Schedule: Tuesday, May 1

Zoubin Ghahramani explores the foundations of the field of probabilistic, or Bayesian, machine learning and details current areas of research, including Bayesian deep learning, probabilistic programming, and the Automatic Statistician. Zoubin also explains how Uber organizes AI research and where probabilistic machine learning fits in.

Danny Lange offers an overview of deep reinforcement learning—an exciting new chapter in AI’s history that is changing the way we develop and test learning algorithms that can later be used in real life—and explains how the crossroads between machine learning and gaming offers innovations that are applicable in other fields of technology, such as the robotics and automotive industries.

Kaarthik Sivashanmugam and Wee Hyong Tok share recommendations to address the common challenges in enabling scalable and efficient distributed DNN training and the lessons learned in building and operating a large-scale training infrastructure.

While deep learning has enjoyed widespread empirical success, fundamental bottlenecks exist when attempting to develop deep learning applications at scale. Ameet Talwalkar shares research on addressing two core scalability bottlenecks: tuning the knobs of deep learning models (i.e., hyperparameter optimization) and training deep models in parallel environments.

Entrusted with the financial data of 42 million customers, Intuit is in a unique position to take advantage of AI to solve some of its customers’ biggest financial pains. Ashok Srivastava discusses technology’s role in solving economic problems and details how Intuit is using its unrivaled financial dataset to power prosperity around the world.

Sergey Ermolin details the latest features, real-world use cases, and what's in store for 2018 for BigDL on Intel Xeon processor-based data center and cloud deployments.

Srinivasa Karlapalem demonstrates an approach for high-throughput single-shot multibox object detection (SSD) on edge devices using FPGAs, specifically for surveillance.

Forecasting the long-term values of time series data is crucial for planning. But how do you make use of a recurrent neural network when you want to compute an accurate long-term forecast? How can you capture short- and long-term seasonality or discover small patterns from the data that generate the big picture? Mustafa Kabul shares a scalable technique addressing these questions.

Episource is building a scalable NLP engine to help summarize medical charts and extract medical coding opportunities and their dependencies to recommend best possible ICD10 codes. Manas Ranjan Kar offers an overview of the wide variety of deep learning algorithms involved and the complex in-house training-data creation exercises that were required to make it work.

What is the impact of AI and deep learning on clinical workflows? Enhao Gong and Greg Zaharchuk offer an overview of AI and deep learning technologies invented at Stanford and applied in the clinical neuroimaging workflow at Stanford Hospital, where they have provided faster, safer, cheaper, and smarter medical imaging and treatment decision making.

Julie Zhu and Dima Rekesh share a deep learning approach for imputing a medical condition based on a multiyear history of prescriptions filled by an individual, using Python and Keras.

David Martin explores cognitive function in conjunction with edge computing and IoT sensors and actuators for eldercare scenarios—specifically the identification of individuals, daily activity monitoring, and aberration detection performed on-premises using HomeAssistant, the Intu open source project, and IBM's Watson cognitive services.

Intelligent applications learn from data to provide improved functionality to users. William Benton examines the confluence of two development revolutions: almost every exciting new application today is intelligent, and developers are increasingly deploying their work on container application platforms. Join William to learn how these two revolutions benefit one another.

Deep learning has fueled the emergence of many practical applications and experiences. Meanwhile, container technologies have been maturing, allowing organizations to simplify the development and deployment of applications in various environments. Join Wee Hyong and Danielle Dean as they walk you through using the Cognitive Toolkit (CNTK) with Kubernetes clusters.

Data scientists and machine learning professionals face a quandary of choices when trying to figure out how to scale their data science experiments. Arshak Navruzyan details the landscape of available options and explains how to make best use of the free and open source tools available.

Sameer Wadkar and Nabeel Sarwar explain how to seamlessly integrate model development and model deployment processes to enable rapid turnaround times from model development to model operationalization in high-velocity data streaming environments.

From determining the most convenient rider pickup points to predicting the fastest routes, Uber uses data-driven machine learning to create seamless trip experiences. Zhenxiao Luo explains how Uber tackles data caching in large-scale machine learning, exploring Uber's machine learning architecture, how Uber uses big data to power machine learning, and how to use data caching to speed up AI jobs.

