Presented By O’Reilly and Intel AI
Put AI to work
Sep 4-5, 2018: Training
Sep 5-7, 2018: Tutorials & Conference
San Francisco, CA

Speaker slides & video

Presentation slides will be made available after the session has concluded and the speaker has given us the files. Check back if you don't see the file you're looking for—it might be available later! (However, please note some speakers choose not to share their presentations.)

If you are looking for slides and video from 2017, visit the AI Conference 2017 site.

Ofer Ronen (Chatbase)
For developers building a bot or virtual agent, the critical question is which bot to build and why? Today, most can’t answer it without a manual intent discovery process, largely based on guesswork, that uncovers only a percentage of possible opportunities. Ofer Ronen demonstrates techniques, based on machine learning, for faster, more efficient intent discovery.
David Patterson (UC Berkeley)
High-level, domain-specific languages and architectures and freeing architects from the chains of proprietary instruction sets will usher in a new golden age. David Patterson explains why, despite the end of Moore’s law, he expects an outpouring of codesigned ML-specific chips and supercomputers that will improve even faster than Moore’s original 1965 prediction.
KC Tung (Microsoft)
KC Tung explains why LSTM provides great flexibility to model the consumer touchpoint sequence problem in a way that allows just-in-time insights about an advertising campaign's effectiveness across all touchpoints (channels), empowering advertisers to evaluate, adjust, or reallocate resources or investments in order to maximize campaign effectiveness.
Huma Abidi (Intel)
Huma Abidi discusses the importance of optimization to deep learning frameworks and shares Xeon performance results and work that Intel is doing with its framework partners, such TensorFlow.
Ankit Jain (Uber)
Personalization is a common theme in social networks and ecommerce businesses. Personalization at Uber involves an understanding of how each driver and rider is expected to behave on the platform. Ankit Jain explains how Uber employs deep learning using LSTMs and its huge database to understand and predict the behavior of each and every user on the platform.
Dawn Song (UC Berkeley)
Dawn Song details challenges and exciting new opportunities at the intersection of AI and security and explains how AI and deep learning can enable better security and how security can enable better AI. You'll learn about secure deep learning and approaches to ensure the integrity of decisions made by deep learning.
Varun Arora (Baidu USA)
We haven't figured out how to make the perfect robot tutors. But we have figured out how make them much more effective in improving student learning outcomes with modern AI techniques. Varun Arora covers some of those important techniques, along with real-world examples.
Jake Saper (Emergence Capital)
Much attention in enterprise AI today is focused on automation. Jake Saper explains why the more interesting applications focus on worker augmentation and offers an overview of coaching networks, which gather data from a distributed network of workers and identify the best techniques for getting things done.
Daniel Whitenack (Pachyderm)
Kubernetes—the container orchestration engine used by all of the top technology companies—was built from the ground up to run and manage highly distributed workloads on huge clusters. Thus, it provides a solid foundation for model development. Daniel Whitenack demonstrates how to easily deploy and scale AI/ML workflows on any infrastructure using Kubernetes.
Julie Shin Choi (Intel AI), Ariel Pisetzky (Taboola)
In this keynote, Julie Choi explores three real-world use cases featuring a diverse set of data centric problems and Intel solutions. Julie also welcomes Ariel Pisetzky, the Vice President of IT at Taboola to discuss how AI is transforming their business as they create personalized content through their predictive recommendation engine.
Kristian Hammond (Northwestern Computer Science)
Even as AI technologies move into common use, many enterprise decision makers remain baffled about what the different technologies actually do and how they can be integrated into their businesses. Rather than focusing on the technologies alone, Kristian Hammond provides a practical framework for understanding your role in problem solving and decision making.
Kai-Fu Lee (Sinovation Ventures)
With the US leading the AI revolution for decades, it would almost seem inconceivable that China could catch up. But China is rapidly catching up with the US in AI applications. Kai-Fu Lee discusses the five key factors that enabled this rapid ascension: tenacious entrepreneurs, speed and execution, tremendous capital, a big market and bigger data, and techno-utilitarian government policies.
