Synced Global AI Weekly | 2018.7.14-7.20

The Conference Week

Yann LeCun's IJCAI Keynote: We Need A World Model

LeCun proposed that current AI systems are weak in inference, common sense and cognition of the mission background because they lack a “World Model.” Hence, along with attempts at improving memory, researchers must also build up a representation of the real world so AI can better understand the environment and perform more intelligently.

For more details

IJCAI 2018 Kicks Off; DeepMind AlphaGo Wins Marvin Minsky Award 

The 2018 International Joint Conference On Artificial Intelligence (IJCAI) opened today in Stockholm, Sweden, with a record-setting 2,500 attendees gathering at the Stockholm International Fairs convention centre.

For more details

Facebook Researchers Win Test of Teim Award at ICML 2018

The Alan Turing Institute in the UK and France's DATAIA Institute have concluded an arrangement to collaborate in the arena of artificial intelligence and data science, part of a five-year commitment between the two countries to improve digital services.

For more details

SIGIR 2018 Reveals Best Papers; Struggles With Visa Issues

The 41st International ACM SIGIR Conference on Research and Development in Information Retrieval is now in full swing at the University of Michigan campus in Ann Arbor, US - although some leading researchers won't be attending. Yesterday the respected conference presented its Best Paper Awards.

For more details

Baidu's Haifeng Wang Will Chair New ACL Asia-Pacific Chapter

At the 56th annual meeting of the Association for Computational Linguistics (ACL) in Melbourne, Australia this week President Marti Hearst announced the establishment of the Asia-Pacific Chapter of the Association for Computational Linguistics (AACL).

For more details


Announcing Cirq: An Open Source Framework for NISQ Algorithms

Cirq is focused on near-term questions and helping researchers understand whether NISQ quantum computers are capable of solving computational problems of practical importance. Cirq is licensed under Apache 2, and is free to be modified or embedded in any commercial or opensource package.

For more details

OpenAI Five Benchmark

The OpenAI Five Benchmark match will be held 12:30pm Pacific Time on August 5th in San Francisco. The human team will include Blitz, Cap, Fogged, and Merlini, some of whom are former professionals. The games will be streamed on our Twitch channel and casted by popular casters Purge and ODPixel.

For more details

AutoGraph converts Python into TensorFlow graphs

We'd like to tell you about a new TensorFlow feature called "AutoGraph". AutoGraph converts Python code, including control flow, print() and other Python-native features, into pure TensorFlow graph code.

For more details

Theory IIIb: Generalization in Deep Networks

Here we prove a similar result for nonlinear multilayer DNNs near zero minima of the empirical loss. The result holds for exponential-type losses but not for the square loss. In particular, we prove that the weight matrix at each layer of a deep network converges to a minimum norm solution up to a scale factor (in the separable case).

For more details

Global AI Events

5-10 Aug

18-20 JulNucl.aiVienna, Austria
25 JulAnthill InsideBangalore, India
26-27 JulThe Fifth Elephant 2018Bangalore, India
31 JulyAI SummitHong Kong
TDWI Anaheim ConferenceAnaheim, USA
6-10 AugUAI 2018Monterey, USA

Global AI Opportunities

Appier needs Machine Learning Scientist, Data Analyst, Frontend/Backend Engineer and DevOps Engineer in Taipei, Taiwan

Apply Now

Appthority is hiring Senior Software Engineer for Data System, Front-end, Cloud Back-end, Android and iOS in Mountain View, CA.

Apply Now

GoogleAI in Ghana is recruiting Machine Learning Research Scientist, Research Software Engineer and AI researcher.

Apply Now

Socure wants Senior Data Scientist, Quality Assurance Engineer and Senior Software Engineer in Chennai, India and New York.

Apply Now










阿尔法围棋是于2014年开始由英国伦敦Google DeepMind公司开发的人工智能围棋程序。AlphaGo是第一个打败人类职业棋手的计算机程序,也是第一个打败围棋世界冠军的计算机程序,可以说是历史上最强的棋手。 技术上来说,AlphaGo的算法结合了机器学习(machine learning)和树搜索(tree search)技术,并使用了大量的人类、电脑的对弈来进行训练。AlphaGo使用蒙特卡洛树搜索(MCTS:Monte-Carlo Tree Search),以价值网络(value network)和策略网络(policy network)为指导,其中价值网络用于预测游戏的胜利者,策略网络用于选择下一步行动。价值网络和策略网络都是使用深度神经网络技术实现的,神经网络的输入是经过预处理的围棋面板的描述(description of Go board)。


TensorFlow是一个开源软件库,用于各种感知和语言理解任务的机器学习。目前被50个团队用于研究和生产许多Google商业产品,如语音识别、Gmail、Google 相册和搜索,其中许多产品曾使用过其前任软件DistBelief。


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。