Deep learning is the driving force behind the current AI revolution and will impact every industry on the planet. However, success requires an AI strategy. Chris Benson walks you through creating a strategy for delivering deep learning into production and explores how deep learning is integrated into a modern enterprise architecture.

AI will fundamentally change (and power) the way the world works together. So what does the future of AI in the enterprise look like? Faizan Buzdar explains how intelligence is being applied to enterprise content in practical ways that will revolutionize the most important business processes for companies of all sizes and across all industries.

AI delivers value to many facets of the automotive value chain, including smart manufacturing, supply chain management, and customer engagement. Andre Luckow discusses how to assess AI technologies, validate use cases, and foster fast adoption and shares lessons and best practices learned from developing computer vision and natural language understanding applications.

Expensify is using AI to streamline and improve customer service, reducing customer wait time from 15 hours to 3 minutes. David Barrett leads a deep dive into the process of building Concierge, a hybrid machine learning-driven chatbot, covering the challenges faced, results to date, and what he sees for the future of AI and customer service.

Jan Neumann and Jeanine Heck explain how Comcast uses deep learning to build virtual assistants that allow its customers to contact the company with questions or concerns and how it uses contextual information about customers and systems in a reinforcement learning framework to identify the best actions that answer these customers' questions or resolve their concerns.

Drawing on his experience leading two successful AI companies that implemented machine learning and NLP solutions in over a hundred organizations, Robbie Allen details patterns and characteristics of successful machine learning implementations (and those that predict failure) and explains how to build and cultivate ML talent within your organization in an increasingly competitive job market.

Advancements in computer vision are creating new opportunities across business verticals, from programs that help the visually impaired to extracting business insights from socially shared pictures, but the benefits of applied AI in computer vision are only beginning to emerge. Ophir Tanz explores the tools and image technology utilizing AI that you can apply to your business today.

Few organizations have mastered integrating AI technology into their business processes and offerings, and many who want to don’t fully understand the work that lies ahead. David Kiron shares surprising insights about businesses’ appetite for and approach to AI, drawn from global collaborative research conducted by MIT Sloan Management Review and The Boston Consulting Group.

Large enterprises struggle to apply deep learning and other machine learning technologies successfully because they lack the mindset, processes, or culture for an AI-first world. AI requires a radical shift. Kathryn Hume explores common failure models that hinder enterprise success and shares a framework for building an AI-first enterprise culture.

Susan Etlinger shares use cases, emerging best practices, and design and CX principles from organizations building consumer-facing chatbots, covering the risks and opportunities of conversational interfaces, the strategic implications for customer experience, business models, brand strategy, and recent innovations.

Tolga Kurtoglu walks you through the advanced technology needed to implement cyberphysical systems, covering the right hardware to sense the right data, explainable AI, and designing security for trustworthy operations. Along the way, Tolga shares case studies and examples of advanced tech deployments.

AI is about more than automating tasks; it's about augmenting and extending human capabilities. James Guszcza discusses principles of human-computer collaboration, organizes them into a framework, and offers several real-life examples in which human-centered design has been crucial to the economic success of an AI project.

Kayvaun Rowshankish and Alexis Trittipo explore the extent to which firms have addressed the EU's General Data Protection Regulation (GDPR) (the deadline being imminent) and how they might build further sustainability into their capabilities, especially through use of AI and other innovative technologies.

Scott Zoldi discusses innovations in explainable AI, such as Reason Reporter, which explains the workings of neural network models used to detect fraudulent payment card transactions in real time, and offers a comparative study with local interpretable model-agnostic explanations (LIME) that demonstrates why the former are better at providing explanations.

Chatbots are having a moment, and banks across the world are utilizing them for everything from basic customer service to assisting internal IT support. But chatbots only skim the AI landscape. Brian Pearce explains how AI helps Wells Fargo use data in a smarter way, from developing custom experiences to uncovering new insights—with customers and employees at the center of it all.

The stock market is well known to be extremely random, making investment decisions difficult, but deep learning can help. Drawing on a concrete financial use case, Aurélien Géron explains how LSTM networks can be used for forecasting.