Joseph Sirosh (Microsoft)
Will artificial intelligence revolutionize prosthetic care and assistance? Join Microsoft’s Joseph Sirosh for an intriguing story on AI-infused prosthetics that are able to see, grip, and feel and discover how this is enabling affordable and functional prosthetic care.
Levent Besik (Google)
Levent Besik explains how can enterprises stay ahead of the game with customized ML in our ever-changing world of AI capabilities and limited data science resources.
Avesh Singh (Cardiogram), Kevin Wu (Cardiogram)
Deep learning is often called a black box, so how do you diagnose and fix problems in a deep neural network (DNN)? Avesh Singh and Kevin Wu explain how they systematically debugged DeepHeart, a DNN that detects cardiovascular disease from heart rate data. You'll leave with an arsenal of tools for debugging DNNs, including Jacobian analysis, TensorBoard, and DNN unit tests.
Ting-Fang Yen (DataVisor)
Online fraud is often orchestrated by organized crime rings, who use malicious user accounts to actively target modern online services for financial gain. Ting-Fang Yen shares a real-time, scalable fraud detection solution backed by deep learning and built on Spark and TensorFlow and demonstrates how the system outperforms traditional solutions such as blacklists and machine learning.
Chris Butler (IPSoft)
Purpose, a well-defined problem, and trust from people are important factors to any system, especially those that employ AI. Chris Butler leads you through exercises that borrow from the principles of design thinking to help you create more impactful solutions and better team alignment.
Mary Wahl (Microsoft), Banibrata De (Microsoft)
High-resolution land cover maps help quantify long-term trends like deforestation and urbanization but are prohibitively costly and time intensive to produce. Mary Wahl and Banibrata De demonstrate how to use Microsoft’s Cognitive Toolkit and Azure cloud resources to produce land cover maps from aerial imagery by training a semantic segmentation DNN—both on single VMs and at scale on GPU clusters.
Simon Crosby (SWIM Inc.)
Simon Crosby details an architecture for learning on time series data using edge devices, based on the distributed actor model. This approach flies in the face of the traditional wisdom of cloud-based, big-data solutions to ML problems. You'll see that there are more than enough resources at “the edge” to cost-effectively analyze, learn from, and predict from streaming data on the fly.
Vikram Saletore (Intel), Lucas Wilson (Dell EMC)
Vikram Saletore and Luke Wilson discuss a collaboration between SURFSara and Intel to advance the state of large-scale neural network training on Intel Xeon CPU-based servers, highlighting improved time to solution on extended training of pretrained models and exploring how various storage and interconnect options lead to more efficient scaling.
Jian Wu (NIO)
Jian Wu discusses an end-to-end engineering project to train and evaluate deep Q-learning models for targeting sequential marketing campaigns using the 10-fold cross-validation method. Jian also explains how to evaluate trained DQN models with neural network-based baseline models and shows that trained deep Q-learning models generally produce better-optimized long-term rewards.
Sharad Gupta (Blue Shield of California)
AI-powered chatbots are increasingly becoming viable solutions for customer service use cases. Technology leaders must consider adopting a multichannel chatbot strategy to avoid siloed chatbot solutions. Sharad Gupta shares a framework to ensure long-term strategic investment in chatbots.
Paco Nathan (
Deep learning works well when you have large labeled datasets, but not every team has those assets. Paco Nathan offers an overview of active learning, an ML variant that incorporates human-in-the-loop computing. Active learning focuses input from human experts, leveraging intelligence already in the system, and provides systematic ways to explore and exploit uncertainty in your data.
Amanda Casari (SAP Concur)
Data-driven companies making intelligent products must design for security and privacy to be competitive globally. Amanda Casari details the high-level changes that EU General Data Protection Regulation (GDPR)-compliant businesses face and how this translates to teams designing products driven by machine learning and artificial intelligence.
