OPTIMIZING EXPECTATIONS: FROM DEEP REINFORCEMENT LEARNING TO STOCHASTIC COMPUTATION GRAPHS

PDF Publication Title:

OPTIMIZING EXPECTATIONS: FROM DEEP REINFORCEMENT LEARNING TO STOCHASTIC COMPUTATION GRAPHS ( optimizing-expectations-from-deep-reinforcement-learning-to- )

Previous Page View | Next Page View | Return to Search List

Text from PDF Page: 052

Total num. policy params Vine: Sim. steps per iter. SP: Sim. steps per iter. Policy iter. Stepsize (DKL ) Discount (γ) Vine: rollouts per state Vine: computation time SP: computation time All games 33500 400K 100K 500 0.01 0.99 ≈4 ≈ 30 hrs ≈ 30 hrs 3.15 learning curves for the atari domain 44 Table 3: Parameters used for Atari domain. 3.15 learning curves for the atari domain 400 600 800 1000 1200 1400 1600 beam rider 0 5 10 15 20 25 30 35 40 45 0 vine 1000 2000 3000 4000 5000 6000 7000 breakout 100 single path vine 0 100 200 300 400 500 enduro single path vine single path vine 0 100 200 300 400 500 number of policy iterations 600 0 100 200 300 400 500 0 100 200 300 400 500 number of policy iterations qbert single path vine 8000 0 100 200 300 400 500 0 100 200 300 400 500 0 100 200 300 400 500 30 20 10 0 10 20 30 pong number of policy iterations seaquest single path vine single path 500 0 500 1000 1500 2000 number of policy iterations number of policy iterations number of policy iterations 100 200 300 400 500 600 space invaders vine single path cost cost cost cost cost cost cost 0 100 200 number of policy iterations 500 Figure 6: Learning curves for the Atari domain. For historical reasons, the plots show cost = negative reward. 300 400

PDF Image | OPTIMIZING EXPECTATIONS: FROM DEEP REINFORCEMENT LEARNING TO STOCHASTIC COMPUTATION GRAPHS

PDF Search Title:

OPTIMIZING EXPECTATIONS: FROM DEEP REINFORCEMENT LEARNING TO STOCHASTIC COMPUTATION GRAPHS

Original File Name Searched:

thesis-optimizing-deep-learning.pdf

DIY PDF Search: Google It | Yahoo | Bing

Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info

Cruising Review Topics and Articles More Info

Software based on Filemaker for the travel industry More Info

The Burgenstock Resort: Reviews on CruisingReview website... More Info

Resort Reviews: World Class resorts... More Info

The Riffelalp Resort: Reviews on CruisingReview website... More Info

CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com (Standard Web Page)