Aug
15
4:25 PM16:25

"Identifying Call Intent from Our Clients" by Justin Federico (Vanguard) & Jeff Emberger (Vanguard)

Clients call Vanguard for many reasons which increases the complexity of our call routing. In order to better serve our clients, Vanguard utilized AI to create an experimental model that identifies call intention.

This talk will present a high level overview of how client calls were used in an AI model. The talk will also cover the full development of how we ingest client calls into cloud, transcribed them into text format, and built a AI topic model that used the transcriptions.


View Event →
Aug
15
4:25 PM16:25

"Credit Card Limit Adjudication: A Contrafactual Regression Study" by Matthew Wander (TD Bank)

Credit Card Limit Adjudication: a Contrafactual Regression Study

The purpose of this talk is to discuss the intersection of big data and business particularly as it impacts the banking sector. This talk will discuss the major issue of applied machine learning and AI techniques to credit card adjudications, a major opportunity area for the banking industry. In fact in the future more and more loan decisions will be made algorithmically, with extraordinary implications for the consumers of these products.

For TD Bank this presents us with a unique opportunity to improve access to these produces as well as fairness in the decision making process. While this is more of an applied rather than theoretical talk, it will focus on a number of aspects of modeling and business decision making. In particular it will focus on the key tenet of data science that one should not intercede in the very thing you are attempting to predict with your model. Unfortunately, many business decisions do precisely that. As such techniques are required to keep the model improving making sure that the new data it is receiving is actually "new". This is where contrafactual regression comes in. While it involves sub-optimal decision making from a business perspective it allows for the best possible business outcome for a given strategy.


View Event →
Aug
15
4:25 PM16:25

"The Humane Machine: Building AI Systems with Emotional Intelligence" by Arpit Mathur (Comcast)

AI based systems have become more and more a part of our everyday lives in the last few years. We now interact routinely with AI based systems like Siri and Alexa. However most such systems today are designed to be functional with little attention paid to the emotional context of these interactions. This is the space of the emerging science of Affective Computing, sometimes referred to as "Emotional AI".

In this session we'll look at the state of the Artificial Emotional Intelligence space as well as the components of modern AEI systems especially in the domain of digital conversational agents. We'll also look at how humans react to such systems and conclude with the risks and rewards of building systems that are capable of understanding and simulating emotions.


View Event →
Aug
15
3:45 PM15:45

"Enhancing Customer Experience with Xfinity Home Intelligence" by HongCheng Wang (Comcast)

With millions of smart cameras and IoT devices in our customers’ home, Xfinity Home is becoming a leading platform to improve our customers’ safety and security. Our teams are using computer vision and deep learning techniques to understand videos and sensor data from our customers’ home, to improve user experience by delivering relevant events and customizing event notification. We detect and recognize objects, identify region of interest and recognize activities by exploring the spatial-temporal relationship among object, place and actions. We built an efficient, scalable and robust video and data analytics platform for smart home applications. I will discuss the challenges of data analytics, summarize our recent efforts, and identify future opportunities for smart home.


View Event →
Aug
15
3:45 PM15:45

"Natural Language, Ask Data, and Other Tableau Stories" by Vidya Setlur (Tableau Software)

Natural language processing has garnered interest in helping people interact with computer systems to make sense and meaning of the world. In the area of visual analytics, natural language has proven to help improve the overall cognition of a visualization task to the user. In this talk, with a sampling of research and engineering projects at Tableau, I will discuss how natural language can be leveraged in various aspects of the analytical workflow ranging from smarter data transformations, visual encodings to asking questions of one's data.


View Event →
Aug
15
3:45 PM15:45

"Choreograph best customer outcomes with AI, ML, and the CX Matrix" by Michael Hoffman (CXC)

Artificial intelligence and machine learning enable companies to move “at the speed of customer” to address individual customer needs, sculpt personalized service, and transform experiences.

