Moving beyond the 'beads-on-a-string' model of speech

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Deep Learning: Methods and Applications (2014, background): “Likewise, it has also been well understood for a long time that the use of phonetic or its finer state sequences, even with contextual dependency, in engineering speech recognition systems, is inadequate in representing such rich structure [86, 273, 355], and thus leaving a promising open direction ”
Dynamic Speech Models: Theory, Algorithms, and Applications (2006, background): “” This contrast has been discussed in some detail in [11] with useful insight.”
A Bottom-Up Modular Search Approach to Large Vocabulary Continuous Speech Recognition (2013, background): “The notion that a word is composed of a sequence of phonemes is often referred to as “beads-on-astring” [30], [31].”
Self-Attentional Acoustic Models (2018, methodology): “After this, traditionally a hidden Markov model is applied over polyphones, modeling only local dependencies (beads-on-a-string view [1]).”
Deep Belief Networks for phone recognition (2009, background): “Substantial research effo rt has been devoted to going beyond the “beads-on-a-string” view of speech to representing str uctu e in speech above the level of the phonetic segment [5, 6, 7].”
Discriminative acoustic word embeddings: Tecurrent neural network-based approaches (2016, background): “This is helpful since, despite decades of work on sub-word modeling [1, 2], it still poses signiﬁcant challenges.”