A lot of ink has been spilled on the great things that AI promises for forward thinking companies. Transform 2019, July 10 & 11 in San Francisco, is all about what goes missing too often in the conversation, and in other conferences: a look at the AI technology that’s actually producing returns on investment, and a deep granular dive into how to reproduce those results in your own business.

We’re bringing together the smart and savvy VIPs to show us exactly what’s under the hood, and take us all the way through to how the rubber hits the road. We have several individual tracks at Transform (around NLP, computer vision, intelligent RPA, business AI integration,AI at the edge, and implementing AI), but one track to consider is the “Tech” track, where we feature sessions specifically focused on the tech behind AI.

Here’s a look at some of the dig-in-deep sessions worth your time.

Design thinking around AI projects

First up, we’re debunking the myth that AI implementation takes months or years — this hands-on workshop is all about how to do it in weeks. The design thinking methodology is a solution-based approach to solving problems. The workshop will look at how to bring together the needs of both IT and business stakeholders to create a viable strategy with real business value; tackling complex infrastructure challenges at speed by leveraging accelerator tools; and ways to adopt a hands-on culture of iterative innovation to produce working AI prototypes.


AI Unleashed

An exclusive invite-only evening of insights and networking, designed for senior enterprise executives overseeing data stacks and strategies.


Learn More

How Reinforcement Learning is changing the games industry and beyond

The Unity Technologies game development platform is used by half of the world’s mobile games, and Danny Lange, VP of AI and Machine Learning at Unity, has created a machine learning agents to play leading games.

Lange believes that gaming is overlooked as the number one driver for leading-edge AI. This year, the company launched the Obstacle Tower Challenge: a first-of-its-kind AI challenge designed to test the capabilities of intelligent agents and accelerate the research and development of AI. It’s a gamified way of testing the vision, control, planning, and generalization abilities of AI agents all at once. The challenge is in phase 2.0 as of May, and Lange will be on stage to talk about how and why it’s changing the game for AI in every industry.

We have two other sessions that will also dealing reinforcement learning trend: First, a main-stage talk from the co-founders of OpenAI, an organization that has also built agents that can beat the best human players at complex games, and is inspiring a generation of executives to take these learnings to other industries; second, a session led by recognized expert on the topic, the Chief Data Scientist, Accenture.​

Recent advancements in AI algorithms and systems

Over the past year we’ve seen major algorithmic developments, improved applications in natural language processing and visual perception, and the rich open source ecosystem for model development — like the Kubeflow and PyTorch software frameworks — keeps growing. The implications are huge for any ambitious company’s tech strategy. Gil Arditi, head of product, Machine Learning at Lyft, is a long-time pioneer in deep learning, and he’ll be providing an overview of these major developments and what they mean for your company right now, as well as how to make the most of evolving technologies like AutoML and Reinforcement Learning.

Standardizing and scaling AI in your organization

Airbnb is one of those companies that has Just Been Doing It for a long time, when it comes to implementing AI into every part of their business and seeing explosive results; Adobe has been embracing AI for all of its creative products, and promising new innovations along the way. So Andrew Hoh, Product Manager, Applied Machine Learning & Machine Learning at Airbnb, and Anil Kamath, VP of Technology, at Adobe, will be joining us to talk about how smart organizations are standardizing processes for machine learning and AI projects, which allows them to scale training and deployment without having to reinvent the wheel each time. Here’s how to be smart.

Semantic Search: Success stories

Semantic search is the next level of search accuracy — it’s search that understands the searcher’s intent as well as the context, can concept match, and process natural language queries. Zappos has seen some big success with their semantic search technology, says Ameen Kazerouni, Lead Data Scientist at Zappos, and will be talking about the bleeding edge of semantic search, as well as how they implemented the technology, are achieving their strategic goals, and what’s next on the semantic search agenda.

Deep Learning 101, Transfer Learning and more

The above are just a few of the tech-specific sessions at Transform. They’re really just the start. We’ll have a ton more on tap too, from a Deep Learning 101 session, for executives to learn how deep learning is done, to a talk from Ken Goldberg, a professor at UC Berkeley, who has been leading a revolution in “robotic grasping” technology by short-circuiting the time it takes to collect data by using synthetic data and a form of Transfer Learning. There will be a session oriented around the “IT architecture” needed to do AI in real settings, that includes Chris Chappo of Gap Inc and execs from IBM, as well as a workshop that includes a look at how you can detect bias in datasets for ML projects.

As well, in the session “Big Picture: Where are we in understanding human speech,” senior research scientist Bryan McCann will pull back the curtain on the state of natural language tech. He’ll address the relationship between representation and prediction in deep learning and machine learning more broadly. This will serve as a foundation for understanding transfer learning in NLP, the momentous impact large, pretrained systems have had on the representations used for NLP tasks, and two important directions for future research. That first future direction is multitask learning, while the second involves extracting world knowledge from large, pretrained systems for commonsense reasoning, one of the hardest tasks in NLP. With the vast improvement in representation learning, continued research will likely bring increased improvements in multitask learning, commonsense reasoning, and explainable models.

Only a few tickets are left to Transform 2019. To hear these great speakers and a hundred more, plus network with 1000 business execs, grab one of the remaining tickets here.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.