Academia and Industry Oil and Water or Bread

Academia and Industry Oil and Water or Bread and Cheese? Michael Kirby Department of Mathematics Colorado State University

Two Cultures? • • Expand frontiers Open discussions Basic research Run overnight • Make Money • Mission Statement • • Pillage Plead 5 th Time to market Every op counts • Make Money • Mission Statement

Potential Benefits of Relationship between University and Industry • Impact on Education • Impact on Research • Program Development: CSU niche?

Industry and Education • • • Real World Problems Very practical definition of an “answer” Non-traditional class: industrial seminar Introduction to Intellectual Property Student Benefit: Get a job.

Program for Industry Applied and Computational MS • • • Computing skills (4 courses) Modeling (2 courses) Statistical tools (2 courses) Project development and communication(1) Area of specialization (2 courses)

Industrial Seminar 2003: SRC • • • Brain Storming Group activities Patent Literature Software Project Mathematical Analysis Innovation

The Problem • When is a sensored system behaving nominally? Failure modes. Time to failure. • Examples of systems – Nuclear power plants – X-ray tube data – Turbines – Space shuttle

The Solution • Explore Data Representations (E. g. ) – MSET (Patented) – SVD – MNF – CCA – PLS – Wavelets

The Implementation • Matlab: not ideal but fast implemenation • Graphical User Interface: Platform for discovery • Data Structure: function handles • Distributed code development • Code development system

Matlab code demo

Code development system

Summary: Course Goals Simulate industrial work environment – Interacting groups on large project – Exercise ability to contribute in group brainstorming – Innovation – Interim reporting – Code development and documentation – Frustrations of having an idiot boss

Industry and Research • Real World Problems • Student Research Support • Technology Transition

Research Interests • Optimal Modeling of Massive Data – Model reduction – Geometric and Nonlinear Signal Processing – Image processing and visualization

Industrial Collaboration • • Honeywell Corporation Technical Concepts Management Inc. IBM Euclid Discoveries, LLC Siemens Corporate Research US Forest Service Raytheon

Oil and Water • • Short term support (no 3 year contracts) Problem a good match? Academic freedom Publishing results, Intellectual property Patent ownership Basic research versus development Can’t talk about results (NDAs)

Bread and Cheese • Real World Problems – Drives mathematics – Develops problem solving ability • Support for Research in Mathematics • Realistic Preparation for Students