深度加強學習（Deep Reinforcement Learning）在計算機視覺領域的前景如何？

01-31

尤其是SLAM和Deep Reinforcement Learning。
SLAM現在貌似到達了一個瓶頸期，都說深度學習是下一個突破點，無論是在優化方向，還是在其與語義信息的結合上。
還有一個疑問是，不太清楚究竟是什麼樣的場景下Reinforcement Learning比監督、非監督學習更加適合？

請問各位有什麼看法？

我覺得RL適合『控制』『決策』等領域，比如機器人。傳統的語音圖像nlp領域還沒想到有什麼特別適合RL的菜。

視覺領域應該是無監督的前景更好，無監督的潛力還有很大。

有一個證據：

如果RL能在傳統領域搞出大新聞，Deepmind那幫開掛的，早就放衛星了！

大部分人對深度強化學習的認識來自於Deepmind在nature上的那篇Atari Game Controller：

Human-level control through deep reinforcement learning : Nature : Nature Research

DRL的原理主要就是用CNN去完成了傳統RL過程中，比如Q-learning中方程估計的的那個部分。看似很暴力的方法，但效果拔群。

在三巨頭合寫的nature綜述里關於Future of deep learning有這麼兩段話:

Unsupervised learning had a catalytic effect in reviving interest in deep learning, but has since been overshadowed by the successes of purely supervised learning. Although we have not focused on it in this Review, we expect unsupervised learning to become far more important in the longer term. Human and animal learning is largely unsupervised: we discover the structure of the world by observing it, not by being told the name of every object.
Human vision is an active process that sequentially samples the optic array in an intelligent, task-specific way using a small, high-resolution fovea with a large, low-resolution surround. We expect much of the future progress in vision to come from systems that are trained end-to- end and combine ConvNets with RNNs that use reinforcement learning to decide where to look. Systems combining deep learning and reinforcement learning are in their infancy, but they already outperform passive vision systems at classification tasks and produce impressive
results in learning to play many different video games.

http://www.nature.com/nature/journal/v521/n7553/full/nature14539.html

綜合大多數人對Supervised learning看了太多數據的詬病，Unsupervised 和 Reinforcement Learning，應該是一個可以值得期待的領域吧。

而且已經有很多機器人控制，和視覺伺服領域的人在向此進發了。

比如：Pieter Abbeel---Associate Professor UC Berkeley---Co-Founder Gradescope---

DRL更適合planning，難點在於state representation space太大。

至於題主問題，每一種演算法都有適用的領域，就好比RNN和CNN，沒有可比性。

個人覺得DRL在cv方向可能會結合無人駕駛來做，而不太容易單純做圖像。

去看看《大漢光武》就知道了