User contributions for Felipefelixarias

00:3300:33, 9 September 2024 diff hist +169‎ N Translations:Diffusion Models Are Real-Time Game Engines/110/zh ‎ Created page with "* Ho & Salimans (2022) Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance, 2022. URL: [https://arxiv.org/abs/2207.12598](https://arxiv.org/abs/2207.12598)." current
00:3300:33, 9 September 2024 diff hist +230‎ N Translations:Diffusion Models Are Real-Time Game Engines/109/zh ‎ Created page with "* Hafner et al. (2020) Danijar Hafner, Timothy Lillicrap, Jimmy Ba, and Mohammad Norouzi. Dream to control: Learning behaviors by latent imagination, 2020. URL: [https://arxiv.org/abs/1912.01603](https://arxiv.org/abs/1912.01603)." current
00:3300:33, 9 September 2024 diff hist +79‎ N Translations:Diffusion Models Are Real-Time Game Engines/108/zh ‎ Created page with "* Ha & Schmidhuber (2018) David Ha and Jürgen Schmidhuber. World models, 2018." current
00:3300:33, 9 September 2024 diff hist +267‎ N Translations:Diffusion Models Are Real-Time Game Engines/107/zh ‎ Created page with "* Gupta et al. (2023) Agrim Gupta, Lijun Yu, Kihyuk Sohn, Xiuye Gu, Meera Hahn, Li Fei-Fei, Irfan Essa, Lu Jiang, and José Lezama. Photorealistic video generation with diffusion models, 2023. URL: [https://arxiv.org/abs/2312.06662](https://arxiv.org/abs/2312.06662)." current
00:3300:33, 9 September 2024 diff hist +329‎ N Translations:Diffusion Models Are Real-Time Game Engines/106/zh ‎ Created page with "* Girdhar et al. (2023) Rohit Girdhar, Mannat Singh, Andrew Brown, Quentin Duval, Samaneh Azadi, Sai Saketh Rambhatla, Akbar Shah, Xi Yin, Devi Parikh, and Ishan Misra. Emu video: Factorizing text-to-video generation by explicit image conditioning, 2023. URL: [https://arxiv.org/abs/2311.10709](https://arxiv.org/abs/2311.10709)." current
00:3200:32, 9 September 2024 diff hist −446‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "我们注意到，类似于 NVidia 的经典 SLI 交替帧渲染（AFR）技术，通过在额外硬件上并行生成多个帧，可以显著提高图像生成速率。然而，与 AFR 类似，实际的仿真速率不会提高，输入延迟也不会减少。"
00:3200:32, 9 September 2024 diff hist +530‎ N Translations:Diffusion Models Are Real-Time Game Engines/105/zh ‎ Created page with "* Bruce et al. (2024) Jake Bruce, Michael Dennis, Ashley Edwards, Jack Parker-Holder, Yuge Shi, Edward Hughes, Matthew Lai, Aditi Mavalankar, Richie Steigerwald, Chris Apps, Yusuf Aytar, Sarah Bechtle, Feryal Behbahani, Stephanie Chan, Nicolas Heess, Lucy Gonzalez, Simon Osindero, Sherjil Ozair, Scott Reed, Jingwei Zhang, Konrad Zolna, Jeff Clune, Nando de Freitas, Satinder Singh, and Tim Rocktäschel. Genie: Generative interactive environments, 2024. URL: [https://arxiv..." current
00:3200:32, 9 September 2024 diff hist +396‎ N Translations:Diffusion Models Are Real-Time Game Engines/104/zh ‎ Created page with "* Brooks et al. (2024) Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luhman, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. Video generation models as world simulators, 2024. URL: [https://openai.com/research/video-generation-models-as-world-simulators](https://openai.com/research/video-generation-models-as-world-simulators)." current
