Intelligent Autonomous Trading Systems - Are We There Yet?
summary
Argues that fully autonomous, set-and-forget trading is not yet practical and gives concrete successes and limits: satellite imaging for crop acreage car counts and oil tank volumes produces tradable signals but everyone now uses it; human-scored model training replaces brittle metric formulas and speeds selection; neural-network derivative pricers with Monte Carlo fallback yield fast, pragmatic pricing. Lists key failures: reinforcement learning overfits, reward functions mislead, noise and hidden liquidity break models, trading costs and latency often erase academic gains.
AI-written summary · about
chapters (5)
- 0:00Start
- 1:25Obvious Answer
- 11:04Human-trained machines
- 14:17Derivative Pricing
- 16:49What about reinforcement learning
related talks
55mAlgorithmic Trading and Machine Learning
13mArtificial General Intelligence: Why aren't we there yet?
26mWhen Crypto Becomes Intelligent
30mHow to build a stock market money printing machine with AI
27mThe Double-Edged Sword of Forex Trading: Maximizing Returns while Managing Risk
15mAuto-Adaptive Systems: Changing the Game for Marketplace Management
19mNow in AI - How We Got Here
59mAI in Finance: Scope and Examples