O’REILLY、INTEL AI主办

English中文
将人工智能用起来
2019年6月18-21日
北京,中国
 
紫金大厅B(Grand Hall B)
13:10
Add Deep prediction: A year in review for deep learning for time series to your personal schedule
14:00 Deep prediction: A year in review for deep learning for time series Aileen Nielsen (Skillman Consulting)
Add ML Ops and Kubeflow Pipeline to your personal schedule
14:50 ML Ops and Kubeflow Pipeline Kaz Sato (Google)
报告厅(Auditorium)
Add AVA: a Cloud-Native Deep Learning Platform at Qiniu to your personal schedule
14:00 AVA: a Cloud-Native Deep Learning Platform at Qiniu Chaoguang Li (Qiniu), Bin Fan (Alluxio)
Add 保险中的机器学习实践 to your personal schedule
16:20 保险中的机器学习实践 鞠芳 (中国人寿研发中心)
多功能厅2(Function Room 2)
Add Achieving Salesforce-Scale Machine Learning in Production to your personal schedule
11:15 Achieving Salesforce-Scale Machine Learning in Production Sarah Aerni (Salesforce Einstein)
13:10
14:00
Add Best practice of building data science platform in Rakuten to your personal schedule
14:50 Best practice of building data science platform in Rakuten 安敖日奇朗 (Rakuten, Inc.), TzuLin Chin (Rakuten, Inc.)
Add Using ML for personalizing Food Recommendations to your personal schedule
16:20 Using ML for personalizing Food Recommendations Maulik Soneji (Go-jek), Jewel James (Go-jek)
多功能厅5A+B(Function Room 5A+B)
Add 自动驾驶技术是如何应用于新潮传媒、新零售行业 to your personal schedule
11:15 自动驾驶技术是如何应用于新潮传媒、新零售行业 Li Yuan (Perceptin 深圳普思英察科技有限公司)
Add How China Telecom combats financial frauds with Adversarial AutoEncoder? to your personal schedule
13:10 How China Telecom combats financial frauds with Adversarial AutoEncoder? Weisheng Xie (China Telecom BestPay Co., Ltd)
Add 基于目标检测的智能化成矿异常信息提取 to your personal schedule
16:20 基于目标检测的智能化成矿异常信息提取 李苍柏 (中国地质科学院矿产资源研究所)
多功能厅6A+B (Function Room 6A+B)
11:15
Add 在边缘实现深度学习 to your personal schedule
13:10 在边缘实现深度学习 陈玉荣 (Intel)
Add PAI Tensor Accelerator and Optimizer---Yet Another Deep Learning Compiler to your personal schedule
14:50 PAI Tensor Accelerator and Optimizer---Yet Another Deep Learning Compiler 杨军 (阿里巴巴), 龙国平 (Alibaba)
Add 快速社交 (Speed Networking) to your personal schedule
08:00 快速社交 (Speed Networking) | Room: 3楼序厅(3rd Floor Foyer)
Add Friday opening remarks to your personal schedule
08:45 Friday opening remarks | Room: 紫金大厅A(Grand Hall A) Ben Lorica (O'Reilly Media), Roger Chen (Computable), Jason (Jinquan) Dai (Intel)
Add Keynotes to come to your personal schedule
08:50 Keynotes to come | Room: 紫金大厅A(Grand Hall A)
10:45 Morning Break | Room: 报告厅序厅 (Auditorium Foyer)
15:30 Afternoon Break | Room: 报告厅序厅 (Auditorium Foyer)
11:55 Lunch | Room: 彩虹厅及国际厅 (Rainbow Room & Ballroom)
11:15-11:55 (40m) 英文讲话 (Presented in English)
Bringing Research And Production Together With PyTorch 1.0
Joseph Spisak (Facebook)
Learn how PyTorch 1.0 enables you to take state-of-the-art research and deploy it quickly at scale in areas from autonomous vehicles to medical imaging. We'll deep dive on the latest updates to the PyTorch framework including TorchScript and the JIT compiler, deployment support, the C++ interface. We will also cover how PyTorch 1.0 is utilized at Facebook to power AI across a variety of products.