00:3200:32, 9 September 2024 diff hist +298‎ N Translations:Diffusion Models Are Real-Time Game Engines/103/zh ‎ Created page with "* Blattmann et al. (2023b) Andreas Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja Fidler, and Karsten Kreis. Align your latents: High-resolution video synthesis with latent diffusion models, 2023b. URL: [https://arxiv.org/abs/2304.08818](https://arxiv.org/abs/2304.08818)." current
00:3200:32, 9 September 2024 diff hist +373‎ N Translations:Diffusion Models Are Real-Time Game Engines/102/zh ‎ Created page with "* Blattmann et al. (2023a) Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram Voleti, Adam Letts, Varun Jampani, and Robin Rombach. Stable video diffusion: Scaling latent video diffusion models to large datasets, 2023a. URL: [https://arxiv.org/abs/2311.15127](https://arxiv.org/abs/2311.15127)." current
00:3200:32, 9 September 2024 diff hist +399‎ N Translations:Diffusion Models Are Real-Time Game Engines/101/zh ‎ Created page with "* Bar-Tal et al. (2024) Omer Bar-Tal, Hila Chefer, Omer Tov, Charles Herrmann, Roni Paiss, Shiran Zada, Ariel Ephrat, Junhwa Hur, Guanghui Liu, Amit Raj, Yuanzhen Li, Michael Rubinstein, Tomer Michaeli, Oliver Wang, Deqing Sun, Tali Dekel, and Inbar Mosseri. Lumiere: A space-time diffusion model for video generation, 2024. URL: [https://arxiv.org/abs/2401.12945](https://arxiv.org/abs/2401.12945)." current
00:3200:32, 9 September 2024 diff hist +200‎ N Translations:Diffusion Models Are Real-Time Game Engines/100/zh ‎ Created page with "* Alonso et al. (2024) Eloi Alonso, Adam Jelley, Vincent Micheli, Anssi Kanervisto, Amos Storkey, Tim Pearce, and François Fleuret. Diffusion for world modeling: Visual details matter in Atari, 2024." current
00:3200:32, 9 September 2024 diff hist +187‎ N Translations:Diffusion Models Are Real-Time Game Engines/99/zh ‎ Created page with "* Akenine-Möller et al. (2018) Tomas Akenine-Möller, Eric Haines, and Naty Hoffman. ''Real-Time Rendering, Fourth Edition''. A. K. Peters, Ltd., USA, 4th edition, 2018. ISBN 0134997832." current
00:3200:32, 9 September 2024 diff hist +18‎ N Translations:Diffusion Models Are Real-Time Game Engines/98/zh ‎ Created page with "== 参考文献 ==" current
00:3100:31, 9 September 2024 diff hist −151‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "我们将代理生成的数据训练与使用随机策略生成的数据训练进行比较。对于随机策略，我们根据与观测结果无关的均匀分类分布对动作进行采样。我们通过对两个模型及其解码器进行"
00:3100:31, 9 September 2024 diff hist +106‎ N Translations:Diffusion Models Are Real-Time Game Engines/97/zh ‎ Created page with "如需联系，请发送邮件至 <code>shlomif@google.com</code> 和 <code>leviathan@google.com</code>。" current
00:3100:31, 9 September 2024 diff hist +607‎ N Translations:Diffusion Models Are Real-Time Game Engines/96/zh ‎ Created page with "* '''Dani Valevski''': 开发了大部分代码库，调整了整个系统的参数和细节，增加了自动编码器微调、代理训练和蒸馏功能。 * '''亚尼夫·列维坦''': 提出了项目、方法和架构，开发了初始实现，是实现和撰写的主要贡献者。 * '''Moab Arar''': 领导了自回归稳定化与噪声增强，进行了许多消融实验，并创建了人类游戏数据的数据集。 * '''Shlomi Fruchter''': 提出了项目、方..." current
