ByteDance Applies Bidirectional AI Model SeedRealtime to Doubao

ByteDance has applied its bidirectional AI model SeedRealtime, which processes voice, video, and text simultaneously, to the Doubao app. The AI research organization Seed unveiled SeedRealtime on the 10th. This model integrates audio, video, and text into an end-to-end structure, processing recognition, understanding, judgment, and expression in parallel. SeedRealtime supports bidirectional interaction that assesses the flow of conversation while receiving input without waiting for the user to finish speaking, and it includes a turn-taking feature within the model that replaces the external voice activity detector (VAD). According to ByteDance's internal evaluation, the issue of conversation rhythm has been reduced by half compared to the existing cascade structure. This model is an extension of the bidirectional voice model Seeduplex, which was unveiled last April, and Seeduplex reduced the false response rate and false interruption rate by half in complex noise environments. The rate at which the AI interrupts while the user hesitates has decreased by 40%, and the conversation fluency metric has improved by 12%. However, no direct connection to the cryptocurrency and blockchain market has been confirmed with this announcement. The information released by ByteDance does not include plans for token issuance or integration with blockchain networks.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.