Sep 9–12, 2019

Schedule: Implementing AI sessions

Add to your personal schedule
9:00am12:30pm Tuesday, September 10, 2019
Location: LL21 A/B
In this workshop, you will get hands-on experience in developing intelligent AI assistants based entirely on machine learning and using only open source tools - Rasa NLU and Rasa Core. You will learn the fundamentals of conversational AI and the best practices of developing AI assistants that scale and learn from real conversational data. Read more.
Add to your personal schedule
9:00am12:30pm Tuesday, September 10, 2019
Location: LL21 C/D
Skyler Thomas (MapR)
The popular open source Kubeflow project is one of the best ways to start doing machine learning and AI on top of Kubernetes. However, Kubeflow is a huge project with dozens of large complex components. In this hands-on session, we will learn about the Kubeflow components and how they interact with Kubernetes. We explore the machine learning lifecycle from model training to model serving. Read more.
Add to your personal schedule
9:00am12:30pm Tuesday, September 10, 2019
Location: 230 C
Paris Buttfield-Addison (Secret Lab), Tim Nugent (lonely.coffee), Mars Geldard (University of Tasmania)
Are you a scientist who wants to test a research problem without building costly and complicated real-world rigs? A self-driving car engineer who wants to test their AI logic in a constrained virtual world? A data scientist who needs to solve a thorny real-world problem without touching a production environment? Have you considered simulation-driven ML problem solving with a game engine? Read more.
Add to your personal schedule
1:30pm5:00pm Tuesday, September 10, 2019
Location: 230 C
Robert Nishihara (UC Berkeley), Philipp Moritz (UC Berkeley), Ion Stoica (UC Berkeley)
Ray is a general purpose framework for programming your cluster. We will lead a deep dive into Ray, walking you through its API and system architecture and sharing application examples, including several state-of-the-art AI algorithms. Read more.
Add to your personal schedule
1:30pm5:00pm Tuesday, September 10, 2019
Location: LL21 A/B
Neil Conway (Determined AI), Yoav Zimmerman (Determined AI)
Success with deep learning requires understanding more than just TensorFlow or Keras. In this tutorial, we will describe a range of practical problems faced by DL practitioners and the software tools and techniques needed to address them, including data prep/augmentation, GPU scheduling, hyperparameter tuning, distributed training, metrics management, deployment, and mobile/edge optimization. Read more.
Add to your personal schedule
1:30pm5:00pm Tuesday, September 10, 2019
Location: LL21 C/D
Boris Lublinsky (Lightbend), Dean Wampler (Lightbend)
This hands-on tutorial examines production use of ML in streaming data pipelines; how to do periodic model retraining and low-latency scoring in live streams. We'll discuss Kafka as the data backplane, pros and cons of microservices vs. systems like Spark and Flink, tips for Tensorflow and SparkML, performance considerations, model metadata tracking, and other techniques. Read more.
Add to your personal schedule
1:30pm5:00pm Tuesday, September 10, 2019
Location: LL21 E
Joel Grus (Allen Institute for Artificial Intelligence)
This tutorial will briefly discuss what modern neural NLP looks like, after which we'll train some models, write some code, and learn how you can apply these techniques to your own datasets and problems. Read more.
Add to your personal schedule
11:05am11:45am Wednesday, September 11, 2019
Location: 230 C
Jasjeet Thind (Zillow)
Advances in AI & deep learning are enabling new technologies to mimic how the human brain interprets scenes, objects & images. This progress has major implications for businesses that need to extract meaning from overwhelming quantities of unstructured data. In this session, learn how implementing computer vision based in deep neural networks allows machines to “see” images in an entirely new way. Read more.
Add to your personal schedule
11:05am11:45am Wednesday, September 11, 2019
Location: Expo Hall 3
Huaixiu Zheng (Uber)
In this talk, I will cover how Uber applies Deep Learning in the domain of NLP and Conversational AI. In particular, I will go into details of how we implement AI solutions in a real-world environment, as well as cutting edge research we are doing in end-to-end dialogue systems. Read more.