00:3100:31, 9 September 2024 diff hist −123‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "== 致谢 =="
00:3100:31, 9 September 2024 diff hist +12‎ N Translations:Diffusion Models Are Real-Time Game Engines/95/zh ‎ Created page with "== 贡献 ==" current
00:3000:30, 9 September 2024 diff hist +290‎ N Translations:Diffusion Models Are Real-Time Game Engines/94/zh ‎ Created page with "我们衷心感谢 Eyal Segalis、Eyal Molad、Matan Kalman、Nataniel Ruiz、Amir Hertz、Matan Cohen、Yossi Matias、Yael Pritch、Danny Lumen、Valerie Nygaard、Theta Labs 和 Google Research 团队以及我们的家人，感谢他们提供的深刻反馈、创意、建议和支持。" current
00:3000:30, 9 September 2024 diff hist −1,030‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "=== 5.2 消融实验 ==="
00:3000:30, 9 September 2024 diff hist +12‎ N Translations:Diffusion Models Are Real-Time Game Engines/93/zh ‎ Created page with "== 致谢 ==" current
00:3000:30, 9 September 2024 diff hist +1,252‎ N Translations:Diffusion Models Are Real-Time Game Engines/92/zh ‎ Created page with "'''迈向互动视频游戏的新范式''' 如今，视频游戏是由人类''编程''的。而''GameNGen''则是新范式的一部分概念验证，在这一新范式中，游戏被视为神经模型的权重，而不是代码行。''GameNGen''展示了一种架构和模型权重，使神经模型能够有效地在现有硬件上互动运行复杂的游戏（如 DOOM）。虽然仍有许多重要问题，但我们希望这种范式能带来显著益处。例如，在这..." current
00:3000:30, 9 September 2024 diff hist +859‎ N Translations:Diffusion Models Are Real-Time Game Engines/91/zh ‎ Created page with "'''未来工作''' 我们在经典游戏《DOOM》上演示了''GameNGen''。在其他游戏或更广泛的交互式软件系统上进行测试会非常有趣。我们注意到，除了强化学习代理的奖励函数外，我们的技术没有任何内容是《DOOM》特有的。我们计划在未来的工作中解决这个问题。虽然''GameNGen''能够准确地维护游戏状态，但如上文所述，它并不完美。可能需要一个更复杂的架构来缓解..." current
00:3000:30, 9 September 2024 diff hist +1,104‎ N Translations:Diffusion Models Are Real-Time Game Engines/90/zh ‎ Created page with "'''局限性。''' GameNGen 存在内存有限的情况。该模型只能访问稍超过3秒的历史记录，因此许多游戏逻辑能够在更长的时间跨度内被保存，这一点令人惊讶。虽然部分游戏状态是通过屏幕像素（如弹药和健康统计、可用武器等）来持久化的，但模型可能学习了强大的启发式方法，从而能够进行有意义的概括。例如，从渲染视图中，模型可以学习推断玩家的位置..." current
00:3000:30, 9 September 2024 diff hist +210‎ N Translations:Diffusion Models Are Real-Time Game Engines/89/zh ‎ Created page with "'''总结。''' 我们介绍了''GameNGen''，并证明在神经模型上可以实现每秒20帧的高质量实时游戏。我们还提供了将计算机游戏等交互式软件转换为神经模型的方法。" current
00:3000:30, 9 September 2024 diff hist +14‎ N Translations:Diffusion Models Are Real-Time Game Engines/88/zh ‎ Created page with "== 7 讨论 ==" current
00:3000:30, 9 September 2024 diff hist +534‎ N Translations:Diffusion Models Are Real-Time Game Engines/87/zh ‎ Created page with "DOOM 于 1993 年发布，掀起了游戏行业的一场革命。它引入了开创性的 3D 图形技术，成为第一人称射击类游戏的基石，影响了无数其他游戏。许多研究工作都对 DOOM 进行了研究。它提供了开放源码的实现和足够低的原生分辨率，适合小型模型的模拟，同时也足够复杂，可以作为一个具有挑战性的测试案例。最后，作者在这款游戏上花费了无数的青春时光，因此..." current
00:3000:30, 9 September 2024 diff hist −329‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "Stable Diffusion v1.4 的预训练自动编码器将 8x8 像素块压缩为 4 个潜通道，在预测游戏帧时会导致有意义的伪影，影响小细节，尤其是底栏 HUD（“抬头显示”）。为了在提高图像质量的同时利用预训练的知识，我们仅使用针对目标帧像素计算的 MSE 损失来训练潜在自动编码器的解码器。使用 LPIPS（Zhang 等人（[https://arxiv.org/html/2408.14837v1#bib.bib40 2018]））等感知损失..."
