How To Use Artificial Intelligence in Enterprise?

Artificial Intelligence in the Enterprise

Presented By Ronald Bodkin
Ronald Bodkin
Ronald Bodkin
CTO, Services & Architecture at Teradata

Ron is responsible for leading the global emerging technology team focusing on Artificial Intelligence, GPU and Blockchain. Previously at Teradata. Ron was the founding CEO of Think Big Analytics. Think Big was acquired by Teradata in 2014 

Presentation Description

Presenter: Ronald Bodkin, CTO, Services & Architecture at Teradata

Title: Artificial Intelligence in the Enterprise

Level: Intermediate

Type: Talk

Length: 42:00 Minutes


This talk looks at how deep Learning is affecting the enterprise including use cases like smart navigation, fraud detection, mobile personalization based on individual behavior, face recognition for authentication and data center optimization. It dives deep into real-world experiences from a project at a large automotive conglomerate that collects and analyzes video data from dash mounted video cameras for the purpose of assisting drivers with navigation and safety.

Learning Outcomes:

  • You will understand how AI is being used in a commercial setting.
  • You will know about enterprise applications of Artificial Intelligence.
  • You will learn how to get started with AI at your journey.

0:00 - 3:23: Introduction & Overview

3:23 - 17:34: Why is AI Hot Right now?

  • Evolution of analytics from decision making to increasing sophistication
  • Deep learning is being used to drive more automation
  • The resurgence of AI
  • Increased performance of AI and decreased cost.
  • Significant investment in AI
  • Rapid progress in research of Neural Networks
  • Tremendous uptick in interest of the C-Suite
  • “How do you harness AI for good?”
  • Computer vision was stalled until the advent of deep learning.
  • What is different about the resurgence of AI?
  • We care because of the ability to create value with little or no domain knowledge.
  • Classic Artificial Neural Network
  • AI Technology Landscape: point solutions, specialized APIs, and general purpose frameworks

17:34 - 22:08: Enterprise Applications

  • Industries: automotive, retail, manufacturing & high-tech, healthcare, financial, and insurance
  • Mobile personalization application. Generalize rules and memorize exceptions.
  • Google play store good example of AI in mobile personalization.

22:08 - 26:36 Anti-fraud example

26:36 - 29:49: Key elements of a solution

  • Key requirement is model interpretation, which isn’t new to DL but is compounded in DL.
  • LIME (Locally Interpretable Model Explanation) is extremely useful for AI.

29:49 - 36:48: Connected car assistance application

  • Provide smart assistance to drivers.
  • Example: build a system that detects objects and scenes from in car video to improve navigation and guidance.
  • Customer data -> video files -> frames -> models
  • Approach to solving problem includes survey state of research and open source code
  • Models: single shot multibox detector (SSD). 
  • Another model used was you only look once.
  • Infrastructure was based on AWS

36:48 - 42:49: How to get started on your AI journey?

  • What are the challenges?
  • Technology is still immature
  • Good quality data is key
  • Focus first on pilot into production.
  • Validate -> discovery -> live test -> production
  • Analytics Ops for cross-functional AI
  • Integrate analytics with the Ops team
  • AI will be pervasive in the economy by 2021

Presentation Curriculum

Artificial Intelligence in the Enterprise
Hide Content