-Specific Language for defining financial logic.
  • Hybrid Execution: Integrates user-defined rules with state-of-the-art RL agents (PPO, DQN).
  • What-If Analysis: Evaluates strategy performance under alternative market conditions.

2. Market Simulation Engine

  • GBM + Jump Diffusion: High-fidelity price simulation with stochastic jumps and mean reversion.
  • Regime Switching: Dynamic transitions between Bull, Bear, Volatile, and Crash regimes.

3. Portfolio & Risk Intelligence

  • Institutional Metrics: Real-time tracking of Alpha, Beta, Sharpe Ratio, and Sortino Ratio.
  • Advanced Analytics: Peak-to-trough Drawdown analysis and Risk-Return Efficiency profiling.

📊 Observation & Action Spaces

Observation Space

The agent receives a rich state vector representing institutional market data:

  • asset_prices: Current price of all tradable assets.
  • volatility_index: VIX-like market uncertainty indicator.
  • market_trend_score: Momentum signal.
  • portfolio_value: Total valuation of holdings and cash.
  • risk_exposure_score: Current peak-to-trough drawdown.
  • news_sentiment_score: NLP-derived alternative data signal.
  • macro_indicators: Inflation and Interest Rate proxies.

Action Space

The agent can execute seven institutional-grade financial operations:

  • 0: BUY: Allocate capital to a specific asset.
  • 1: SELL: Exit or reduce positions.
  • 2: HOLD: Maintain current allocation.
  • 3: REBALANCE: Equal-weight portfolio realignment.
  • 4: ANALYZE: Spend budget to improve signal clarity.
  • 5: HEDGE: Execute defensive moves using non-correlated assets.
  • 6: LIQUIDATE: Complete exit to cash reserve.

🧪 OpenEnv Tasks

Task IDNameDifficultyObjective
bull_market_growthBull Market GrowthEasyMaximize profit in a steady upward trending market.
volatile_market_navigationVolatile NavigationMediumAchieve high Sharpe ratio despite extreme price swings.
market_crash_survivalCrash SurvivalHardProtect capital and survive a severe market liquidation event.

📈 Baseline Performance

Results from the inference.py baseline script:

TaskBaseline Score (0-1)Status
Bull Market Growth0.85PASS
Volatile Navigation0.62PASS
Crash Survival0.45EVAL

🚀 Execution

Institutional Terminal (Gradio)

bash
python3 gradio_app.py

OpenEnv Inference Script

bash
python3 inference.py

FinomIQThe future of autonomous financial strategy intelligence.