00:3000:30, 9 September 2024 diff hist +16‎ N Translations:Diffusion Models Are Real-Time Game Engines/86/zh ‎ Created page with "===== DOOM =====" current
00:3000:30, 9 September 2024 diff hist +2,404‎ N Translations:Diffusion Models Are Real-Time Game Engines/85/zh ‎ Created page with "有几项研究试图利用动作输入来训练游戏仿真模型。Yang 等人（[https://arxiv.org/html/2408.14837v1#bib.bib38 2023]）建立了一个包含真实世界和模拟视频的多样化数据集，并训练了一个扩散模型，根据前一个视频片段和动作的文字描述来预测后续视频。Menapace 等人（[https://arxiv.org/html/2408.14837v1#bib.bib18 2021]）和 Bruce 等人（[https://arxiv.org/html/2408.14837v1#bib.bib7 2024]）专注于..." current
00:3000:30, 9 September 2024 diff hist −143‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "仅使用 4 个去噪步骤导致 U-Net 总耗时为 40 毫秒（包括自动编码器的推理总耗时为 50 毫秒），即每秒 20 帧。我们推测，在我们的案例中，较少步骤对质量影响可忽略不计，是由于以下因素的结合：(1) 受限的图像空间，以及 (2) 前一帧的强条件作用。"
00:3000:30, 9 September 2024 diff hist +39‎ N Translations:Diffusion Models Are Real-Time Game Engines/84/zh ‎ Created page with "===== 游戏模拟与世界模型 =====" current
00:2900:29, 9 September 2024 diff hist +1,106‎ N Translations:Diffusion Models Are Real-Time Game Engines/83/zh ‎ Created page with "扩散模型在文本到图像生成中取得了最先进的成果（Saharia 等人，[https://arxiv.org/html/2408.14837v1#bib.bib27 2022]；Rombach 等人，[https://arxiv.org/html/2408.14837v1#bib.bib26 2022]；Ramesh 等人，[https://arxiv.org/html/2408.14837v1#bib.bib25 2022]；Podell 等人，[https://arxiv.org/html/2408.14837v1#bib.bib23 2023]），这一研究领域也被应用于文本到视频生成任务（Ho 等人，[https://arxiv.org/html/2408.14837v1#bib.b..." current
00:2900:29, 9 September 2024 diff hist −459‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "重建三维表示的神经方法在过去几年中取得了重大进展。NeRFs（Mildenhall 等人，[https://arxiv.org/html/2408.14837v1#bib.bib20 2020]）使用深度神经网络对辐射场进行参数化，该网络针对从不同相机姿态拍摄的一组图像的特定场景进行了专门优化。训练完成后，可通过体积渲染方法对场景的新视角进行采样。Gaussian Splatting（Kerbl 等人，[https://arxiv.org/html/2408.14837v1#bib.bib15 202..."
00:2900:29, 9 September 2024 diff hist +30‎ N Translations:Diffusion Models Are Real-Time Game Engines/82/zh ‎ Created page with "===== 视频扩散模型 =====" current
00:2900:29, 9 September 2024 diff hist +742‎ N Translations:Diffusion Models Are Real-Time Game Engines/81/zh ‎ Created page with "重建三维表示的神经方法在过去几年中取得了重大进展。NeRFs（Mildenhall 等人，[https://arxiv.org/html/2408.14837v1#bib.bib20 2020]）使用深度神经网络对辐射场进行参数化，该网络针对从不同相机姿态拍摄的一组图像的特定场景进行了专门优化。训练完成后，可通过体积渲染方法对场景的新视角进行采样。Gaussian Splatting（Kerbl 等人，[https://arxiv.org/html/2408.14837v1#bib.bib15 202..." current
00:2900:29, 9 September 2024 diff hist +30‎ N Translations:Diffusion Models Are Real-Time Game Engines/80/zh ‎ Created page with "===== 神经三维仿真 =====" current
00:2900:29, 9 September 2024 diff hist +1,012‎ N Translations:Diffusion Models Are Real-Time Game Engines/79/zh ‎ Created page with "模拟二维和三维环境的视觉和物理过程，并允许对其进行交互式探索，是计算机图形学中一个广泛发展的领域（Akenine-Möller等人，[https://arxiv.org/html/2408.14837v1#bib.bib1 2018]）。像虚幻和Unity这样的游戏引擎是可以处理场景几何表示并根据用户交互渲染图像流的软件。游戏引擎负责跟踪所有世界状态，例如玩家的位置和移动、物体、角色动画和光照。它还负责跟..." current
00:2900:29, 9 September 2024 diff hist −60‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "视频质量我们使用第[https://arxiv.org/html/2408.14837v1#S2 2]节中描述的自回归设置，按照真实轨迹所定义的动作序列对帧进行迭代采样，同时将模型自身的过往预测作为条件。自回归采样时，预测轨迹和真实轨迹常常在几步后发生偏离，这主要是由于不同轨迹的帧间积累了少量不同的运动速度。因此，如图[https://arxiv.org/html/2408.14837v1#S5.F6 6]所示，每帧的PSNR和LPIPS值..."
