Homophone Reveals the Truth: A Reality Check for Speech2Vec
This work exposes potential flaws in a foundational study for speech processing, impacting researchers relying on Speech2Vec for ASR and speech translation improvements.
The paper investigates the authenticity of Speech2Vec, a seminal work on spoken word embeddings, by proposing a homophone-based inspection method and reproducing the model, finding that the released embeddings appear fabricated and the model fails to learn effective semantics, with correlation scores as low as 0.08 and 0.15 on benchmarks.
Generating spoken word embeddings that possess semantic information is a fascinating topic. Compared with text-based embeddings, they cover both phonetic and semantic characteristics, which can provide richer information and are potentially helpful for improving ASR and speech translation systems. In this paper, we review and examine the authenticity of a seminal work in this field: Speech2Vec. First, a homophone-based inspection method is proposed to check the speech embeddings released by the author of Speech2Vec. There is no indication that these embeddings are generated by the Speech2Vec model. Moreover, through further analysis of the vocabulary composition, we suspect that a text-based model fabricates these embeddings. Finally, we reproduce the Speech2Vec model, referring to the official code and optimal settings in the original paper. Experiments showed that this model failed to learn effective semantic embeddings. In word similarity benchmarks, it gets a correlation score of 0.08 in MEN and 0.15 in WS-353-SIM tests, which is over 0.5 lower than those described in the original paper. Our data and code are available.