The NTT DCASE2020 Challenge Task 6 system: Automated Audio Captioning with Keywords and Sentence Length Estimation
arXiv:2007.00225
Abstract
This technical report describes the system participating to the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 Challenge, Task 6: automated audio captioning. Our submission focuses on solving two indeterminacy problems in automated audio captioning: word selection indeterminacy and sentence length indeterminacy. We simultaneously solve the main caption generation and sub indeterminacy problems by estimating keywords and sentence length through multi-task learning. We tested a simplified model of our submission using the development-testing dataset. Our model achieved 20.7 SPIDEr score where that of the baseline system was 5.4.
Technical Report of DCASE2020 Challenge Task 6
References in corpus (4)
Cited by in corpus (9)
- BYOL for Audio: Exploring Pre-trained General-purpose Audio Representations
- Audio Captioning using Pre-Trained Large-Scale Language Model Guided by Audio-based Similar Caption Retrieval
- DECAR: Deep Clustering for learning general-purpose Audio Representations
- Effects of Word-frequency based Pre- and Post- Processings for Audio Captioning
- Enhance Temporal Relations in Audio Captioning with Sound Event Detection
- A Transformer-based Audio Captioning Model with Keyword Estimation
- Text-to-Audio Grounding: Building Correspondence Between Captions and Sound Events
- Sound Event Detection with Adaptive Frequency Selection
- Audio Caption in a Car Setting with a Sentence-Level Loss