Attachment0013ResponsesToQuestionsAndAnswers.pdf

Machine Foreign Language Translation System (MFLTS) Text-to-Text (T2T) Language Packs (LPs)

Questions and Answers Solicitation Number: W909MY-18-R-0011 RFP Release Date: 30 April 2018

Question 1: For the Demonstration Plan, please confirm that the intention of the demo will be to test the vendor’s standalone application with the language component(s), and not the integration into the

MFLTS environment. In other words, do we have to develop a Language Pack / Language Component for integration into the MFLTS stack for the purpose of the pre-award demo?

Answer: Under the MFLTS the Demonstration is for the vendor to demonstrate their standalone application with the language components and not the integration into the MFLTS environment.

Question 2: For each of the language packs, are we required to provide file extraction functionality?

Answer: No, the Government will provide the offeror with an input file in a foreign language and the Offeror shall provide an output file in English.

Question 3: Contrary to the Government's response to Question 6, there do not appear to be any sample documents for processing in the reading room area. If the government believes otherwise, could they identify these sample documents?

Answer: There is a new project title string in the reading room that contains the samples.

Name: W909MY-17-R-A023 Title: MFLTS Technical Demonstration Description: Government Corpora for MFLTS Technical Demonstration

Question 4: We cannot find any evidence of PDF support in the Architecture documents, SDK, MFLTS core, or sample apps, and in fact, the documentation for T2T (stored here: \CILABLSDK\MFLTS\app-release\T2TGUI\Help) explicitly states *no* support for PDF. Could the Government point us at the changed documents referred to in Question 16?

Answer: Changes were made to the MFLTS CILABL Language Component Development Guide and MFLTS CILABL Architecture Specification located in the reading room.

Question 5: The government's answer to Question 12 raises additional questions. There is overlap between Simplified and Traditional representations in Chinese, yet they are significantly different graphically. The first portion of the government's answer -- "Yes, for image documents" -- implies the government will be testing the vendor systems with a mixed folder of Simplified and Traditional images for the OCR process. Is that the case, or will the documents be processed separately? We have the same question for Machine Translation. Also, should another line be added to Attachment 10, Pricing Model, to accommodate two separate LPs?

Answer: The Image documents will be Simplified Chinese only. Simplified Chinese Sample documents are located in the reading room.

Question 6: How are you going to assess discounted pricing if a vendor is acceptable and lowest price on, for example, languages 1 and 3, but is not eligible for language 2? This is not captured in the Pricing Model.

Answer: For evaluation purposes, the Initial LP awarded pricing will be used as the first tier pricing, for each additional LP the discounted prices will be used to determine the LPTA. For example if an offeror proposed LP1, LP3, and LP8 etc. The initial pricing will be used for the first awarded LP, each additional awarded LP will use the discounted tier pricing to determine LPTA.