Research talk: Evaluating human-like navigation in 3D video games
On the path to developing agents that learn complex human-like behavior, a key challenge is the need to quickly and accurately quantify human-likeness. While human assessments of such behavior can be highly accurate, speed and scalability are limited. The researchers address these limitations through a novel automated Navigation Turing Test (NTT) that learns to predict human judgments of human-likeness. They demonstrate the effectiveness of their automated NTT on a navigation task in a complex 3D environment. They investigated six classification models to shed light on the types of architectures best suited to this task, and they validated them against data collected through a human NTT. The best models achieve high accuracy when distinguishing true human and agent behavior. At the same time, the researchers show that predicting finer-grained human assessment of agents’ progress towards human-like behavior remains unsolved. Their work takes an important step towards agents that more effectively learn complex human-like behavior.
Learn more about the 2021 Microsoft Research Summit: https://Aka.ms/researchsummit (opens in new tab)
- Évènement :
- Microsoft Research Summit 2021
- Piste :
- Reinforcement Learning
- Date:
- Haut-parleurs:
- Raluca Georgescu, Ida Momennejad
- Affiliation:
- Microsoft Research Cambridge, Microsoft Research NYC
-
-
Raluca Stevenson
Senior Research Scientist
-
Ida Momennejad
Principal Researcher in Reinforcement Learning
-
-
Reinforcement Learning
-
-
-
Research talk: Reinforcement learning with preference feedback
Speakers:- Aadirupa Saha
-
-
-
-
-
-
Panel: Generalization in reinforcement learning
Speakers:- Mingfei Sun,
- Roberta Raileanu,
- Harm van Seijen
-
-
Research talk: Successor feature sets: Generalizing successor representations across policies
Speakers:- Kiante Brantley
-
Research talk: Towards efficient generalization in continual RL using episodic memory
Speakers:- Mandana Samiei
-
Research talk: Breaking the deadly triad with a target network
Speakers:- Shangtong Zhang
-
Panel: The future of reinforcement learning
Speakers:- Geoff Gordon,
- Emma Brunskill,
- Craig Boutilier
-