Triplet Entropy Loss: Improving The Generalisation Of Short Speech Language Identification Systems

Abstract

We present several methods to improve the generalisation of language identification (LID) systems to new speakers and to new domains. These methods involve Spectral augmentation, where spectrograms are masked in the frequency or time bands during training and CNN architectures that are pre-trained on the Imagenet dataset. The paper also introduces the novel Triplet Entropy Loss training method, which involves training a network simultaneously using Cross Entropy and Triplet loss. It was found that all three methods improved the generalisation of the models, though not significantly. Even though the models trained using Triplet Entropy Loss showed a better understanding of the languages and higher accuracies, it appears as though the models still memorise word patterns present in the spectrograms rather than learning the finer nuances of a language. The research shows that Triplet Entropy Loss has great potential and should be investigated further, not only in language identification ta

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyvandermerwe2020triplet

Related papers

• Learning Efficient Representations For Keyword Spotting With Triplet Loss (2021)11.76
• Triplet Loss Based Embeddings For Forensic Speaker Identification In Spanish (2021)2.26
• Generative Data Augmentation Guided By Triplet Loss For Speech Emotion Recognition (2022)3.58
• Triplet Based Embedding Distance And Similarity Learning For Text-independent Speaker Verification (2019)5.24
• Two-stage Triplet Loss Training With Curriculum Augmentation For Audio-visual Retrieval (2023)0.00
• Triplet Network With Attention For Speaker Diarization (2018)7.16
• Latent Space Representation For Multi-target Speaker Detection And Identification With A Sparse Dataset Using Triplet Neural Networks (2019)5.24
• Scenario Aware Speech Recognition: Advancements For Apollo Fearless Steps & Chime-4 Corpora (2021)5.84