After hands-on work re-engineering customer processes with  teams of modelers and developers to craft industry changing customer experiences at over 100 Fortune 500 companies and tens of mid-size and start-up companies, Michael Hoffman compiled the successful best practices and wrote Customer Worthy, Why and How Everyone One in Your Organization Must Think Like a Customer! This text presents a method for every company to depict each customer-to-company interaction in the customer’s context.

During this session, Michael will discuss

  • How to visualize, monetize, analyze, prioritize, optimize customer interactions across the customer lifecycle

  • Practical applications designed to enhance employee performance in every interaction

  • Real life learnings: Utopia versus Business Reality

  • What’s next: virtual reality, enhanced reality, internet of everything, autonomous experience and digital twins


View Event →
Aug
15
3:05 PM15:05

"Chaos Engineering & Machine Learning - Practical Applications for Improving Customer Experience" by Sudhir Borra (Comcast)

The presentation goes over how we can integrate principles of chaos engineering into building machine learning models. The output from these integrated models are different from traditional ML models and can be used across a wide spectrum of industry use cases.

The focus is on micro-services based applications with emphasis on cloud native and serverless environments. We cover several scenarios and real examples using insights from our implementation. We dwell into how these models can be deployed at the edge for improving customer experience, realizing intrusion detection use cases, as well as customizations that can be easily incorporated with this approach.


View Event →
Aug
15
3:05 PM15:05

"How to Incorporate AI in Telecom to Build an Intelligent Dispatch" by Sepideh Seifzadeh (IBM)

Facing its own path toward digital transformation, network communication companies such as Sprint have rose to the challenge to make sure the vision of 5G is a indeed a reliable one. Sprint started preparing their data for Artificial Intelligence (AI) with the goal of using machine learning algorithms to gain near real-time insights and increase responsiveness to customers.

The success of Sprint's digital transformation depends on the ability to quickly discover, organize and present the right data at the right time to those teams that make decisions that impact the customer journey. IBM Cloud Private for Data, a leading enterprise insight platform provided the right solution for Sprint, enabling AI projects in a shorter timeframe through unifying and simplifying four critical stages in the journey to AI: the collection, organization, analysis, and modeling of data.

Sprint currently uses IBM Netcool to collect and understand network alarms and events. Based upon specific alarms, trouble tickets and dispatches of people may occur. The dispatches often require equipment to be carried with the dispatched resources, where often the incorrect or less than optimal equipment and resources are sent, causing multiple truck rolls to resolve issues. In this work, the objective was to use Artificial Intelligent and train a Machine learning algorithm to predict the resolution of faults and alarms and, when necessary, to dispatch resources and make recommendations on the correct parts needed based on historical data.


View Event →
Aug
15
3:05 PM15:05

"Lies, Damned Lies, and Accuracy Metrics" by Joe Morrison (Azavea)

This session is a beginner’s guide to:

  • Gaining fluency in common statistics used to describe machine learning models and their limitations.

  • Understanding how they can be manipulated, both intentionally and unintentionally.

  • Seeing some real examples of accuracy metrics done right and wrong.

Advances in machine learning technology, especially the sub-field of deep learning, have dramatically accelerated the adoption of computer vision methods in nearly every industry. In the world of drones and satellite imagery, companies are being formed to help with everything from counting trees in an orchard, to calculating the volume of stockpiles, to estimating the rust on aging infrastructure. However, describing the accuracy of a machine learning model is far less straightforward than many of these companies make it seem when advertising their work. In practice, statistical measures used to evaluate machine learning models are often more misleading than they are helpful.

This talk will use real examples projects that leveraged machine learning and drone or satellite imagery to automate a task--these examples will be used to illustrate how different evaluation techniques can tell different stories about the same data, and to suggest best practices for portraying the accuracy of a model... accurately.


View Event →
Aug
15
2:25 PM14:25

"Using AI to Improve Customer Experience by Measuring Service Reliability Grade" by Preethi Bojja (Comcast)

We propose to enhance customer experience by leveraging unsupervised learning to classify severity levels - to define the intensity of service reliability grade experienced by customer in products usage, call center, etc. The techniques used as a resolution are not limited to auto encoders, clustering algorithms, and latent variable models. This effort will help us move toward maximizing product engagement, retention, mitigating churn, increasing customer lifetime value. AI methodologies help to unearth hidden factors that adversely affect quality customer engagement.  