13:10-13:50 (40m)
Session
14:00-14:40 (40m) 模型与方法 (Models and Methods)
Deep prediction: A year in review for deep learning for time series
Aileen Nielsen (Skillman Consulting)
Deep learning for time series analysis has made rapid progress in 2018 and 2019, with advances in the use of both convolutional and recurrent neural network architectures. The state of the art in deep forecasting will be summarized for 2018 and 2019, including use cases in both forecasting and generating time series.
14:50-15:30 (40m) 实施人工智能 (Implementing AI), 模型与方法 (Models and Methods), 英文讲话 (Presented in English)
ML Ops and Kubeflow Pipeline
Kaz Sato (Google)
Creating an ML model is just a starting point. To bring the technology into production service, you need to solve various real-world issues such as: building a data pipeline for continuous training, automated validation of the model, version control of the model, scalable serving infra, and ongoing operation of the ML infra with monitoring and alerting.
16:20-17:00 (40m) 执行简报/最佳实践 (Executive Briefing/Best Practices), 文化与组织 (Culture and Organization), 案例研究 (Case Studies), 英文讲话 (Presented in English)
AI at ING: the why, how, and what of a data-driven enterprise
Bas Geerdink (ING)
AI is at the core of ING’s business. We are a data-driven enterprise, with ‘analytics skills’ as a top strategic priority. We are investing in AI, big data, and analytics to improve business processes such as balance forecasting, fraud detection and customer relation management. In this talk, Bas will give an overview of the use cases and technology to inspire the audience!
11:15-11:55 (40m) 实施人工智能 (Implementing AI), 英文讲话 (Presented in English)
Analytics Zoo: Distributed TensorFlow in Production on Apache Spark
Yang Wang (Intel)
We will introduce Analytics Zoo, a unified analytics + AI platform for distributed TensorFlow, Keras and BigDL on Apache Spark, designed for production environment. It enables easy deployment, high performance and efficient model serving for deep learning applications.
13:10-13:50 (40m) 实施人工智能 (Implementing AI)
Sparkling: 基于Apache Spark进行一站式机器学习
Yiheng Wang (Tencent)
机器学习项目在企业中实际落地往往涉及到复杂工作流构建和数据管理,以及多种工具的整合。而且随着数据规模的增加,团队规模的扩大,这一任务更具挑战性。Apache Spark是业界流行的大数据框架,被广泛的应用在海量数据的分析处理。本议题将介绍我们在腾讯云上如何基于Apache Spark为客户建立一个一站式机器学习平台的相关工作。主要内容包括多种数据源的接入,构建复杂数据管线,利用数据可视化理解数据,通过可插拔的机制使用各种流行的机器学习框架,以及部署和监控模型。我们也会分享在这一过程中遇到的问题和挑战。听众也可以了解到,通过这种和大数据紧密结合的一站式机器学习,用户可以怎样更加高效的建立和管理他们的机器学习项目,从而加速了机器学习在业务中的落地。
14:00-14:40 (40m) 实施人工智能 (Implementing AI)
AVA: a Cloud-Native Deep Learning Platform at Qiniu
Chaoguang Li (Qiniu), Bin Fan (Alluxio)
Atlab Lab at Qiniu Cloud focuses on deep learning for computer vision. Our team has built a high-performance and cost-effective training platform based on Cloud for deep learning, called AVA, which deeply integrates open source software stack including Tensorflow, Caffe, Alluxio and KODO our own cloud object storage.
14:50-15:30 (40m) 实施人工智能 (Implementing AI)
Query the planet: Geospatial big data analytics at Uber
Zhenxiao Luo (Uber)
One of the distinct challenges for Uber is analyzing geospatial big data. Locations and trips provide insights that can improve business decisions and better serve users. Geospatial data analysis is particularly challenging, especially in a big data scenario. For these analytical requests, we must achieve efficiency, usability, and scalability in order to meet user needs and business requirements.
