“ASR Status Report 2018-1-8”版本间的差异

• Update the APSIPA paper: experiments results;

                          description of database;
                          analysis of results

• Collect the photo and made the photo book.

• Finish collecting photo and the photo book.
• Buy the gifts.
• Plan the time capsule project for the annual meeting.
• Plan the "special offer" project.
• Help to make tests for Parrot project.

• copy-right for voice-checker and voice-printer
• finish m2asr annual summary
• i-vector system for m2asr(train down test is in progress)
• contract of Deep Curious 深度好奇(haven't got response)

• DNN-i-vector Lid
• build my personal home page
• CTC for social science academy(this task is very samilar with zero resource decoding which list in m2asr project)

• Commercial deep speaker model training still in process. [1]
• Phone-aware scoring on deep speaker feature. [2]
• Phonetic speaker embedding still in process. [3]
• Overlap training for speaker features. [4]
• Voice QR code design [Assign to Shouyi Dai].

• Commercial deep speaker model training and evaluation.
• Phonetic speaker embedding.
• Overlap training for speaker features.

• Technical description of Parrot for a patent

• Configuration adaptation for Parrot

• Commercial deep speaker model training. [5]
• Phone-aware scoring on deep speaker feature. [6]
• Phonetic speaker embedding. [7]
• Overlap training for speaker features. [8]

• Commercial deep speaker model training.
• Phone-aware scoring on deep speaker feature.
• Phonetic speaker embedding.
• Overlap training for speaker features.

• A brief survey on oral language evaluation, both on application and algorithm

• A test version of Parrot
• Patent of Parrot