Armen Donigian (ZestFinance)
What does it mean to explain a machine learning model, and why is it important? Armen Donigian addresses those questions while discussing several modern explainability methods, including traditional feature contributions, LIME, and DeepLift. Each of these techniques offers a different perspective, and their clever application can reveal new insights and solve business requirements.
Mayank Kejriwal (USC Information Sciences Institute)
Human trafficking is a form of modern-day slavery. Online sex advertisement activity on portals like Backpage provide important clues that, if harnessed and analyzed at scale, can help resource-strapped law enforcement crack down on trafficking activity. Mayank Kejriwal details an AI architecture called DIG that law enforcement have used (and are using) to combat sex trafficking.
Kai-Fu Lee (Sinovation Ventures), Tim O'Reilly (O'Reilly Media)
Fireside chat with Tim O'Reilly and Kai-Fu Lee
AI is real and has immense value potential for enterprises. However, there is a lot of hype and confusion around AI, creating a critical need for every business to be thoughtful about developing the right strategy and vision for AI within the organization. Join Manish Goyal to explore four success factors for an AI journey and learn how you can best unlock the value of enterprise AI.
Joshua Dillon (Google Research), Wahid Bhimji (NERSC)
Join in for two talks on TensorFlow in space and mathematics. Josh Dillon discusses TensorFlow Probablity (TFP), and Wahid Bhimji discusses deep learning for fundamental sciences using high-performance computing.
David Kearns (IBM), Ari Kaplan (Aginity), Erin Ledell (, Christopher Coad (Aginity)
Join Ari Kaplan, Erin LeDell, Chris Coad, and David Kearns to see where AI meets business intelligence, as they explore the latest ML technologies and concepts powering today's decisions, including Hortonworks, Aginity Amp,, IBM Data Science Experience, and more—using real-life baseball data to illustrate the concepts.
Rachael Rekart (Autodesk )
Rachael Rekart offers an overview of Autodesk Virtual Agent (AVA), which has revolutionized the way Autodesk approaches customer service. Customers chat with AVA as they would a human, in natural language, and AVA processes transactions quickly, returns accurate answers, or gathers information to pass to a human counterpart to resolve the query.
Ramesh Sridharan (Captricity)
Captricity has deployed a machine learning pipeline that can read handwriting at human-level accuracy. Ramesh Sridharan discusses the big ideas the company learned building and deploying this system, using data to identify specific problems to solve using AI and to evaluate and validate the algorithm itself and the overall system once deployed.
Danielle Dean (Microsoft), Wee Hyong Tok (Microsoft)
Transfer learning enables you to use pretrained deep neural networks and adapt them for various deep learning tasks (e.g., image classification, question answering, and more). Join Wee Hyong Tok and Danielle Dean to learn the secrets of transfer learning and discover how to customize these pretrained models for your own use cases.
Michelle Casbon (Google)
Michelle Casbon offers an overview of Kubeflow. By providing a platform that reduces variability between services and environments, Kubeflow enables applications that are more robust and resilient, resulting in less downtime, quality issues, and customer impact. It also supports the use of specialized hardware such as GPUs, which can reduce operational costs and improve model performance.
There are immense opportunities to apply machine learning to investment management if you know where to look. Unlike many Silicon Valley challenges, it's not simply a matter of throwing capital and PhDs at the financial markets. Michael Weinberg explains why you must exploit domain expertise to achieve disruptive success.
To help enterprises formulate their strategies for actionable and effective use of AI, HfS and IBM have jointly developed a practical guide to starting your AI journey, leveraging insights from IBM’s Institute for Business Value (IBV) and recent HfS research, as well as real-world experiences, gleaned from interviews with clients and field practitioners.