16:20-17:00 (40m) 实施人工智能 (Implementing AI)
保险中的机器学习实践
鞠芳 (中国人寿研发中心)
分析保险行业人工智能发展情况及现有数据特性,评估机器学习模型构建的主流工具、语言、算法。总结基于机器学习技术,实现一个保险业人工智能场景的全流程——从场景研讨、数据加工提取到模型构建、模型效果评估、模型落地实施。以一个真实的机器学习模型项目为例,介绍整个方法论不同环节中各方人员的参与工作内容和比例,探讨特征稳定性、样本不均衡、参数选择、模型可解释性等环节的难点及尝试方案。为金融或者其他行业的机器学习项目落地提供参考和指导。
11:15-11:55 (40m) 执行简报/最佳实践 (Executive Briefing/Best Practices), 英文讲话 (Presented in English)
Achieving Salesforce-Scale Machine Learning in Production
Sarah Aerni (Salesforce Einstein)
At Salesforce Einstein data science is an agile partner to over 100,000 customers. How do we achieve this scale? We share lessons learned in business, technology and process along the way. Via use cases, oft-missed foundational elements for deployment, and the evaluations that must happen along the way, we will share how to achieve and sustain models in production, and where to go from there.
13:10-13:50 (40m)
Session
14:00-14:40 (40m) 实施人工智能 (Implementing AI), 模型与方法 (Models and Methods)
Session
14:50-15:30 (40m) 与人工智能交互 (Interacting with AI), 实施人工智能 (Implementing AI), 英文讲话 (Presented in English)
Best practice of building data science platform in Rakuten
安敖日奇朗 (Rakuten, Inc.), TzuLin Chin (Rakuten, Inc.)
Data Science Platform is a suite of tools for exploring data, training models, and running GPU/CPU compute jobs in an isolated container environment. It provides one click machine learning environment creation, powerful job scheduler and flexible "function as a service" component. It runs on Kubernetes and supports both on-premises and cloud environment, as well as hybrid mode.
16:20-17:00 (40m) 模型与方法 (Models and Methods), 英文讲话 (Presented in English)
Using ML for personalizing Food Recommendations
Maulik Soneji (Go-jek), Jewel James (Go-jek)
The story of how we prototyped the search framework that personalizes the restaurant search results by using ML to learn what constitutes a relevant restaurant given a user's purchasing history
11:15-11:55 (40m) 实施人工智能 (Implementing AI)
自动驾驶技术是如何应用于新潮传媒、新零售行业
Li Yuan (Perceptin 深圳普思英察科技有限公司)
如何令自动驾驶技术落地并结合新潮传媒以及新零售业务,相关的技术是如何实现,商业模式是什么以及如何通过人工只能技术提升行业的效率。
13:10-13:50 (40m) 案例研究 (Case Studies)
How China Telecom combats financial frauds with Adversarial AutoEncoder?
Weisheng Xie (China Telecom BestPay Co., Ltd)
We exploit the good representation capability of AAE (Adversarial AutoEncoder) in our risk factors modeling in fighting a special kind of financial frauds. It's one step of our long stack of unsupervised tasks, yet it's proved to be efficient and effective in our practice.
14:00-14:40 (40m) 与人工智能交互 (Interacting with AI), 实施人工智能 (Implementing AI)
人工智能病理影像辅助诊断系统——从方法到落地
王书浩 (透彻影像)
