Loading…

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

Room D [clear filter]
Sunday, October 22
 

8:30am PDT

Welcoming Speech
Speakers
avatar for Jason Geng

Jason Geng

CEO, Wise IOT Solutions
Guest Lecturer at University of Southern California August 2016 - Present Guest Lecturer for Database and Data Science Committee Chair May 2016 - Present Working as committee chair in NPO board and was organizer for "SoCal Data Science Conference 2016"; Organizer of "Dallas Data Science... Read More →


Sunday October 22, 2017 8:30am - 8:40am PDT
Room D Room D

8:40am PDT

Keynote-Data-Driven AI for Entertainment and Healthcare
Speakers
avatar for Demetri Terzopoulos

Demetri Terzopoulos

FRS, FACM, FIEEE; DISTINGUISHED PROFESSOR OF COMPUTER SCIENCE, UCLA; CO-FOUNDER & CHIEF SCIENTIST, VOXELCLOUD INC.
Demetri Terzopoulos is a Chancellor's Professor of Computer Science at the University of California, Los Angeles, where he holds the rank of Distinguished Professor and directs the UCLA Computer Graphics & Vision Laboratory. He is also Co-Founder and Chief Scientist of VoxelCloud... Read More →


Sunday October 22, 2017 8:40am - 9:00am PDT
Room D Room D

9:00am PDT

Keynote: How to Turn an AI Technology into an AI Business
Moderators
avatar for Steve Schoch

Steve Schoch

Former CEO, Miramax Films
Schoch served as CEO of Miramax Films for five years through the end of 2016, and concurrently served as CFO, a position he held from 2010. He led the building of a new studio company from the Miramax assets purchased from The Walt DisneyCompany in late 2010. During this period of... Read More →

Speakers
avatar for Xavier Kochhar

Xavier Kochhar

Founder and CEO, The Video Genome Project (Hulu)
Xavier Kochhar is the Founder and CEO of The VideoGenome Project® (The VGP), a company whose mission is deeply rooted in the belief that the world's data should be accessible for all to use. The VGP is the largest, broadest, and most granular structured database of video content... Read More →


Sunday October 22, 2017 9:00am - 9:40am PDT
Room D Room D

9:45am PDT

Keynote: Artificial Intelligence: Hype, Reality, Vision.
There has been tremendous progress in AI over the past few years. For example, speech recognition in devices like Siri and Alexa, language translation, image categorization, and semi-autonomous driving are now performing at levels that engineers could only dream about just 5 years ago.

Almost all of these improvements are driven by Machine Learning, Big Data, and Massive Computing. These rapid advances have spawned an AI investment mania driven by often quite unrealistic hype and projections – including such claims as fully autonomous driving in 2 years, or 50% unemployment within 10 years. Such claims, plus other misconceptions, are fostering a cadre of ‘prophets of doom’ with warnings of demons and WW3. 

A more sober view, one held by most computer scientists actually working on AI, is that today’s ‘AI’ technologies actually possess very little intelligence – they are very narrow and quite brittle. We seem to be a long way from achieving, for example, the navigational or learning ability of a bee. Never mind the conversational skills of a 4-year old. 

Recently, an increasing number of prominent researchers have begun to publicly acknowledge some core problems of current approaches – namely, the inability of systems to learn incrementally and autonomously, and their lack of generality, transfer learning, and reasoning ability. 

I order to break through this intelligence bottleneck, another approach is starting to gain momentum: ‘Cognitive Architectures’. This paradigm is closely related to what has been called the ‘Third Wave of AI’ – computers that can think, learn, and reason like humans.

Speakers
avatar for Peter Voss

Peter Voss

Founder & CEO, AGI Innovations Inc.
Peter Voss' careers include being an entrepreneur, engineer, and scientist. His experience includes growing a computer solutions company from zero to a 400-person IPO. For the past 15 years, his focus has been on developing AGI (artificial general intelligence). In 2009 Peter founded... Read More →


Sunday October 22, 2017 9:45am - 10:15am PDT
Room D Room D

10:20am PDT

Keynote: Battling Skynet: The Role of Humanity in Artificial Intelligence
The Artificial Intelligence revolution is simultaneously exhilarating and frightening. Human beings will be able to achieve unprecedented advances in every industry, from medicine to manufacturing. However, we fear to become irrelevant as robots are able to perform our tasks better than us. How can humanity survive in a world of robots?

