DQN-based linear model.

<p>(A) Behavioral data were collected while subjects played the arcade games Breakout (top), Space Invaders (middle), and Enduro (bottom). The human-generated videos were processed by a baseline DQN (gray), Ape-X (green), and SEED (red). Each neuron of the output layer of a DQN corresponds to...

وصف كامل

محفوظ في:

التفاصيل البيبلوغرافية
المؤلف الرئيسي:	Sabine Haberland (22783613) (author)
مؤلفون آخرون:	Hannes Ruge (5815964) (author), Holger Frimmel (22783616) (author)
منشور في:	2025
الموضوعات:	Sociology Science Policy Mental Health Biological Sciences not elsewhere classified xlink "> humans significant performance improvements remained unclear whether recorded motor responses humans transform high appropriate motor responses conventional experimental approach human motor responses playing arcade games grained temporal scale human behavior across dqn &# 8217 continuous visual stimuli rl ), enable compare prediction accuracy modeling human behavior third baseline dqn model human behavior continuous experimental tasks deep learning used human behavior experimental tasks prediction accuracy deep rl baseline dqn human participants human data used features three games temporal resolution improved modeling everyday tasks deep q better prediction varying degrees trial structure thereby opening term memory results suggest response probabilities reinforcement learning long short linear model interesting avenue future research dqns ), double q directed actions considerably improved chance level
الوسوم:	إضافة وسم لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!

الوصف
الملخص:	<p>(A) Behavioral data were collected while subjects played the arcade games Breakout (top), Space Invaders (middle), and Enduro (bottom). The human-generated videos were processed by a baseline DQN (gray), Ape-X (green), and SEED (red). Each neuron of the output layer of a DQN corresponds to the Q-value of a possible type of action, such as ‘no operation’ (‘noop’), move left, move right, fire, and brake. The time series of Q-values were used as predictors in a GLM, while the actual time series of human motor responses served as the predicted variable. For each DQN, the GLMs were fitted to six out of seven sessions, and the Pearson correlation between the predicted human time series and the actual one was calculated on the left-out session. (B) In the preprocessing step, human actions were binary coded and smoothed. A separate GLM was fitted for each type of human action, with the corresponding human time series as the dependent variable.</p>

DQN-based linear model.

مواد مشابهة