病理学是医学诊断的“金标准”,病理报告对于临床医生提供进一步治疗策略至关重要。一位能够独立发病理报告的病理医师需要10年以上的培养周期,我国目前共有约1万名注册在案的病理医师,根据WHO的要求,人才缺口为4-9万人。使用人工智能来辅助病理医师对样本进行诊断,不仅能够大幅提高医师的诊断效率,而且可以减少漏诊,提高诊断准确率。数字化的病理影像能够观察到组织的细胞形态,在最高倍数字扫描时,文件尺寸达到GB量级,需要从人工智能和系统工程的层面去应对这些挑战。在这个演讲中,我们将从人工智能系统的构建方法入手,介绍透彻影像与中国人民解放军总医院在消化道病理影像辅助系统研发过程中的技术细节。同时,我们将分享诊断系统从部署到落地使用的一些经验。
14:50-15:30 (40m) 人工智能对商业及社会的影响 (Impact of AI on Business and Society)
运用自动化AI技术打击“智能化”网络欺诈
Hongyu Cui (DataVisor)
AI技术在赋能各个产业的同时,也被网络黑产所利用,使得黑产攻击更加自动化,更加隐蔽,难于检测。 DataVisor在互联网反欺诈领域研究发现,目前黑产的攻击模型呈现以下趋势:攻击方法多样化而变化快,攻击手段趋于模拟正常用户,攻击账号主要来源由大规模注册渐渐转向ATO账号。传统的规则系统和有监督的模型,由于对欺诈案例以及标签数据的强依赖,往往无法及时应对迅速演化的黑产攻击,在反欺诈中一直处于被动防守的状态。DataVisor的无监督算法,通过全局分析,在高维空间聚类,可以在无标签情况下,自动发现大规模关联欺诈团伙。无监督算法在提前预警以及检测快速演变欺诈模式方面体现了显著的优势。
16:20-17:00 (40m) 实施人工智能 (Implementing AI)
基于目标检测的智能化成矿异常信息提取
李苍柏 (中国地质科学院矿产资源研究所)
矿床所在的位置往往伴随着地质、地球物理、地球化学、遥感异常,因此,这些异常所在的位置也往往伴随着矿床的存在。所以,在找矿工作当中,一个重要的过程便是在地、物、化、遥数据中寻找异常,并将其整合,得出该区域成矿的概率,从而推断出靶区所在的位置。但传统方法并未考虑空间中点与点之间的相关关系。而卷积神经网络中的卷积和池化方法,充分考虑了点与点之间的相关关系。但单纯使用卷积神经网络只能进行特征提取,不能圈定异常所在的区域。因此,特将目标检测的相关算法引入其中,从而圈定异常所在的区域。
11:15-11:55 (40m)
Session
13:10-13:50 (40m) 模型与方法 (Models and Methods)
在边缘实现深度学习
陈玉荣 (Intel)
深度学习在许多领域尤其是视觉识别/理解方面取得了巨大突破,但它在训练和部署方面都存在一些挑战。本讲座将介绍我们通过高效CNN算法设计、领先DNN模型压缩技术和创新部署时DNN网络结构优化来解决深度学习部署挑战的前沿研究成果。
14:00-14:40 (40m) 模型与方法 (Models and Methods)
非监督学习在大规模图谱上的案例应用和开源算法剖析
Mingxi Wu (TigerGraph)
图数据上的非监督学习在激活大数据的经济价值上有着广泛和不可替代的作用。 PageRank能够发掘重要的实体, 社区发掘(community detection)可以找到具有某种特性的群体,紧密度中心性算法(Closeness Centrality)可以自动找到远离群体的个体。所有这些算法都是非监督的学习。 我们分享一些具体客户案例来展示他们的价值,同时分享怎样在大数据上灵活应用这些开源算法。
14:50-15:30 (40m) 与人工智能交互 (Interacting with AI), 实施人工智能 (Implementing AI)
PAI Tensor Accelerator and Optimizer---Yet Another Deep Learning Compiler
杨军 (阿里巴巴), 龙国平 (Alibaba)
本次演讲会介绍阿里计算平台PAI团队过去一年多时间里在深度学习编译器领域的技术工作进展----PAI TAO(Tensor Accelerator and Optimizer)。PAI-TAO采用通用编译优化技术,来解决PAI平台所承载的多样性AI workload面临的训练及推理需求的性能优化问题,在部分workload上获得了20%到4X不等的显著加速效果,并且基本作到用户层全透明,在显著提升平台效率性能的同时也有效照顾了用户的使用惯性。目前PAI-TAO已经先后用于支持阿里内部搜索、推荐、图像、文本等多个业务场景的日常训练及推理需求。
16:20-17:00 (40m) 实施人工智能 (Implementing AI)
Low precision inference on Intel Architecture
Lei Xia (Intel)
Vector Neural Network Instructions or VNNI is the new Intel instruction set for low precision AI inference inside next generation Xeon platform. This lecture is to introduce the features of the VNNI and Intel software tools to support developers to use this new instruction set to accelerate inference with INT8.
08:00-08:30 (30m)
快速社交 (Speed Networking)
在本次人工智能大会上与寻求联系的与会者会面。会议将在周四主题演讲之前举行一个非正式的快速社交活动。一定要带上自己的名片来享受社交活动。
08:45-08:50 (5m)
Friday opening remarks
Ben Lorica (O'Reilly Media), Roger Chen (Computable), Jason (Jinquan) Dai (Intel)
Opening keynote remarks by Program Chairs Ben Lorica, Jason Dai, and Roger Chen
08:50-10:40 (1h 50m)
Keynotes to come
Keynotes to come
10:45-11:15 (30m)
Break: Morning Break
15:30-16:20 (50m)
Break: Afternoon Break
11:55-13:10 (1h 15m)
Break: Lunch