Investigating Human Priors for Playing Video Games

By Rachit Dubey, Pulkit Agrawal, Deepak Pathak, T. Griffiths, Alexei A. Efros. 2018

155 citations recorded by the source corpus.

What makes humans so good at solving seemingly complex video games? Unlike computers, humans bring in a great deal of prior knowledge about the world, enabling efficient decision making. This paper investigates the role of human priors for solving video games. Given a sample game, we conduct a series of ablation studies to quantify the importance of various priors on human performance. We do this by modifying the video game environment to systematically mask different types of visual information that could be used by humans as priors. We find that removal of some prior knowledge causes a drastic degradation in the speed with which human players solve the game, e.g. from 2 minutes to over 20 minutes. Furthermore, our results indicate that general priors, such as the importance of objects and visual consistency, are critical for efficient game-play. Videos and the game manipulations are available at this https URL

Subfields: Artificial Intelligence, Sociology and Political Science

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Sociology and Political Science

Citations in the corpus

Exploratory Not Explanatory: Counterfactual Analysis of Saliency Maps for Deep Reinforcement Learning (2019, background): “Current theory and evidence from cognitive science implies that humans learn complex processes, such as video games, by categorizing objects into abstract classes and by inferring causal relationships among instances of those classes (Tenenbaum & Niyogi, 2003; Dubey et al., 2018).”
The Laplacian in RL: Learning Representations with Efficient Approximations (2018, background): “In reinforcement learning (RL), the choice of state representation may affect generalization (Rafols et al., 2005), exploration (Tang et al., 2017; Pathak et al., 2017), and speed of learning (Dubey et al., 2018).”
Toybox: A Suite of Environments for Experimental Evaluation of Deep Reinforcement Learning (2019, background): “Human players rely on unique problem-solving capabilities that deep RL agents have not yet achieved, while deep networks are undeterred by the kind of noise that can confuse humans (Szegedy et al., 2014; Dubey et al., 2018).”
On Inductive Biases in Deep Reinforcement Learning (2019, background): “…is evidence that we, as humans, posses speciﬁc types of biases, and that these have a role in enabling efﬁcient learning (Spelke and Kinzler, 2007; Dubey et al., 2018); however, it is far from clear the extent to which these are essential for intelligent behaviour to arise, what form these priors…”
Task-Agnostic Dynamics Priors for Deep Reinforcement Learning (2019, background): “L G ] 15 J un 2019 of new environments quickly (Dubey et al., 2018) – one such prior is an understanding of physical laws of motion.”
Investigating Simple Object Representations in Model-Free Deep Reinforcement Learning (2020, unknown): “Dubey et al. (2018) explicitly examine the importance of the visual object prior to human and artiﬁcial agents, discovering that the human agents exhibit strong reliance on the objectness of the environment, while deep RL agents suffer no penalty when it is removed.”
Unsupervised Video Object Segmentation for Deep Reinforcement Learning (2018, background): “In contrast, humans immediately recognize salient objects (e.g., spaceships, monsters, balls) and exploit texture cues to guess a strategy [5].”