Add to your personal schedule
11:05am11:45am Wednesday, September 11, 2019
Location: LL21 C/D
Danielle Dean (Microsoft), Wee Hyong Tok (Microsoft)
In this session, you will learn best practices and reference architectures (which have been validated in real-world AI/ML projects for customers globally) for implementing AI. Join us in this session as Wee Hyong and Danielle share the lessons distil from working with large global customers on AI/ML projects, and the challenges that they overcome. Read more.
Add to your personal schedule
11:55am12:35pm Wednesday, September 11, 2019
Location: 230 B
Nagendra Shishodia (EXL), Solmaz Torabi (EXL Service), Chaithanya Manda (EXL Service)
Every NLP based document processing solution depends on converting scanned documents/ images to machine readable text using an OCR solution. However, accuracy of OCR solutions is limited by quality of scanned images. We show that generative adversarial networks can be used to bring significant efficiencies in any document processing solution by enhancing resolution and de-noising scanned images. Read more.
Add to your personal schedule
11:55am12:35pm Wednesday, September 11, 2019
Location: Expo Hall 3
Hagay Lupesko (Amazon Web Services)
In this session, you will learn how Lex, Amazon's cloud-based AI-powered chatbot service, was architected, built and deployed. You will learn practical considerations for deploying and maintaining deep learning models in production, and how Lex used Apache MXNet and MXNet Model Server to build and scale the successful service. Read more.
Add to your personal schedule
11:55am12:35pm Wednesday, September 11, 2019
Location: LL21 C/D
Michael Bauer (Sylabs, Inc.)
Containerization technology can be used to build distributed, scalable, and complex neural networks by leveraging decoupled resource pools - pools that would not traditionally be amenable to such a task. Using Singularity, we demonstrate the approach by treating a container as a Decoupled Neural Interface to enable novel applications for neural networks which were previously impractical. Read more.
Add to your personal schedule
11:55am12:35pm Wednesday, September 11, 2019
Location: LL21 E/F
Huaiyu Zhu (IBM Research - Almaden), Dulce Ponceleon (IBM Research - Almaden), Yunyao Li (IBM Research - Almaden)
Natural Language Understanding (NLU) underlies a wide range of applications and services. Rich resources available for English do not exist for most other languages. Is it possible to avoid duplicating the effort? Further, can NLU-dependent applications be developed language-agnostically (write once, applicable to multiple languages)? We will show a vision to answering yes to both questions. Read more.
Add to your personal schedule
1:45pm2:25pm Wednesday, September 11, 2019
Location: 230 B
Akhilesh Kumar (Adobe)
Photographic defects such as noise, exposure(underexposure/overexposure), blur can ruin the perfect shot. We have developed a solution based on GAN that can identify the region of defectiveness in images and fix these defective images. This solution is better than traditional algorithms. It can also be applied to fix videos. No more spending hours of time manually editing the images. Read more.
Add to your personal schedule
1:45pm2:25pm Wednesday, September 11, 2019
Location: Expo Hall 3
Paige Bailey (Google)
TensorFlow 2.0 has landed! During this session, you will learn all about TensorFlow 2.0's new features, usability enhancements, performance increases, and focus on developer productivity. We will use the TF 2.0 migration tool to transition a model from TensorFlow 1.x to 2.0, and deploy an end-to-end open-source machine learning model. Read more.
Add to your personal schedule
1:45pm2:25pm Wednesday, September 11, 2019
Location: LL21 C/D
Roshan Sumbaly (Facebook)
There aren't many systems in the world that need to run hundreds of computer vision models (from classification to segmentation) on billions of visual entities (images, videos, 3D) daily. This talk walks through the challenges we faced while building such a platform and how, surprisingly, a lot of the answers were found in traditional software engineering best practices. Read more.
Add to your personal schedule
2:35pm3:15pm Wednesday, September 11, 2019
