Benefits Of The Monte Carlo Method

1524 Words4 Pages

In the Bonawitz paper, she argues that children learn causation through the Monte Carlo method. This is similar to a Bayesian inference in that the posterior hypothesis is proportional to the likelihood that the data would occur if the hypothesis is true and prior probability of the hypothesis. However, instead of sampling all the possible hypotheses that could explain the observed data children randomly sample a few. The chance that a hypothesis is randomly sampled is according to the likelihood it is correct. This is called the Monte Carlo method also known as the sampling hypothesis. This way the brain does less computation to get to the same answer. However, it may lead to the wrong answer occasionally due to the random sampling nature. …show more content…

In the short wait the children had two trials back to back; while in the long wait condition there was a two week period between the two trials. Children were presented an assorted of red and blue chips in an 80:20 ratio and were both of which activated a machine. Chips were placed in the bag and it tip over into the machine. Children were asked which of the chips activated the machine. The results showed that the group in the long wait condition guesses as to what chip more closely reflected the proportion of red chip to blue chips. This supports that independence between sampling is necessary for accurate probability matching. Next this experiment was expanding in way to test the noisy maximizing theory. The children were presented three different “conditions”: a 95:5 condition, a 75:25 condition and a 50:50 condition. If noisy maximizing theory were true then children would have the same response for the 95:5 condition and the 75:25 condition. This is because in both conditions children processing at ceiling levels which in this case is set to anything about ~72%. The results show that children’s guess about the red chip had a linear relationship with the proportion of the red chip to the blue chip. This showed that children were not in fact using this strategy to solve causation problems leaving naïve frequency matching and sampling hypothesis. The third experiment was the same as the second only with three conditions instead of two. So the children had three possible hypotheses with differing probabilities. The results showed that the children’s response can still be predicted via the sampling hypothesis when there are multiple choices. The final experiment tested whether children use sampling hypothesis or naïve frequency matching. Children were presented two bags one had a red blue ratio of 14:6 and the second bag had a red blue

Open Document