Breaking News

Razer Unveils the Ultra-Lightweight DeathAdder V4 Pro Sony launches a high-resolution shotgun microphone with superior sound quality and compact design. Arctic announces New Liquid Freezer III Pro 280 and Pro 420 Silicon Power Launches Hypera microSDXC Express Card Samsung announces Watch8, Z Fold7 and Z Flip7

logo

  • Share Us
    • Facebook
    • Twitter
  • Home
  • Home
  • News
  • Reviews
  • Essays
  • Forum
  • Legacy
  • About
    • Submit News

    • Contact Us
    • Privacy

    • Promotion
    • Advertise

    • RSS Feed
    • Site Map

Search form

DeepMind Researchers Create Deep RL Agent That Outperforms Humans in the Atari Human Benchmark

DeepMind Researchers Create Deep RL Agent That Outperforms Humans in the Atari Human Benchmark

Enterprise & IT Mar 31,2020 0

Researchers at Alphabet's DeepMind have developed Agent57, a deep reinforcement learning (RL) agent that outperforms the standard human benchmark on 57 Atari games.

The Atari57 suite of games is a long-standing benchmark to gauge agent performance across a wide range of tasks. This benchmark was proposed to test general competency of RL algorithms. Previous work has achieved good average performance by doing outstandingly well on many games of the set, but very poorly in several of the most challenging games.

Researchers have developed Agent57, the first deep reinforcement learning agent to obtain a score that is above the human baseline on all 57 Atari 2600 games. Agent57 combines an algorithm for efficient exploration with a meta-controller that adapts the exploration and long vs. short-term behaviour of the agent.


To achieve this result, the researchers train a neural network which parameterizes a family of policies ranging from very exploratory to purely exploitative. They propose an adaptive mechanism to choose which policy to prioritize throughout the training process.

Additionally, the researchers utilize a novel parameterization of the architecture that allows for more consistent and stable learning.

DeepMind's Agent57 builds on our previous agent Never Give Up, and instantiates an adaptive meta-controller that helps the agent to know when to explore and when to exploit, as well as what time-horizon it would be useful to learn with. A wide range of tasks will naturally require different choices of both of these trade-offs, therefore the meta-controller provides a way to dynamically adapt such choices.

According to DeepMind's detailed blog post and this paper, Agent57 was able to scale with increasing amounts of computation: the longer it trained, the higher its score got.

However, while this enabled Agent57 to achieve strong general performance, it takes a lot of computation and time; the data efficiency can certainly be improved.

Additionally, this agent shows better 5th percentile performance on the set of Atari57 games. This by no means marks the end of Atari research, not only in terms of data efficiency, but also in terms of general performance.

The researchers offer two views on this: firstly, analyzing the performance among percentiles gives them new insights on how general algorithms are. While Agent57 achieves strong results on the first percentiles of the 57 games and holds better mean and median performance than previous agents NGU or R2D2, it could still obtain a higher average performance.

Secondly, all current algorithms are far from achieving optimal performance in some games. To that end, key improvements to use might be enhancements in the representations that Agent57 uses for exploration, planning, and credit assignment.

Tags: deepmindArtificial Intelligence
Previous Post
Fitbit Charge 4 Comes with GPS, Spotify and Heart Metrics
Next Post
Xiaomi Fourth Quarter Revenue Jumps 27%

Related Posts

  • What Is Explainable AI?

  • Fujitsu AI-Video Recognition Technology Promotes Hand Washing Etiquette and Hygiene in the Workplace

  • PAC-MAN Recreated with AI by NVIDIA Researchers

  • Chinese Sogou Introduces 3D AI News Anchor

  • Microsoft Announces New AI Supercomputer

  • Sony and Microsoft to Create AI-powered Smart Cameras

  • Researchers Use Analog AI hardware to Support Deep Learning Inference Without Great Accuracy

  • Nvidia Unveils New Ampere Data Center Chips, Ampere Computers, and More

Latest News

Razer Unveils the Ultra-Lightweight DeathAdder V4 Pro
PC components

Razer Unveils the Ultra-Lightweight DeathAdder V4 Pro

Sony launches a high-resolution shotgun microphone with superior sound quality and compact design.
Cameras

Sony launches a high-resolution shotgun microphone with superior sound quality and compact design.

Arctic announces New Liquid Freezer III Pro 280 and Pro 420
Cooling Systems

Arctic announces New Liquid Freezer III Pro 280 and Pro 420

Silicon Power Launches Hypera microSDXC Express Card
Cameras

Silicon Power Launches Hypera microSDXC Express Card

Samsung announces Watch8, Z Fold7 and Z Flip7
Smartphones

Samsung announces Watch8, Z Fold7 and Z Flip7

Popular Reviews

be quiet! Light Loop 360mm

be quiet! Light Loop 360mm

be quiet! Dark Mount Keyboard

be quiet! Dark Mount Keyboard

be quiet! Light Mount Keyboard

be quiet! Light Mount Keyboard

Noctua NH-D15 G2

Noctua NH-D15 G2

Soundpeats Pop Clip

Soundpeats Pop Clip

be quiet! Light Base 600 LX

be quiet! Light Base 600 LX

Crucial T705 2TB NVME White

Crucial T705 2TB NVME White

be quiet! Pure Base 501

be quiet! Pure Base 501

Main menu

  • Home
  • News
  • Reviews
  • Essays
  • Forum
  • Legacy
  • About
    • Submit News

    • Contact Us
    • Privacy

    • Promotion
    • Advertise

    • RSS Feed
    • Site Map
  • About
  • Privacy
  • Contact Us
  • Promotional Opportunities @ CdrInfo.com
  • Advertise on out site
  • Submit your News to our site
  • RSS Feed