SpeechCLIP+: Self-supervised multi-task representation learning for speech via CLIP and speech-image data
This work addresses incremental improvements in self-supervised representation learning for speech and image tasks, benefiting researchers in multimodal AI.
The paper tackles improving SpeechCLIP, a model for bridging speech and text via images without transcription, by introducing a Continuous Integrate-and-Fire module and a hybrid multi-task architecture. Results show the CIF-based model outperforms the previous version in speech keyword extraction, and the hybrid architecture boosts performance in image-speech retrieval tasks on Flickr8k and SpokenCOCO datasets.
The recently proposed visually grounded speech model SpeechCLIP is an innovative framework that bridges speech and text through images via CLIP without relying on text transcription. On this basis, this paper introduces two extensions to SpeechCLIP. First, we apply the Continuous Integrate-and-Fire (CIF) module to replace a fixed number of CLS tokens in the cascaded architecture. Second, we propose a new hybrid architecture that merges the cascaded and parallel architectures of SpeechCLIP into a multi-task learning framework. Our experimental evaluation is performed on the Flickr8k and SpokenCOCO datasets. The results show that in the speech keyword extraction task, the CIF-based cascaded SpeechCLIP model outperforms the previous cascaded SpeechCLIP model using a fixed number of CLS tokens. Furthermore, through our hybrid architecture, cascaded task learning boosts the performance of the parallel branch in image-speech retrieval tasks.