Closed ivan-v-kush closed 5 years ago
Yes, good idea. Added.
one more from 2016 IEEE Conference on Computational Intelligence and Games (CIG)
, don't know if it contains a good scientific idea, but
In this paper, we propose an algorithm to balance exploration and exploitation. Generally, the number of output nodes in neural network of DQL is the same to the actions that a player can perform.
https://cilab.sejong.ac.kr/home/lib/exe/fetch.php?media=public:paper:ieee_cig16_hyunsoo.pdf
Playing Doom with SLAM-Augmented Deep Reinforcement Learning https://arxiv.org/abs/1612.00380
A number of recent approaches to policy learning in 2D game domains have been successful going directly from raw input images to actions. However when employed in complex 3D environments, they typically suffer from challenges related to partial observability, combinatorial exploration spaces, path planning, and a scarcity of rewarding scenarios. Inspired from prior work in human cognition that indicates how humans employ a variety of semantic concepts and abstractions (object categories, localisation, etc.) to reason about the world, we build an agent-model that incorporates such abstractions into its policy-learning framework. We augment the raw image input to a Deep Q-Learning Network (DQN), by adding details of objects and structural elements encountered, along with the agent's localisation. The different components are automatically extracted and composed into a topological representation using on-the-fly object detection and 3D-scene reconstruction.We evaluate the efficacy of our approach in Doom, a 3D first-person combat game that exhibits a number of challenges discussed, and show that our augmented framework consistently learns better, more effective policies.
`
on the site wrong link to Training Agent for First-Person Shooter Game with Actor-Critic Curriculum Learning
should be https://openreview.net/pdf?id=Hk3mPK5gg
It turns out that the link was correct but markdown didn't like the '?id=Hk3mPK5gg' part but now it should work ok.
guys test their method in VizDoom
Curiosity-driven Exploration by Self-supervised Prediction https://arxiv.org/pdf/1705.05363.pdf
Guys you may also like this one paper too, check it out.......https://arxiv.org/pdf/1707.03902.pdf
I think it would be good to add subj
https://openreview.net/pdf?id=Hk3mPK5gg
http://www.aaai.org/ocs/index.php/AAAI/AAAI17/paper/download/14455/14027