1 |
Gradations of interpretability in spoken complex word recognition
|
|
|
|
BASE
|
|
Show details
|
|
2 |
Pertinacity in loanwords: same underlying systems, different outputs
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Beyond decomposition: Processing zero-derivations in English visual word recognition
|
|
|
|
BASE
|
|
Show details
|
|
5 |
Metrical grouping and cliticisation in Middle Dutch: Evidence from verse
|
|
|
|
BASE
|
|
Show details
|
|
6 |
The neural correlates of morphological complexity processing: Detecting structure in pseudowords
|
|
|
|
BASE
|
|
Show details
|
|
7 |
Phonological feature-based speech recognition system for pronunciation training in non-native language learning
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Open syllable lengthening in Middle Dutch: Evidence from verse
|
|
|
|
BASE
|
|
Show details
|
|
9 |
The unabashed typologist: A Frans Plank Schubertiade: Prefac
|
|
|
|
In: Linguistic Typology, vol 21, iss 2017 (2017)
|
|
BASE
|
|
Show details
|
|
10 |
Asymmetric processing of consonant duration in Swiss German
|
|
|
|
BASE
|
|
Show details
|
|
12 |
Attribute based shared hidden layers for cross-language knowledge transfer
|
|
|
|
Abstract:
Deep neural network (DNN) acoustic models can be adapted to under-resourced languages by transferring the hidden layers. An analogous transfer problem is popular as few-shot learning to recognise scantily seen objects based on their meaningful attributes. In similar way, this paper proposes a principled way to represent the hidden layers of DNN in terms of attributes shared across languages. The diverse phoneme sets of different languages can be represented in terms of phonological features that are shared by them. The DNN layers estimating these features could then be transferred in a meaningful and reliable way. Here, we evaluate model transfer from English to German, by comparing the proposed method with other popular methods on the task of phoneme recognition. Experimental results support that apart from providing interpretability to the DNN acoustic models, the proposed framework provides efficient means for their speedy adaptation to different languages, even in the face of scanty adaptation data.
|
|
Keyword:
cross-lingual ASR; Deep neural networks adaptation; knowledge transfer; phonological features; zeroshot learning
|
|
URL: https://doi.org/10.1109/SLT.2016.7846327
|
|
BASE
|
|
Hide details
|
|
13 |
Phonological Feature Based Mispronunciation Detection and Diagnosis using Multi-Task DNNs and Active Learning
|
|
|
|
BASE
|
|
Show details
|
|
14 |
"Fake" gemination in suffixed words and compounds in English and German
|
|
|
|
In: "Fake" gemination in suffixed words and compounds in English and German (2016)
|
|
BASE
|
|
Show details
|
|
15 |
"Fake" gemination in suffixed words and compounds in English and German
|
|
|
|
In: Symplectic Elements at Oxford ; Added by author ; ORA review team (2016)
|
|
BASE
|
|
Show details
|
|
16 |
Macroscopic and microscopic typology: Basic Valence Orientation, more pertinacious than meets the naked eye
|
|
|
|
BASE
|
|
Show details
|
|
17 |
Mutation in Breton verbs: Pertinacity across generations
|
|
|
|
In: Symplectic Elements at Oxford ; CrossRef ; ORA review team (2015)
|
|
BASE
|
|
Show details
|
|
18 |
Height Differences in English Dialects: Consequences for Processing and Representation
|
|
|
|
In: Symplectic Elements at Oxford ; CrossRef (2010)
|
|
BASE
|
|
Show details
|
|
19 |
Phonological phrasing in Germanic: the judgement of history, confirmed through experiment
|
|
|
|
In: Symplectic Elements at Oxford ; CrossRef (2010)
|
|
BASE
|
|
Show details
|
|
20 |
Distinctive features: Phonological underspecification in representation and processing
|
|
|
|
In: Symplectic Elements at Oxford ; CrossRef (2010)
|
|
BASE
|
|
Show details
|
|
|
|