View Event →
Aug
15
1:50 PM13:50

Workshops and Tours [PM Session]

IBM Workshop

Watson Assistant Hands-on Lab

The Watson Assistant lab/workshop will walk you through the steps needed to create a virtual agent using Watson Assistant Lite on IBM Cloud.

Prerequisite: Attendees will need to sign up for a free IBM Cloud account prior to the conference.


H2O Workshop

Introduction to Driverless AI v1.7.0

In this hands-on training, we will introduce you to automated feature engineering, model building, visualization, and interpretability. Additionally, we will showcase automatic report generation and one-click model deployment, and highlight new features of the 1.7.0 release. We will conduct this hands-on training using H2O.ai's training platform.

Prerequisite: Attendees will need to sign up for a free H2O account prior to the conference.


Ascend.io Workshop

Hands on with Ascend Autonomous Dataflow Service

Participants will compete to solve complex data challenges by building continuously-optimized, Spark-based data pipelines against real-world datasets. Using the Ascend Autonomous Dataflow Service, participants will be able to:

  • Build large-scale pipelines using declarative configurations and 85% less code, to get up and running in a matter of minutes.

  • Handle unexpected data changes from upstream systems and APIs, to decrease pipeline maintenance.

  • Collaborate and iterate within their team by tapping into live data feeds to fuel downstream analytics and machine learning on-demand.

Prerequisite: Attendees must be comfortable working in SQL


DataRobot Demonstration

Enabling the AI-Driven Enterprise with Automated Machine Learning

Learn how DataRobot’s industry-leading automated Machine Learning platform addresses the predictive model building and deployment needs for a broad spectrum of users - from advanced Data Scientists to beginner Data Analysts and everyone in between. In this session, Gourab De, VP of Data Science, will focus on a Customer Lifetime Value use case to demonstrate the workflow from model training and tuning all the way to deployment. 

 No laptops needed, but please bring your curiosity and questions to make the session very interactive!


Databricks Workshop

Building Data Pipelines for Apache Spark™ with Delta Lake

Delta is Databricks’ next-gen engine built on top of Apache Spark. This course is for data engineers, architects, data scientists and software engineers who want to use Databricks Delta for building pipelines for data lakes with high data reliability and performance. The course will cover typical data reliability and performance challenges that data lakes face and teach how to address them using Delta. The course ends with a capstone project building a complete data pipeline using Databricks Delta. Topics Covered Include:

  • Creating Delta tables for a data lake

  • Appending records to a Databricks Delta table

  • Performing UPSERTs of data into existing Databricks Delta tables

  • Reading and writing streaming data into a data lake

  • Optimizing a data pipeline and optimization best practices

Architecture:

  • Comparison with Lambda architecture

  • Getting streaming Wikipedia data into a data lake via Kafka broker

  • Writing streaming data into a raw table

  • Cleaning up bronze data and generating normalized query tables

  • Creating summary tables of key business metrics

  • Creating plots/dashboards of business metrics

*PREREQUISITES

Completed the Getting Started with Apache Spark™ SQL, Getting Started with Apache Spark™ DataFrames, or ETL Part 1 course, or already have similar knowledge.


Tour of the Comcast Technology Center

This guided tour showcases the unique Comcast urban campus and educates employees, business partners, and community partners about Comcast’s history, corporate culture, and products.


View Event →
Aug
15
11:35 AM11:35

"Better AI with Better Data" by Nikhil Kishore (Comcast)

Data is one of the most important aspects for training a model. Although there have been significant improvements in end-to-end, deep learning models, the ability to explain predictions and transparency remains a challenge. Nikhi and collaborators, Kevin Bohinski and Jimit Patel, are dedicated to a feature-focused approach for better machine Learning. They curate better data which helps to provide better predictions to improve customer experience. From the bottom up, they also collaborate with data producers to ensure data quality with help from domain experts. The data is then availed for making data driven decisions using our highly scalable platform.