Financial econometric models are usually handcrafted using a combination of statistical methods, stochastic calculus, and dynamic programming techniques. Ambika Sukla explains how recent advancements in AI can help simplify financial model building by carefully replacing complex mathematics with a data-driven incremental learning approach.

Pensieve is a natural language processing (NLP) project that classifies reviews for their sentiment, reason for sentiment, high-level content, and low-level content. Megan Yetman offers an overview of Pensieve as well as ways to improve model reporting and the ability for continuous model learning and improvement.

Historically, the consumer loan industry has restricted itself to using relatively simple machine learning models and techniques to accept or deny loan applicants. However, more powerful (but also more complicated) methods can significantly improve business outcomes. Sean Kamkar shares a framework for evaluating, explaining, and managing these more complex methods.

Drawing on Affectiva's experience building a multimodal emotion AI that can detect human emotions from face and voice, Taniya Mishra outlines various deep learning approaches for building multimodal emotion detection. Along the way, Taniya explains how to mitigate the challenges of data collection and annotation and how to avoid bias in model training.

Alex Jaimes explains how the cloud can be used effectively to deploy deep learning and the factors that allow you to do so cost effectively. Along the way, Alex shares examples of when and how to deploy deep learning in the cloud as well as the corresponding benefits, challenges, and opportunities.

Activity-based intelligence (ABI) is the art and science of understanding normal patterns of life to enhance the ability of a system to detect anomalous behavior (e.g., to identify cases of credit card fraud). Jamie Irza demonstrates how machine learning can be used to implement ABI for detecting threatening behavior from unmanned aerial systems, commonly known as drones.

The adversarial nature of security makes applying machine learning complicated. If attackers can evade signatures and heuristics, what is stopping them from evading ML models? Yacin Nadji evaluates, breaks, and fixes a deployed network-based ML detector that uses graph clustering. While the attacks are specific to graph clustering, the lessons learned apply to all ML systems in security.

Drawing on NVIDIA’s system for detecting anomalies on various NVIDIA platforms, Joshua Patterson and Aaron Sant-Miller explain how to bootstrap a deep learning framework to detect risk and threats in operational production systems, using best-of-breed GPU-accelerated open source tools.

Jake Porway explores AI’s true potential to impact the world in a positive way. Drawing on his experience as the head of DataKind, an organization applying AI for social good, Jake shares best practices, discusses the importance of using human-centered design principles, and addresses ethical concerns and challenges you may face in using AI to tackle complex humanitarian issues.

Erika Menezes and Serina Kaye share a data science process for music synthesis, including preprocessing, model architecture, training, and prediction, using Microsoft’s Azure Machine Learning.

Amazon SageMaker is a fully managed machine learning platform that enables developers and data scientists to quickly and easily build, train, and deploy machine learning models in the cloud, at any scale. Randall Hunt offers an overview of SageMaker and demonstrates an end-to-end machine learning workflow by building an ML-powered Twitter bot that you can interact with in real time.

Daniel Raskin and Jonathan Greenberg explain what the extreme data economy is about and how machine learning advances along with accelerated parallel computing will play a key role in translating data into instant insight to power business in motion.

Indigenous trackers all over the world can look at a single footprint in the dirt and intuitively know what animal species that print belongs to. Mary Beth Ainsworth explains how biologists, zoologists, machine learning and computer vision experts have come together to develop, automate, and scale a noninvasive approach to monitoring endangered wildlife by analyzing where animals have walked.

Analytic techniques leveraging artificial intelligence can result in dramatic improvements in crime detection and interdiction across diverse attack modalities. Will Griffith and Ben MacKenzie share AI models and operational techniques they’ve used with major banking clients to substantially strengthen and accelerate their responses to criminal attacks.

Fortune 500 companies are building conversational AI in-house to create a competitive edge. Alan Nichol shares a case study of a successful customer acquisition chatbot built by a large corporation and demonstrates how to build a useful, engaging conversational AI bot based entirely on machine learning using Rasa NLU and Rasa Core, the leading open source libraries for building conversational AI.

