Case Study: HyperSentience achieves better AI audio analysis with Twine

How Twine Enhanced Audio Analysis of HyperSentience's AI model

HyperSentience, a London-based AI firm specializing in context awareness technology, needed a diverse set of authentic, natural audio recordings of daily home activities to train its AI model. The primary challenge was ensuring the data was genuine and unstaged. To meet this need, they partnered with vendor Twine to obtain this specialized training data.

Twine implemented a comprehensive solution that guided participants to capture high-quality, natural audio. Key factors included the phone's distance from the activity, how it was mounted, the room location, and detailed descriptions of the background noise and activity. This approach provided HyperSentience with the authentic dataset required, significantly enhancing their AI's audio analysis capabilities. Twine also managed all aspects of freelancer recruitment and secure payment, leading to a successful partnership praised by the client.

HyperSentience

Ian Sherwin

Product Owner