Message-ID: <19439745.1075858462900.JavaMail.evans@thyme> Date: Tue, 15 May 2001 15:34:07 -0700 (PDT) From: tlimperis@grabparts.com To: kamins@enron.com Subject: the special pattern recognition problems Cc: bondie@aol.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Bcc: bondie@aol.com X-From: "Tom Limperis" @ENRON X-To: kamins@enron.com X-cc: bondie@aol.com X-bcc: X-Folder: \Vince_Kaminski_Jun2001_10\Inbox X-Origin: Kaminski-V X-FileName: vkamins.pst Hi Vince Here's a rough draft for your perusal and comments. I can be reached at (517) 423 5617. We at ICAST have had a meeting to explore the two problems we discussed with you and Vasant over the phone last week, and we have decided we feel confident we can find optimum solutions to these problems given the necessary support. Outlined below is the approach we have used for solving difficult pattern recognition problems in the past for Eli Lilly and GM. If Enron is interested in this approach, we would be happy to submit a proposal to Enron to take on this project for one or both of the problems you described. Task 1. Our company will analyze the data sets available from Enron for one week at our own expense. The purpose of this analysis is to estimate the cost (amount of time and the skill set needed) to develop proof-of-concept prototype software. Task 2. ICAST will then demonstrate the POC software to the Enron scientists at the Enron facilities using the training set(s) and one or more new data sets. Task 3. Given successful demonstrations of the POC solution(s), ICAST and Enron will meet to discuss the efficacy of the approaches used and will determine jointly the specifications of planned production software, and the method of providing this capability to Enron for routine every day use. Task 4. ICAST will submit proposals for developing the production versions of the softare that will meet the agreed upon specifications. The cost of developing the proof-of-concept software, and the labor needed to demonstrate it to Enron is estimated to be $150K (plus out-of-pocket expenses) for each of the problems, and, it is estimated at this time that the calendar time needed for each of the projects is about 4-months. It may be possible to do the two Proof-of-concepts concurrently. I will call you before the end of the week to discuss this recommended approach. Tom Limperis CEO ICAST