Ben Vigoda offers an overview of idea learning, a new approach to deep learning that has been funded since 2013 as one of DARPA's largest investments in next-generation machine learning. Ben details the process of teaching machines with ideas instead of labeled data and demonstrates use cases with state-of-the-art performance on applications in unstructured enterprise data.

AI will change—and in some ways already is changing—the way we work, live, and play at a scale the world has never experienced. Join Tatiana Mejia to see how marketers, designers, and creative professionals can gain huge benefits in productivity, content scale, and workflow efficiencies while unleashing expanded career opportunities for workers in these industries.

In manufacturing, software development, and aerospace, tech-op teams need to make critical decisions on the spot with very little information. In this session, presented by Intel Saffron, the speakers share actual use cases of cognitive AI-based applications helping technical professionals make more confident decisions to solve the pressing issues in their day-to-day work.

New technologies make Bayesian inference and generative modeling more accessible to business analysts, but this also creates new communications challenges. Richard Tibbetts shares techniques for capturing domain knowledge and making findings actionable for decision makers utilizing the explanatory powers of transparent AI.

Drawing on his experience bringing AI to the public sector, Sumeet Vij offers perspectives on public sector AI trends, dispelling myths around barriers to entry and sharing approaches and opportunities as he highlights examples of successful AI adoptions.

Regardless of industry, every executive is concerned with the same thing: their customers. Omar Tawakol details the building blocks of speech technologies, including natural language processing, automatic speech recognition, and neural networks, that are necessary to implement voice-activated artificial intelligence and more importantly, enable a customer-centric enterprise.

Artificial Intelligence program chairs Ben Lorica and Roger Chen open the first day of keynotes.

In this fireside chat, Justin Herz and Fiaz Mohammed discuss how artificial intelligence can improve content discovery and monetization. In collaboration with Intel AI technologies, Warner Bros. is just scratching the surface of what’s possible.

Keynote by program chairs Ben Lorica and Roger Chen

Comprehensive and sustainable wildlife monitoring technologies are key to maintaining biodiversity. Mary Beth Ainsworth offers an overview of SAS deep learning and computer vision capabilities that can rapidly analyze animal footprints to help map wildlife presence and scale conservation efforts around the world.

Autonomy—consisting of extensive data processing, decision making and execution, and learning from experience—creates the need for a new interaction between humans and AI. Manuela Veloso delves into the roles humans can have in such interactions, as well as the underlying challenges to AI, in particular in terms of collaboration and interpretability.

Food production needs to double by 2050 to feed the world’s growing population. Jennifer Marsman details a solution that uses sensors in the soil, aerial imagery from drones, machine learning, and networking research in television whitespaces and discusses the AI for Earth grant program, which supports similar work in the areas of clean water, agriculture, biodiversity, and climate change.

The Intel AI portfolio includes hardware and software solutions that span use cases and edge-to-cloud implementations, rooted in extensive expertise in data science and research. Fiaz Mohamed explains how Intel AI solves today’s business problems and how Intel’s partner ecosystem is accelerating the adoption of solutions built on Intel technology.

Fireside chat with Peter Norvig and Kavya Kopparapu

For more than 20 years, Amazon has invested in experimenting and deploying AI at scale. Dan Mbanga explores how accelerating AI experimentation has influenced innovations such as Amazon Alexa, Prime Air, and Go and how developers and data scientists from startups to large-scale enterprises have benefited from this innovation.

Zoubin Ghahramani discusses fundamental concepts and recent advances in artificial intelligence, highlighting research on the frontiers of deep learning, probabilistic programming, Bayesian optimization, and AI for data science. Zoubin concludes by considering the societal implications of this work.

Program chairs Ben Lorica and Roger Chen close the first day of keynotes.

Topic Table discussions are a great way to informally network with people in similar industries or interested in the same topics.

Ready, set, network! Meet fellow attendees who are looking to connect at the AI Conference. We'll gather before Tuesday and Wednesday keynotes for an informal speed networking event. Be sure to bring your business cards—and remember to have fun.

Don't miss AI at Night, happening on Tuesday after the Attendee Reception.

Come enjoy delicious snacks and beverages with fellow AI Conference attendees, speakers, and sponsors at the Attendee Reception, happening immediately after the afternoon sessions on Tuesday.