Applause, the Boston-based crowd-sourced testing specialist, has launched a solution for AI training and testing that is says trains AI algorithms to learn quickly and tests the output to ensure those algorithms are processing and responding appropriately. The solution is designed to leverage Applause’s global community of testers to deliver the widest possible range of training inputs, the company said. Applause claims the results can then be tested across every possible device, location and circumstance to identify issues and provide actionable user feedback in real time.
“Users want AI to be more natural, more human. Applause’s crowdsourced approach delivers what AI has been missing: a diverse and large collection of real human interactions prior to release,” said Kristin Simonini, VP of Product at Applause. “Not only will this improve AI experiences for consumers everywhere, the breadth of the community also has the potential to mitigate bias concerns and make AI more representative of the real world.”
Because the Applause community of testers includes people across different countries, ages, genders, races, and cultures it generates a broad base of data samples and that results in a more representative and unbiased output than if the data were sourced from a smaller group, the company said. Applause added that it can also test the outputs of those algorithms to check for bias. If bias has crept into an algorithm at any stage, the community can identify it when testing the output, something that a smaller or less diverse group of testers might not be able to do.
Applause said that its new solution operates across five unique types of AI engagements:
- Voice: Source utterances to train voice-enabled devices, and test those devices to ensure they understand and respond accurately.
- OCR (Optimized Character Recognition): Provide documents and corresponding text to train algorithms to recognize text, and compare printed docs and the recognized text for accuracy.
- Image Recognition: Deliver photos taken of predefined objects and locations, and ensure objects are being recognized and identified correctly.
- Biometrics: Source biometric inputs like faces and fingerprints, and test whether those inputs result in an experience that’s easy to use and actually works
- Chatbots: Give sample questions and varying intents for chatbots to answer, and interact with chatbots to ensure they understand and respond accurately in a human-like way.