Speakers
avatar for Rumman Chowdhury

Rumman Chowdhury

Senior Manager, Accenture AI
Dr. Chowdhury's passion lies at the intersection of artificial intelligence and humanity. She comes to data science and Artificial Intelligence from a quantitative social science background. At Accenture AI, she works on responsible implementations of enterprise AI. She has been featured... Read More →


Sunday October 22, 2017 10:20am - 10:50am PDT
Room D Room D

10:55am PDT

Deep Learning for Ad Mix Optimization
Online advertising is a highly dynamic and non-linear problem, which makes changes unpredictable, controls unreliable, and the data is often incomplete or censored. Deep learning can address these challenges and can generate optimal ad mixes.

Speakers
avatar for Neal Fultz

Neal Fultz

Principal Data Scientist, System1
Neal Fultz is the Principal Data Scientist at System1 (a Venice ad-tech startup), the Owner-Operator of NJNM Consulting (a boutique statistical consultancy), and the author of several open-source R and Python packages for Bayesian inference and optimization.


Sunday October 22, 2017 10:55am - 11:25am PDT
Room D Room D

11:30am PDT

Hot Dog, Not Hot Dog! Generate new training data without taking more photos.
If you've been a Silicon Valley TV show enthusiast like I am then you've heard about the "hot dog, not hot dog" app.It's true that getting great training data is one of the most arduous tasks but there are simple and proven techniques to generate training data at the comfort of your ergonomic chair without involving a class of college students.

I will share what I've done for one of the self-driving car projects where I generated more image data before feeding it to the Neural Network algorithm for classification.

Speakers
avatar for Annie Flippo

Annie Flippo

Manager of Analytics and Lead Data Scientist, Thinknear by Telenav
Annie is a software engineer, product/project/people manager, and a data scientist. She focuses on the application of machine learning techniques to extract insights in the areas of marketing, media, publishing and consumer behavior. She is an advocate of the STEM for women & minority... Read More →


Sunday October 22, 2017 11:30am - 12:00pm PDT
Room D Room D

12:00pm PDT

Lunch Break
Sunday October 22, 2017 12:00pm - 1:00pm PDT
Room D Room D

1:00pm PDT

Generating Creative Works with AI
In addition to classification and regression capabilities, AI, in particular, Deep Learning can be used for generating new data. A trained deep neural network can be sampled, in different layers, to generate novel data. There has been work in generating "alien-looking" images, novel sounds and text in this way. The challenge in utilizing this technique for creative purposes is shaping the generated samples to fit a user's, or the general audiences' tastes.

I'll be presenting our approach, at EndCue, in utilizing variational auto-encoders, with active learning to elicit user preferences, to generate scripts, dialogues, and speech.

I'll present how we have used these techniques to produce a series of the world's first movies written by AI. I'll also show how we have used these generative AI methods to create a platform for story tellers, to augment their creative process.

Speakers
avatar for Debajyoti (Deb) Ray

Debajyoti (Deb) Ray

CTO, EndCue
Debajyoti (Deb) Ray and his team of Data Scientists and Engineers are building End Cue’s AI platform for analyzing and generating original content.Previously, Deb was the Chief Data Officer for VideoAmp, a cross-screen digital video, and TV advertising platform, where he continues... Read More →


Sunday October 22, 2017 1:00pm - 1:45pm PDT
Room D Room D

1:55pm PDT

Learning to learn Model Behavior: How to use "human-in-the-loop" to explain decisions.
The adoption of Machine Learning or Statistical Models in solving real-world problems has increased exponentially, but users still struggle to derive full potential of the predictive models. There is still a dichotomy between explainability and model performance while choosing the algorithm. Linear Models / Simple Decision Trees are often preferred over more complex models such as Ensembles or Deep Learning models when operationalizing models for ease of interpretation which often results in loss of accuracy.

But, is it necessary to accept a trade-off between model complexity and interpretability?Being able to interpret and explain a model globally faithfully helps in understanding feature contribution on predictions and model variability in a non-stationary environment. This enables trust in the algorithm which drives better collaboration and communication among peers. The need to understand the variability in the predictive power of a model in a human-interpretable way becomes even more important for complex models, e.g., text, image, machine translations.

In this talk, we demonstrate the usefulness of our Model Interpretation library (Skater) for evaluating models using interactiveness and usefulness of Jupiter environment and how it could help practitioners - analysts, data scientists, statisticians - understand the model behavior better without compromising on the choice of algorithm.

