Phonetic and Lexical Discovery of a Canine Language using HuBERT

Xingyuan Li,Sinong Wang,Zeyu Xie,Mengyue Wu,Kenny Q. Zhu

2024-02-25

Abstract:This paper delves into the pioneering exploration of potential communication patterns within dog vocalizations and transcends traditional linguistic analysis barriers, which heavily relies on human priori knowledge on limited datasets to find sound units in dog vocalization. We present a self-supervised approach with HuBERT, enabling the accurate classification of phoneme labels and the identification of vocal patterns that suggest a rudimentary vocabulary within dog vocalizations. Our findings indicate a significant acoustic consistency in these identified canine vocabulary, covering the entirety of observed dog vocalization sequences. We further develop a web-based dog vocalization labeling system. This system can highlight phoneme n-grams, present in the vocabulary, in the dog audio uploaded by users.

Sound,Computation and Language,Machine Learning,Audio and Speech Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to explore the potential communication patterns in canine languages, especially to identify phoneme labels and lexical patterns in dog barks. Traditionally, research on animal sounds has relied on human prior knowledge and limited data sets to search for sound units, which has limited the in - depth study. This paper proposes a self - supervised method, using the HuBERT model, which can accurately classify phoneme labels and identify the possible primary words in dog barks. The study found that these identified canine words are acoustically consistent and cover all the observed dog bark sequences. To achieve this goal, the researchers developed a web - based dog bark annotation system, which can highlight the phoneme n - grams present in the dog audio uploaded by users. Through this method, the research not only improves the understanding of canine languages but also provides an important basis for further research on the meaning of dog languages in the future.

Phonetic and Lexical Discovery of a Canine Language using HuBERT

Towards Lexical Analysis of Dog Vocalizations via Online Videos

Towards Dog Bark Decoding: Leveraging Human Speech Processing for Automated Bark Classification

Does My Dog ''Speak'' Like Me? The Acoustic Correlation between Pet Dogs and Their Human Owners

DogChat: A Pet-centered Smart Collar Prototype Based on Large Language Models and Wechat

Word Discovery in Visually Grounded, Self-Supervised Speech Models

PhonHuBERT: A Phoneme Transcription Tool for Song Datasets

Animal speech and singing synthesis model based on So-VITS-SVC

Spatial HuBERT: Self-supervised Spatial Speech Representation Learning for a Single Talker from Multi-channel Audio

Syllable Discovery and Cross-Lingual Generalization in a Visually Grounded, Self-Supervised Speech Model

Speak Like a Dog: Human to Non-human creature Voice Conversion

How do soundboard-trained dogs respond to human button presses? An investigation into word comprehension

Revisiting speech segmentation and lexicon learning with better features

CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation

Integrating Self-supervised Speech Model with Pseudo Word-level Targets from Visually-grounded Speech Model

Language-Universal Phonetic Representation in Multilingual Speech Pretraining for Low-Resource Speech Recognition

Pushing the Limits of Unsupervised Unit Discovery for SSL Speech Representation

Leveraging Laryngograph Data for Robust Voicing Detection in Speech

Automatic Assessment of Language Background in Toddlers Through Phonotactic and Pitch Pattern Modeling of Short Vocalizations

Text-Guided HuBERT: Self-Supervised Speech Pre-Training via Generative Adversarial Networks

SD-HuBERT: Sentence-Level Self-Distillation Induces Syllabic Organization in HuBERT