00:2900:29, 9 September 2024 diff hist +33‎ N Translations:Diffusion Models Are Real-Time Game Engines/78/zh ‎ Created page with "===== 交互式三维仿真 =====" current
00:2800:28, 9 September 2024 diff hist −445‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "为了消除噪声增强的影响，我们训练了一个不添加噪声的模型。我们对标准噪声增强模型和不添加噪声的模型（经过 200,000 步训练后）进行自回归评估，并计算在随机保留的 512 条轨迹上预测帧与真实帧之间的 PSNR 和 LPIPS 指标。我们在图 [https://arxiv.org/html/2408.14837v1#S5.F7 7] 中报告了每个自回归步骤的平均指标值，最多可达 64 帧。"
00:2800:28, 9 September 2024 diff hist +20‎ N Translations:Diffusion Models Are Real-Time Game Engines/77/zh ‎ Created page with "== 6 相关工作 ==" current
00:2800:28, 9 September 2024 diff hist +582‎ N Translations:Diffusion Models Are Real-Time Game Engines/76/zh ‎ Created page with "{| class="wikitable" ! 难度级别 ! 数据生成策略 ! PSNR <math>\uparrow</math> ! LPIPS <math>\downarrow</math> |- | 简单 | 代理 | <math>20.94 \pm 0.76</math> | <math>0.48 \pm 0.01</math> |- | | 随机 | <math>20.20 \pm 0.83</math> | <math>0.48 \pm 0.01</math> |- | 中等 | 代理 | <math>20.21 \pm 0.36</math> | <math>0.50 \pm 0.01</math> |- | | 随机 | <math>16.50 \pm 0.41</math> | <math>0.59 \pm 0.01</math> |- | 困难 | 代理 | <math>17.51 \pm 0.35</math..." current
00:2800:28, 9 September 2024 diff hist +307‎ N Translations:Diffusion Models Are Real-Time Game Engines/75/zh ‎ Created page with "'''表 2：不同难度级别的表现。''' 我们比较了使用代理生成数据和随机生成数据训练的模型在简单、中等和困难数据集上的表现。简单和中等数据集各有 112 个样本，困难数据集有 232 个样本。在 3 秒后的单帧上计算每个轨迹的指标。" current
00:2800:28, 9 September 2024 diff hist +968‎ N Translations:Diffusion Models Are Real-Time Game Engines/74/zh ‎ Created page with "总体而言，我们观察到在随机轨迹上训练模型的效果出奇地好，但受到随机策略探索能力的限制。在比较单帧生成时，代理的效果稍好，PSNR 为 25.06，而随机策略为 24.42。在比较 3 秒自回归生成后的帧时，差距增大到 19.02 对 16.84。在手动操作模型时，我们发现某些区域对两者都很容易，而某些区域对两者都很困难，而在某些区域，代理的表现要好得多。基于..." current
00:2800:28, 9 September 2024 diff hist −980‎ Diffusion Models Are Real-Time Game Engines/zh ‎ Created page with "我们通过速度参数化训练模型，使得扩散损失最小化（Salimans & Ho, [https://arxiv.org/html/2408.14837v1#bib.bib29 2022b]）："
00:2800:28, 9 September 2024 diff hist +240‎ N Translations:Diffusion Models Are Real-Time Game Engines/73/zh ‎ Created page with "我们将代理生成的数据训练与使用随机策略生成的数据训练进行比较。对于随机策略，我们根据与观测结果无关的均匀分类分布对动作进行采样。我们通过对两个模型及其解码器进行" current
00:2800:28, 9 September 2024 diff hist +28‎ N Translations:Diffusion Models Are Real-Time Game Engines/72/zh ‎ Created page with "==== 5.2.3 代理执行 ====" current
00:2800:28, 9 September 2024 diff hist +268‎ N Translations:Diffusion Models Are Real-Time Game Engines/71/zh ‎ Created page with "center|thumb|600x600px|图 7：噪声增强的影响。图中显示了每个自回归步骤的 PSNR 平均值（越高越好）。不使用噪声增强时，质量在 10-20 帧后迅速下降。噪声增强可以防止这种情况。" current

9 September 2024