Abhishek Tayal (Twitter)
Abhishek Tayal offers insight into how Twitter's ML platform team, Cortex, is developing models, related tooling, and infrastructure with the objective of making entity embeddings a first-class citizen within Twitter's ML platform. Abhishek also shares success stories on how developing such an ecosystem increases efficiency and productivity and leads to better outcomes across product ML teams.
Neta Zmora (Intel AI Lab)
Neta Zmora offers an overview of Distiller, an open source Python package for neural network compression research. Neta discusses the motivation for compressing DNNs, outlines compression approaches, and explores Distiller's design and tools, supported algorithms, and code and documentation. Neta concludes with an example implementation of a compression research paper.
Paco Nathan (
Paco Nathan examines decentralized data markets. With components based on blockchain technologies—smart contracts, token-curated registries, DApps, voting mechanisms, etc.—decentralized data markets allow multiple parties to curate ML training datasets in ways that are transparent, auditable, and secure and allow equitable payouts that take social values into account.
Greg Brockman (OpenAI)
OpenAI has recently demonstrated systems capable of advanced robotic manipulation, holding their own against professionals in the massively complex game Dota 2, as well as unprecedented language understanding. Greg Brockman discusses the increasing generality of these systems and their implication for how we should think about and plan for creating safe AGI.
Zhou Xing (Borgward R&D Silicon Valley)
Predicting driver intention and behavior is of great importance for the planning and decision-making processes of autonomous driving vehicles. Zhou Xing shares a methodology that can be used to build and train a predictive driver system, helping to learn on-road drivers' intentions, behaviors, associated risks, etc.
Karmel Allison (Google)
Building machine learning models is a multistage process. TensorFlow's high-level APIs make this process smooth and easy, whether you're starting small or going big. Karmel Allison walks you through a practical example of building, training, and debugging a model and then exporting it for serving using these APIs.
Alyssa Simpson Rochwerger (Figure-Eight)
AI - everyone is talking about it but who is actually doing it (and generating business results). This session takes an industry by industry perspective on true AI adoption disambiguating the hype from the reality, the theoretical from the practical and the research labs from ROI.
Danny Goodman (Switchback Ventures)
Danny Goodman discusses reinforcement learning and the future of software.
Robert Currie (UCSC Genomics Institute)
Robert Currie offers an overview of the Cancer Genome Trust, which was developed to enable providers to openly share consented patients’ deidentified health data using Ethereum and IPFS at a clinical-relevant time scale. Robert also discusses a pilot at UCSF that includes genetic, clinical, and imaging patient data.
Elizabeth Partridge (milk+honey), Nick Paquin (milk+honey), Byron Freney (milk+honey)
Utilizing AI technologies to advance business goals remains one of the most daunting challenges for many business leaders. Beth Partridge, assisted by Nick Paquin and Annie O'Connor, shares a breakthrough approach that bridges the gap between data science and business. Join in to gain a clear understanding of what AI can do for your business and how to go about implementing it.
Martin Görner (Google)
Martin Görner explores the newest developments in image recognition and convolutional neural network architectures and shares tips, engineering best practices, and pointers to apply these techniques in your projects. No PhD required.
Peter Norvig (Google)
In 2011, we saw a sudden increase in the abilities of computer vision systems brought about by academic researchers in deep learning. Today, Peter Norvig explains, we see continued progress in those fields, but the most exciting aspect is the diversity of applications in fields far astray from the original breakthrough areas, as well as the diversity of the people making these applications.
Caroline Sofiatti (Computable Labs )
Caroline Sofiatti explains how a token economy and a privacy-centric approach can reduce analysis times, cut data preparation costs, solve insufficient data problems, improve data security and governance, and enable the next generation of AI algorithms.
mayukh bhaowal (Salesforce)
Machine learning is eating software. As decisions are automated, model interpretability must become an integral part of the ML pipeline rather than an afterthought. In the real world, the demand for being able to explain a model is rapidly gaining on model accuracy. Mayukh Bhaowal discusses the steps Salesforce Einstein is taking to make machine learning more transparent and less of a black box.