Reverse Hypothesis Machine Learning by Parag Kulkarni

Reverse Hypothesis Machine Learning by Parag Kulkarni

Author:Parag Kulkarni
Language: eng
Format: epub, pdf
Publisher: Springer International Publishing, Cham


2.Sometimes environment response comes very late and Temporal Difference (TD) methods are not effective

3.Crowd intelligence in action but no over dependence

4.Learning and crowd sourcing are separated and hence it has knowledge acquisition as well as learning

5.Crowd Sourcing can be very effective in real-time applications

6.Crowd sourcing is simply used for action selection and reward calculation and hence learning can become more powerful.

Further overall outcome of the decision is not based on a single decision or move. For example while playing basketball, basket may result due to a series of good moves. It is about coordination and randomness of environment too. It is not just about a single right or wrong move, not even about simply couple of right or wrong moves. It may be outcome of a series of moves, dominated by good moves with reference to opposite players’ positions, and position of your team mates can result in basket. It is rather appropriateness and goodness of the move in that scenario and its possible impact on the immediate next move and apparently final outcome that helps in learning and decision-making. Here, all decisions are impacting the moves that are not independent and their mannerism is defined with reference to environment. The key aspects of these types of applications whether it is Basketball, football, or even some business process are the roles of environment, appropriateness of action, measurement of goodness of rewards, and the feedback. How one can measure appropriateness and goodness of a move. It can be measured using different mechanisms:1.Cumulative reward calculations



Download



Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.