Auto Byte

专注未来出行及智能汽车科技

微信扫一扫获取更多资讯

Science AI

关注人工智能与其他前沿技术、基础学科的交叉研究与融合发展

微信扫一扫获取更多资讯

机器之心海外团队作者

Synced Global AI Weekly | 2018.5.26—6.1

Europe Does Not Want To Be Left Behind In AIGDPR Explanation vs AI Performance?

European netizens spent yesterday clicking through privacy policy updates for their various subscribed platforms and services. That's because the European Union's new GDPR - billed as the world's toughest data use policy - came into effect today...

For more details


MoD Launches Flagship AI Lab

The government's obsession with everything AI has continued with Defence Secretary Gavin Williamson announcing the launch of a new artificial intelligence flagship hub to enhance UK’s AI, machine learning capabilities...

For more details


Uber's European Rival Taxify Raises $175M Led by Daimler at A $1B Valuation?

There's a new unicorn in the global ride-hailing space after Taxify, a startup born in Estonia that does battle with Uber across Europe and Africa, closed $175 million in new funding that takes it valuation to the $1 billion mark...

For more details


Russian Search Giant Yandex Made Its Own Smart Speaker

The speaker, called the Yandex Station, is the company’s first attempt at hardware, and it’s equipped with a Russian-speaking voice assistant called Alice that was developed in-house...

For more details


Technology

Gym Retro

OpenAI released the full version of Gym Retro, a platform for reinforcement learning research on games. This brings their publicly-released game count from around 70 Atari games and 30 Sega games to over 1,000 games across a variety of backing emulators.

For more details


Learning to Optimize Tensor Programs

Researchers from Allen Institute, University of Washington, Fudan University and SJTU introduce a learning-based framework to optimize tensor programs for deep learning workloads.

For more details


Self-Attention Generative Adversarial Networks

Researchers from Google Brain and Rutgers University propose the Self-Attention Generative Adversarial Network (SAGAN) which allows attention-driven, long-range dependency modeling for image generation tasks.

For more details


You May Also Like...

World No.1 AI Startup SenseTime Gets Additional $620 Million

China's computer vision startup SenseTime today announced it had raised US$620 million in Series C+ funding, a mere 50 days after it closed its Series C funding with US$600 million...

For more details


Andrew Ng, Lior Pachter & Gary Marcus Twitter Joust on AI Radiology

Ng's team at Stanford got the results with a 169-layer convolutional neural network (CNN) trained on the Stanford lab’s new bone X-ray dataset MURA, for “musculoskeletal radiographs” - said to be the world's largest public radiographic image dataset, with 40,561 labeled images...

For more details


Global AI Events

5-6 JunGartner Data & Analytics SummitMumbai, India.
6-7 JunMachine Intelligence SummitHong Kong, China.
11-12 JunCogX London 2018: Festival of All Things AILondon, UK.
12-14 JunAI Summit LondonLondon, UK.
12-13 JunPredictive analytics World Industry 4.0Munich, Germany.
14-15 JunAI & Machine Learning for Clinical TrialPhiladelphia, USA.

Global AI Opportunities

DeepMotion is looking for Deep Learning Engineer, Path Planning Engineer and Automative Control Engineer in Beijing, China.

career@deepmotion.ai


Nullmax needs full-time/intern Deep Learning Engineer, Computer Vision Engineer and C++ Engineer in Shanghai,China.

jobs@nullmax.ai


Preferred Networks is hiring engineers and researchers in their ML/Robotics/Transportation/BioHealth team in Tokyo and Berkeley, CA.

Apply Now


Liulishuo wants Research Scientist, Android Developers, Data Scientists and Enigeers in NLP in San Mateo and Shanghai.

Apply Now



NewsLetter
相关数据
机器学习技术

机器学习是人工智能的一个分支,是一门多领域交叉学科,涉及概率论、统计学、逼近论、凸分析、计算复杂性理论等多门学科。机器学习理论主要是设计和分析一些让计算机可以自动“学习”的算法。因为学习算法中涉及了大量的统计学理论,机器学习与推断统计学联系尤为密切,也被称为统计学习理论。算法设计方面,机器学习理论关注可以实现的,行之有效的学习算法。

规划技术

人工智能领域的「规划」通常是指智能体执行的任务/动作的自动规划和调度,其目的是进行资源的优化。常见的规划方法包括经典规划(Classical Planning)、分层任务网络(HTN)和 logistics 规划。

张量技术

张量是一个可用来表示在一些矢量、标量和其他张量之间的线性关系的多线性函数,这些线性关系的基本例子有内积、外积、线性映射以及笛卡儿积。其坐标在 维空间内,有 个分量的一种量,其中每个分量都是坐标的函数,而在坐标变换时,这些分量也依照某些规则作线性变换。称为该张量的秩或阶(与矩阵的秩和阶均无关系)。 在数学里,张量是一种几何实体,或者说广义上的“数量”。张量概念包括标量、矢量和线性算子。张量可以用坐标系统来表达,记作标量的数组,但它是定义为“不依赖于参照系的选择的”。张量在物理和工程学中很重要。例如在扩散张量成像中,表达器官对于水的在各个方向的微分透性的张量可以用来产生大脑的扫描图。工程上最重要的例子可能就是应力张量和应变张量了,它们都是二阶张量,对于一般线性材料他们之间的关系由一个四阶弹性张量来决定。

计算机视觉技术

计算机视觉(CV)是指机器感知环境的能力。这一技术类别中的经典任务有图像形成、图像处理、图像提取和图像的三维推理。目标识别和面部识别也是很重要的研究领域。

机器人技术技术

机器人学(Robotics)研究的是「机器人的设计、制造、运作和应用,以及控制它们的计算机系统、传感反馈和信息处理」 [25] 。 机器人可以分成两大类:固定机器人和移动机器人。固定机器人通常被用于工业生产(比如用于装配线)。常见的移动机器人应用有货运机器人、空中机器人和自动载具。机器人需要不同部件和系统的协作才能实现最优的作业。其中在硬件上包含传感器、反应器和控制器;另外还有能够实现感知能力的软件,比如定位、地图测绘和目标识别。之前章节中提及的技术都可以在机器人上得到应用和集成,这也是人工智能领域最早的终极目标之一。

生成对抗网络技术

生成对抗网络是一种无监督学习方法,是一种通过用对抗网络来训练生成模型的架构。它由两个网络组成:用来拟合数据分布的生成网络G,和用来判断输入是否“真实”的判别网络D。在训练过程中,生成网络-G通过接受一个随机的噪声来尽量模仿训练集中的真实图片去“欺骗”D,而D则尽可能的分辨真实数据和生成网络的输出,从而形成两个网络的博弈过程。理想的情况下,博弈的结果会得到一个可以“以假乱真”的生成模型。

推荐文章
暂无评论
暂无评论~