Log in to bookmark your favorites and sync them to your phone or calendar.

Technical Lab Sessions [clear filter]
Friday, January 25


Building Scalable Framework and Environment of Reinforcement Learning
Deep Reinforcement Learning (DRL) has made strong progress in many tasks that are traditionally considered to be difficult, such as complete information games, navigation, architecture search, etc. Although the basic principle of DRL is quite simple and straightforward, to make it work often requires substantial efforts, compared to traditional supervised training. In this talk, we introduce our recent open-sourced ELF platform: efficient, lightweight and flexible frameworks to facilitate DRL research. We show the scalability of our platforms by reproducing and open sourcing AlphaGoZero/AlphaZero framework using 2000 GPUs and 1.5 weeks, achieving super-human performance of Go AI that beats 4 top-30 professional players with 20-0. We also show usability of our platform by training agents in real-time strategy games with only a small amount of resource. The trained agent develops interesting tactics and is able to beat rule-based AIs by a large margin. On the environment side, we propose House3D that makes multi-room navigation easy with fast frame rate. With House3D, we show that model-based agent that plans ahead with uncertain information navigate in unseen environments more successfully.

avatar for Yuandong Tian, Facebook AI Research (FAIR)

Yuandong Tian, Facebook AI Research (FAIR)

Research Scientist and Manager, Facebook AI Research (FAIR)
Yuandong Tian is a Research Scientist and Manager in Facebook AI Research, working on deep reinforcement learning and its applications in games, and theoretical analysis of deep models. He is the lead scientist and engineer for ELF Platform for Reinforcement Learning, OpenGo and DarkForest... Read More →

Friday January 25, 2019 9:10am - 9:55am
Pacific G-H Hyatt Regency San Francisco, 5 Embarcadero Center, San Francisco, CA 94111, USA


Building On-prem GPU Training Infrastructure
You'll learn how to provide your team with GPU training infrastructure at a variety of scales, from a single shared multi-GPU system to a cluster for distributed training.

avatar for Stephen Balaban, Lambda Labs

Stephen Balaban, Lambda Labs

Co-Founder & CEO, Lambda Labs
Stephen Balaban, is the co-founder and CEO of Lambda. Lambda provides GPU servers, workstations and cloud services for training neural networks. Lambda customers include Apple, Tencent, Intel, Microsoft, Harvard, Princeton, Stanford, and the U.S. Federal Government. Prior to Lambda... Read More →

Friday January 25, 2019 10:30am - 11:15am
Pacific G-H Hyatt Regency San Francisco, 5 Embarcadero Center, San Francisco, CA 94111, USA


End-to-End Continuous Machine Learning in Production with PipelineAI, Spark ML, TensorFlow AI, PyTorch, Kafka, TPUs, and GPUs
Traditional machine learning pipelines end with life-less models sitting on disk in the research lab.  These traditional models are typically trained on stale, offline, historical batch data. Static models and stale data are not sufficient to power today's modern, AI-first Enterprises that require continuous model training, continuous model optimizations, and lightning-fast model experiments directly in production. Through a series of open source, hands-on demos and exercises, we will use PipelineAI to breathe life into these models using 4 new techniques that we’ve pioneered:
* Continuous Validation (V)
* Continuous Optimizing (O)
* Continuous Training (T)
* Continuous Explainability (E).

The Continuous "VOTE" techniques has proven to maximize pipeline efficiency, minimize pipeline costs, and increase pipeline insight at every stage from continuous model training (offline) to live model serving (online.)
Attendees will learn to create continuous machine learning pipelines in production with PipelineAI, TensorFlow, and Kafka.

A laptop is required for this session. 

avatar for Chris Fregly, PipelineAI

Chris Fregly, PipelineAI

Founder, PipelineAI
Chris Fregly is Founder at PipelineAI, a Real-Time Machine Learning and Artificial Intelligence Startup based in San Francisco. He is also an Apache Spark Contributor, a Netflix Open Source Committer, founder of the Global Advanced Spark and TensorFlow Meetup, author of the O’Reilly... Read More →

Friday January 25, 2019 11:15am - 12:15pm
Pacific G-H Hyatt Regency San Francisco, 5 Embarcadero Center, San Francisco, CA 94111, USA


Imagination-Based AI: Envision the Future Beyond DL 
In this workshop, we explore an emerging frontier of AI and machine learning research that goes beyond correlational data mining approaches, such as deep learning. In a large variety of challenging business and societal problems, from combating traffic congestion in urban metropolises, and delivering better quality healthcare and education, to improving the success rate of digital marketing and online campaigns, and modeling interventions that combat climate change, what is fundamentally necessary is integrating three types of reasoning by combining correlational, causal, and imagination-based models. Participants in this workshop will be introduced to a powerful class of modeling languages that extend deep learning through a three layer architecture that combines correlational (layer one), causal and interventional decisions (layer two), as well as counterfactual and imagination-based reasoning (layer three).

Most of deep learning is currently restricted to producing a statistical summarization of data, that is, they answer “What is” queries. For example, generative adversarial networks (GANs) can be trained to produce new samples from a given fixed distribution (e.g, new face images given a dataset of face images). We introduce novel and powerful new methods that can also answer “What-if” queries, as well as “Why” queries. Integrating the three fundamental modes of reasoning, from “What is” to “What if” and “Why”, will enable the next generation of AI systems to be far more powerful than current deep learning architectures, like GANs. The workshop presentation will include a detailed overview of the underlying modeling languages, as well as hands-on demonstrations of how to use these new emerging tools in a diverse range of practical real-world problems.

avatar for Sridhar Mahadevan, Adobe Research

Sridhar Mahadevan, Adobe Research

Director of Data Science, Adobe Research
Sridhar Mahadevan is currently Director of Data Science at Adobe Research in San Jose, and holds faculty appointments at Stanford Univeristy (visiting professor in the School of Engineering) and a Full Professor at the College of Information and Computer Sciences at the University... Read More →

Friday January 25, 2019 1:00pm - 3:00pm
Pacific G-H Hyatt Regency San Francisco, 5 Embarcadero Center, San Francisco, CA 94111, USA