Transforming how we interact with digital environments with more realistic teammates and opponents, and with richer experiences. Leading the way with ethical AI that protects the interests of AI users and creators. Exploring the relationship between AI and the physical world. Uncovering where AI inspires unprecedented creativity.

Sony AI was established in April 2020 to pursue groundbreaking research in AI and robotics to unleash human imagination and creativity with AI. We believe in AI that empowers the imagination and creativity of artists, makers, scientists and creators around the world.

Join us to empower imagination and creativity through AI. Sony AI is eager to work with outstanding AI, robotics and computer vision researchers and engineers from around the world, as well as those who excel in their own specialized fields.

2022年2月11日 · GT Sophy 是一种自主型AI，利用索尼 AI、PDI 和 SIE 合作开发的新型深度强化学习平台进行训练。 各方通过汇集基础 AI 研发方面的专业知识、超逼真的真实世界赛车模拟器和大规模 AI 培训的基础设施，为GT Sophy项目的成功做出了贡献。

2019年11月21日 · Sony AI将结合世界一流的基础研发与索尼独有的技术资产，特别是在影像和传感解决方案、机器人技术和娱乐（游戏、音乐和电影）方面，将全面推动现有业务领域的转型，并为创建新业务领域做出贡献。

2019年11月20日 · Sony AI will combine world class fundamental research and development with Sony's unique technical assets, especially in Imaging & Sensing Solutions, Robotics and Entertainment (Games, Music and Movies), driving transformation across all existing business domains and contributing to the creation of new business domains.

SonyAI | 22,943 followers on LinkedIn. Sony AI is Sony's flagship AI organization whose mission is to "Unleash Human Imagination and Creativity with AI."

2023年4月6日 · 索尼研究公司成立之初将专注于传感、ai和虚拟空间领域。 成立于2020年的索尼AI公司将并入索尼研究公司，索尼研究公司将在其已有成果的基础上进一步扩大研发范围，开拓更多的研究领域，应对更大的挑战。

2020年8月31日 · Sony AIは、ゲームやイメージング＆センシング、音楽や映画、エレクトロニクス事業などソニーのアセットを活用したプロジェクトに取り組むことで、AIの研究開発においてグローバルで存在感を発揮していきます。

Unified framework enables diverse samplers and 1-step generation SOTAs. Enhancing GAN with metrizable discriminators. Fast, Efficient, Training-Free, and Controllable diffusion-based generation method. Generalizing hierarchical VQ-VAEs with a Bayesian framework. Improving density estimation of diffusion. Achieving blind inversion using DDPM.