| H92-1037 | phonetic HMM for each speaker in the | speaker-independent training | database . Thus , for each speaker |
| H91-1053 | simultaneously estimated during the | speaker-independent training | process . The speaker - specific |
| H90-1067 | sentences from 109 speakers for | speaker-independent training | . We trained the phonetic HMMs |
| H91-1080 | speaker-independent system . A new paradigm for | speaker-independent training | has been developed . Instead |
| H92-1037 | test-set perplexity of about 60 . The | speaker-independent training | speech database consists of 3990 |
| H92-1037 | clusters are constructed from the | speaker-independent training | database . A codeword-dependent |
| H89-2034 | reference model . Specifi - cally , | speaker-independent training | for the DARPA evaluations utilizes |
| H93-1019 | supports both speaker-dependent and | speaker-independent training | and evaluation . The DARPA TIMIT |
| H93-1009 | improved language modeling and | speaker-independent training | . Phonetic Modeling In the current |
| H90-1079 | system . * A new paradigm for | speaker-independent training | has been developed . Instead |
| H91-1053 | interested in the problems of | speaker-independent training | and speaker adaptation . Therefore |
| H93-1015 | if there is a large amount of | speaker-independent training | avail - able , there is no longer |
| H90-1060 | discuss experimental results . 2 . | Speaker-Independent Training | 109 Speaker SI Training For several |
| H90-1079 | speaker-dependent training or with | speaker-independent training | ( which requires speech from |
| H89-2058 | speaker-dependent training or with | speaker-independent training | ( which requires speech from |
| H94-1079 | speaker dependent ) and x o from | speaker-independent training | data . Thus E \ -LSB- ~ \ -RSB- |
| H92-1078 | Longitudinal Speaker - Dependent , and | Speaker-Independent training | components as well as specifically |
| H91-1054 | reduce speaker differences before | speaker-independent training | is involved such that the speaker-independent |
| H90-1060 | for <title> A New Paradigm for | Speaker-Independent Training | and Speaker Adaptation </title> |
| H90-1039 | data using the RM1 database for | speaker-independent training | ( 109 vs. 72 speakers ) and the |