Location: LL21 C/D
Evan Sparks (Determined AI)
We describe the current gap between the AI haves: Google, Facebook, Amazon, and Microsoft, and the AI have-nots: the rest of the industry, from the perspective of software infrastructure for model development. We discuss opportunities for end-to-end system design to enable rapid iteration and scale in AI application development. Read more.
Add to your personal schedule
4:00pm4:40pm Wednesday, September 11, 2019
Location: LL21 C/D
Ashish Bansal (Twitter)
This talk gives insight into unique recommendation system challenges at Twitter’s scale and what makes this a fun and challenging task. Read more.
Add to your personal schedule
4:50pm5:30pm Wednesday, September 11, 2019
Location: 230 C
Anuradha Gali (Uber)
Learn how Uber is leveraging AI to automate their business model via their unique platform. You'll hear about their technology that evolves based on current market insights and dynamically adjusts for the future. Anu Gali, Engineering Leader of this platform will discuss best practises and the architecture that enables organizations like Uber grow and scale rapidly. Read more.
Add to your personal schedule
4:50pm5:30pm Wednesday, September 11, 2019
Location: Expo Hall 3
Siddha Ganju (Nvidia), Meher Kasam (Square)
Optimizing deep neural nets to run efficiently on mobile devices. Read more.
Add to your personal schedule
11:05am11:45am Thursday, September 12, 2019
Location: 230 C
Chiranjeet Chetia (Visa Inc.), Carolina Barcenas (Visa)
Artificial intelligence has revolutionized the way we live, work and play. Payments is no exception. With the help of AI, electronic payments have become more secure and convenient for consumers globally — regardless of currency or form factor. In this talk, we explore a use case in which data and deep learning converge to root out malicious actors and make the payments ecosystem more secure. Read more.
Add to your personal schedule
11:55am12:35pm Thursday, September 12, 2019
Location: 230 A
Michael Radwin (Intuit)
Design thinking is a methodology for creative problem solving developed at Stanford University d.school. The methodology is used by world-class design firms like IDEO and many of the world's leading brands like Apple, Google, Samsung, and GE. In this session, Michael Radwin, VP of Data Science at Intuit, will offer a recipe for how to apply design thinking to the development of AI/ML products. Read more.
Add to your personal schedule
11:55am12:35pm Thursday, September 12, 2019
Location: 230 B
Vijay Gabale (Infilect)
Beyond computer games and neural architecture search; practical applications of Deep Reinforcement Learning to improve classical classification or detection tasks are few and far between. In this talk, I will share a technique and our experiences of applying D-RL on improving the distribution input datasets to achieve state of the art performance, specifically on object detection tasks. Read more.
Add to your personal schedule
1:45pm2:25pm Thursday, September 12, 2019
Location: 230 B
Shashank Prasanna (Amazon Web Services)
Machine learning involves a lot of experimentation. Data scientists spend several days, weeks or months performing algorithm search, model architecture search, hyperparameter search etc. In this session, we’ll discuss you how you can easily run large-scale machine learning experiments using containers, kubernetes, Amazon ECS and Sagemaker. Read more.
Add to your personal schedule
1:45pm2:25pm Thursday, September 12, 2019
Location: 230 C
Vijay Agneeswaran (Publicis Sapient), Abhishek Kumar (Publicis Sapient)
We illustrate how capsule networks can be industrialized: 1. Overview of capsule networks and how they help in handling spatial relationships between objects in an image. We also learn about how they can be applied to text analytics. 2. We show an implementation of recurrent capsule networks, which are useful in text analytics, especially for some tasks such as summarization or classification. Read more.
Add to your personal schedule
1:45pm2:25pm Thursday, September 12, 2019
Location: LL21 E/F
Holden Karau (Google), Trevor Grant (IBM)
Modeling is easy- productizing models, less so. Distributed training? forget about it. Hellllllloooo Kubeflow- a system that makes it easy for data scientists who know how to containerize their models, to train and serve on Kubernetes. Read more.