Speakers
avatar for Pramit Choudhary

Pramit Choudhary

Lead Data Scientist, DataScience.com
Pramit Choudhary is a Lead Data Scientist at Data Science.com. His focus is on effective ways of optimizing and applying classical (Machine Learning) and Bayesian design strategy to solve real-world problems. Currently, he is leading initiatives on figuring out better ways to explain... Read More →


Sunday October 22, 2017 1:55pm - 2:40pm PDT
Room D Room D

2:50pm PDT

Panel: What are the opportunities for AI and industry in the next 5 years?
Moderators
avatar for Kyle Polich

Kyle Polich

Host, Data Skeptic
Kyle Polich is the host of Data Skeptic, a podcast about data science, statistics, machine learning, and artificial intelligence, all through the eye of scientific skepticism. Outside of the show, he's an advisor for a few early-stage startups and consults with growth companies to... Read More →

Speakers
avatar for Frank Bell

Frank Bell

Principal, IT Strategists
Frank Bell is a Principal at IT Strategists (www.itstrategists.com), a leading business and technology consulting firm in advanced AI, web, and mobile development in Southern California. He has consulted with companies such as Yahoo, Disney, Toyota, Nissan, Deluxe, AEG, Fox, Cisco... Read More →
avatar for Pramit Choudhary

Pramit Choudhary

Lead Data Scientist, DataScience.com
Pramit Choudhary is a Lead Data Scientist at Data Science.com. His focus is on effective ways of optimizing and applying classical (Machine Learning) and Bayesian design strategy to solve real-world problems. Currently, he is leading initiatives on figuring out better ways to explain... Read More →
avatar for Annie Flippo

Annie Flippo

Manager of Analytics and Lead Data Scientist, Thinknear by Telenav
Annie is a software engineer, product/project/people manager, and a data scientist. She focuses on the application of machine learning techniques to extract insights in the areas of marketing, media, publishing and consumer behavior. She is an advocate of the STEM for women & minority... Read More →
avatar for Debajyoti (Deb) Ray

Debajyoti (Deb) Ray

CTO, EndCue
Debajyoti (Deb) Ray and his team of Data Scientists and Engineers are building End Cue’s AI platform for analyzing and generating original content.Previously, Deb was the Chief Data Officer for VideoAmp, a cross-screen digital video, and TV advertising platform, where he continues... Read More →


Sunday October 22, 2017 2:50pm - 3:40pm PDT
Room D Room D

3:50pm PDT

Introduction to Deep Reinforcement Learning
Over the past few years, the AI community has witnessed huge breakthroughs since the development of deep reinforcement learning. For example, using reinforcement learning scientists developed software that learned how to achieve super-human skill in playing different Atari games directly from raw-pixels without being given any game-specific instructions, and the AlphaGo AI system from DeepMind used reinforcement learning to master the Go game and beat the best human champion.

Reinforcement learning is a branch of artificial intelligence that deals with teaching machines how to choose best actions to perform while facing uncertainty. It has a lot of applications in robotics, recommendation systems,  and autonomous systems such as self-driving cars.

In this tutorial, we are going to provide an explanation about reinforcement learning from the ground-up covering the fundamental theory and algorithms. We will also describe how RL meets with the powerful deep-neural networks to form "deep reinforcement learning". In addition, we will demonstrate code examples that show how to apply these algorithms to solve AI problems.

Speakers
avatar for Moustafa Alzantot

Moustafa Alzantot

PhD student in the Networked and Embedded Systems Laboratory (NESL), UCLA


Sunday October 22, 2017 3:50pm - 4:20pm PDT
Room D Room D

4:30pm PDT

AliMe Bot Platform Technical Practice - Alibaba`s Personal Intelligent Assistant in the E-commerce Field
  • With the development of AI, the companies including Google, Facebook, Microsoft, Amazon and many startups have built their own intelligent assistant or bot platform, the intelligent human-computer interaction becomes an important direction of AI. In July 2015 we built an intelligent assistant called AliMe Assistant in Alibaba, focus on three domains: customer service, shopping guide and assistant application in the e-commerce field. 
  • With the development of technology and product, in July 2016 we also published bot platform called AliMe bot platform to help merchants and enterprises which are in the e-commerce ecosystem to enhance their business. The sharing outline:
    • AliMe Bot Platform Introduction
    • Intelligent Interaction Technical Practice In Alibaba
    • The Future and The Challenge

Speakers
avatar for Chen Haiqing

Chen Haiqing

Senior Technical Expert, Alibaba
Alibaba Intelligence Innovation Center Senior Technical Expert, I have worked and studied in the field of the intelligent human-computer interaction in Alibaba for 8 years, lead the technical team to build bot platform and it has covered multi-domain scenes in Alibaba including Taobao... Read More →


Sunday October 22, 2017 4:30pm - 5:00pm PDT
Room D Room D
 
Filter sessions
Apply filters to sessions.