"VKontakte" today announced the launch of the video message transcription function in VK Messenger.
The transcription technology is fully automated and works on the basis of machine learning methods developed by the "VKontakte" team. For high accuracy, video recordings are processed by three neural networks: they recognize sounds, separating them from the video stream, form words from sounds, and determine the boundaries of sentences.
The company notes that messenger users can transcribe both received and sent video messages.
The technology is trained on movie subtitles and is able to understand indistinct speech and informal vocabulary, remove pauses from the recording, and break the text into sentences and semantic blocks. Speech recognition is fully automated and takes place without human intervention.
The technology already works in the web version web.vk.me, and can also be used by some users of the "VKontakte" and "VK Messenger" applications on Android.