Add to your personal schedule
2:35pm3:15pm Thursday, September 12, 2019
Location: 230 A
This session is a result of lessons learnt from productizing enterprise ML services across Vision, Language, Recommendations, Anomaly Detection over the last 5+ years. You will walk away with an actionable framework to bootstrap & scale a machine learning function. We highlight this via a real product journey involving deep learning that we productized in record speed in-spite of no dataset. Read more.
Add to your personal schedule
2:35pm3:15pm Thursday, September 12, 2019
Location: Expo Hall 3
Brennan Saeta (Google)
Swift for TensorFlow is a next-generation machine learning and differential programming framework that unlocks new domains and applications. This talk will dance through the motivations for Swift, the benefits of this toolchain, and how to use Swift for TensorFlow in your projects. Read more.
Add to your personal schedule
2:35pm3:15pm Thursday, September 12, 2019
Location: LL21 E/F
Manasi Vartak (Verta.AI)
Enterprises are investing heavily in integrating AI/ML into their business, and yet it remains challenging to transform these research-oriented initiatives into revenue driving functions due to a lack of efficient tooling. We discuss key methods that enterprise AI teams can leverage with regards to driving revenue including A/B testing, data pipelines, reproducibility. Read more.
Add to your personal schedule
4:00pm4:40pm Thursday, September 12, 2019
Location: 230 B
Jisheng Wang (Mist Systems)
Increased complexity and business demands continue to make enterprise network operation more challenging. In this talk, we will introduce the architecture of the first autonomous network operation solution together with two examples of ML-driven automated actions. We also share experiences and lessons learned applying ML/DL and AI to the development of SaaS-based enterprise solutions. Read more.
Add to your personal schedule
4:00pm4:40pm Thursday, September 12, 2019
Location: Expo Hall 3
Ting-Fang Yen (DataVisor)
We describe a monitor for production machine learning systems that handle billions of requests daily. Our approach discovers detection anomalies, such as spurious false positives, as well as gradual concept drifts when the model no longer captures the target concept. This session presents new tools for detecting undesirable model behaviors early in large-scale online ML systems. Read more.
Add to your personal schedule
4:00pm4:40pm Thursday, September 12, 2019
Location: LL21 E/F
Kai Liu (Microsoft (BING)), Yuqi Wang (Microsoft), Bin Wang (Microsoft)
FrameworkLauncher is built to orchestrate all kinds of workloads on YARN through the same interface without making changes to the workload themselves. These workloads include but not limited to: Large-Scale Long-Running Services (DeepLearning Serving, HBase, Kafka, etc), Batch Jobs (DeepLearning Training, KDTree Building, etc) and Streaming Jobs (Data Processing, Dynamic Rendering, etc). Read more.
Add to your personal schedule
4:50pm5:30pm Thursday, September 12, 2019
Location: 230 C
Tianchu Liang (American Tire Distributors)
Deep Learning has been a sweeping revolution in the world of AI and machine learning. But how does this new, hot, technology help a legacy business everyday? In this talk, I will go over a warehouse staffing solution we deployed in 140 distribution centers, where I implemented LSTM recurrent neural network model to generate staffing level forecasts and to optimize staffing schedules. Read more.
Add to your personal schedule
4:50pm5:30pm Thursday, September 12, 2019
Location: LL21 E/F
Jonathan Peck (Algorithmia)
We’ll look at why Machine Learning is a natural fit for serverless computing, discuss a general architecture for scalable ML, and cover issues we ran into when implementing our own on-demand scaling over GPU clusters, providing general solutions and a vision for the future of cloud-based ML Read more.

Contact us

confreg@oreilly.com

For conference registration information and customer service

partners@oreilly.com

For more information on community discounts and trade opportunities with O’Reilly conferences

Become a sponsor

For information on exhibiting or sponsoring a conference

Contact list

View a complete list of O'Reilly AI contacts