×

Loading...
Ad by
  • 最优利率和cashback可以申请特批,好信用好收入offer更好。请点链接扫码加微信咨询,Scotiabank -- Nick Zhang 6478812600。
Ad by
  • 最优利率和cashback可以申请特批,好信用好收入offer更好。请点链接扫码加微信咨询,Scotiabank -- Nick Zhang 6478812600。

微软靠着云和这一波AI翻身,特别是他对OpenAI的投资以及chatgpt和自身产品的绑定。谷歌LLM的技术路径和GPT不一样,虽然都是基于谷歌transformer思想。Google DeepMind’s 前两天发布Genie,一个基于世界模型可以自己创造游戏的AI, 侧重点不是视觉而是逻辑的合理性

Genie是什么? 根据官方Google DeepMind博客文章,Genie是一个基础世界模型,它是在从互联网获取的视频上进行训练的。该模型能够“从合成图像、照片,甚至草图中生成各种可玩的(可操作的)世界。”

研究论文《Genie: 生成交互式环境》指出,Genie是第一个通过未标记的互联网视频以无监督方式训练的生成交互式环境。在尺寸方面,Genie包括了110亿参数,并且由一个时空视频标记器、自回归动态模型以及一个简单且可扩展的潜在行动模型组成。

这些技术规格使得Genie能够在生成的环境中以逐帧的方式行动,即使在没有训练、标签或其他特定领域要求的情况下也能够实现。

What is Genie?
According to the official Google DeepMind blog post, Genie is a foundation world model that is trained on videos sourced from the Internet. The model can “generate an endless variety of playable (action-controllable) worlds from synthetic images, photographs, and even sketches.”

The research paper ‘Genie: Generative Interactive Environments’ states that Genie is the first generative interactive environment that has been trained in an unsupervised manner from unlabelled internet videos. When it comes to size, Genie stands at 11B parameters and consists of a spatiotemporal video tokenizer, an autoregressive dynamics model, and a simple and scalable latent action model.

These technical specifications let Genie act in generated environments on a frame-by-frame basis even in the absence of training, labels, or any other domain-specific requirements.

Report

Replies, comments and Discussions:

  • 枫下茶话 / 工商经济 / 劈柴悬了? +1
    Helios Capital founder Samir Arora has said Google's CEO Sundar Pichai will either be fired or resign due to controversies surrounding Google's Gemini AI. Responding to social media inquiries, Arora criticized Pichai's handling of AI advancements. Gemini, Google's AI chatbot, faces scrutiny for alleged bias against PM Modi. India's Ministry of Electronics and Information Technology may also issue a notice to Google over these concerns.
    • 一心垄断市场,缺乏创新
    • 领导谷歌从AI的老大沦为AI的笑料,是够废柴的。 +5
      • 当初阿尔法狗下围棋可谓一鸣惊人。。。 +1
    • 谷歌这两年天天玩政治挂帅,有神马能让人记住的新功能和新产品吗? +3
      • 谷歌这几年最大的贡献大概就是这篇文章 chatgpt model built on the transformer architecture, +1
        designed to generate text in a conversational style. The transformer architecture is a type of deep learning model introduced in the paper "Attention is All You Need" by Vaswani et al. in 2017.
        • 谷歌颠覆性的创新不少,可惜赚不到啥钱。能躺着赚钱赚钱的业务又不给力。
          • 微软靠着云和这一波AI翻身,特别是他对OpenAI的投资以及chatgpt和自身产品的绑定。谷歌LLM的技术路径和GPT不一样,虽然都是基于谷歌transformer思想。Google DeepMind’s 前两天发布Genie,一个基于世界模型可以自己创造游戏的AI, 侧重点不是视觉而是逻辑的合理性

            Genie是什么? 根据官方Google DeepMind博客文章,Genie是一个基础世界模型,它是在从互联网获取的视频上进行训练的。该模型能够“从合成图像、照片,甚至草图中生成各种可玩的(可操作的)世界。”

            研究论文《Genie: 生成交互式环境》指出,Genie是第一个通过未标记的互联网视频以无监督方式训练的生成交互式环境。在尺寸方面,Genie包括了110亿参数,并且由一个时空视频标记器、自回归动态模型以及一个简单且可扩展的潜在行动模型组成。

            这些技术规格使得Genie能够在生成的环境中以逐帧的方式行动,即使在没有训练、标签或其他特定领域要求的情况下也能够实现。

            What is Genie?
            According to the official Google DeepMind blog post, Genie is a foundation world model that is trained on videos sourced from the Internet. The model can “generate an endless variety of playable (action-controllable) worlds from synthetic images, photographs, and even sketches.”

            The research paper ‘Genie: Generative Interactive Environments’ states that Genie is the first generative interactive environment that has been trained in an unsupervised manner from unlabelled internet videos. When it comes to size, Genie stands at 11B parameters and consists of a spatiotemporal video tokenizer, an autoregressive dynamics model, and a simple and scalable latent action model.

            These technical specifications let Genie act in generated environments on a frame-by-frame basis even in the absence of training, labels, or any other domain-specific requirements.

            • 大概率openai 会推出自己的genie, openai 原创能力远不如谷歌,但工程能力很强。他的大语言模型,dalle, sora, 都是基于谷歌人的文章。
    • 谷歌全员裁撤Austin YouTube music 部门。做得实在是不好看。 +1
    • 拜托赶快裁了他吧,谷歌股票万绿丛中一点红好几天了,真是急人