View Event →
Aug
15
10:45 AM10:45

"Bolster the Customer Experience with AI" by Diana Shaw (SAS)

Companies around the world recognize the importance and imperative of incorporating AI technologies (notably natural language processing and conversational AI) to improve organizational processes and deliver services more efficiently. But knowing where to start in such a rapidly evolving field can be overwhelming. Diana Shaw will define core AI technologies, practically explain how they work and share real world examples of AI in action. Beyond the software, companies must consider factors such as infrastructure, legacy systems, ethics, accountability and potential hurdles to successful adoption. Diana will address each of these areas in a discussion on best practices for implementing AI strategies.


View Event →
Aug
15
10:45 AM10:45

Workshops and Tours [AM Session]

IBM Workshop

Watson Assistant Hands-on Lab

The Watson Assistant lab/workshop will walk you through the steps needed to create a virtual agent using Watson Assistant Lite on IBM Cloud.

Prerequisite: Attendees will need to sign up for a free IBM Cloud account prior to the conference.


H2O Workshop

Introduction to Driverless AI v1.7.0

In this hands-on training, we will introduce you to automated feature engineering, model building, visualization, and interpretability. Additionally, we will showcase automatic report generation and one-click model deployment, and highlight new features of the 1.7.0 release. We will conduct this hands-on training using H2O.ai's training platform.

Prerequisite: Attendees will need to sign up for a free H2O account prior to the conference.


Ascend.io Workshop

Hands on with Ascend Autonomous Dataflow Service

Participants will compete to solve complex data challenges by building continuously-optimized, Spark-based data pipelines against real-world datasets. Using the Ascend Autonomous Dataflow Service, participants will be able to:

  • Build large-scale pipelines using declarative configurations and 85% less code, to get up and running in a matter of minutes.

  • Handle unexpected data changes from upstream systems and APIs, to decrease pipeline maintenance.

  • Collaborate and iterate within their team by tapping into live data feeds to fuel downstream analytics and machine learning on-demand.

Prerequisite: Attendees must be comfortable working in SQL


Databricks Workshop

Managing the Complete Machine Learning Lifecycle with MLflow

ML development brings many new complexities beyond the traditional software development lifecycle. Unlike in traditional software development, ML developers want to try multiple algorithms, tools and parameters to get the best results, and they need to track this information to reproduce work. In addition, developers need to use many distinct systems to productionize models.

To solve for these challenges, Databricks unveiled last June MLflow, an open source project that aims at simplifying the entire ML lifecycle. MLflow introduces simple abstractions to package reproducible projects, track results, and encapsulate models that can be used with many existing tools, accelerating the ML lifecycle for organizations of any size.

In this tutorial, we will show you how using MLflow can help you:

  • Keep track of experiments runs and results across frameworks.

  • Execute projects remotely on to a Databricks cluster, and quickly reproduce your runs.

  • Quickly productionize models using Databricks production jobs, Docker containers, Azure ML, or Amazon SageMaker.

What you will learn:

  • Understand the 3 main components of open source MLflow (MLflow Tracking, MLflow Projects, MLflow Models) and how each help address challenges of the ML lifecycle.

  • How to use MLflow Tracking to record and query experiments: code, data, config, and results.

  • How to use MLflow Projects packaging format to reproduce runs on any platform.

  • How to use MLflow Models general format to send models to diverse deployment tools.

Prerequisites:

  • A fully-charged laptop (8-16GB memory) with Chrome or Firefox

  • Python 3 and pip pre-installed

  • Pre-register for a Databricks Standard Trial at http://databricks.com/try

  • Basic knowledge of Python programming language.

  • Basic understanding of machine learning concepts.


Tour of the Comcast Technology Center

This guided tour showcases the unique Comcast urban campus and educates employees, business partners, and community partners about Comcast’s history, corporate culture, and products.


View Event →