DEVOPSdigest

Applause announced expanded capabilities in assisting clients with the training, testing and validation of high-quality generative AI models.

Applause has been empowering global organizations to deliver advanced and reliable AI-driven solutions for many years, and is now at the forefront of helping companies test their large-scale generative AI platforms.

Applause offers comprehensive services designed to evaluate the effectiveness of large language models (LLMs) and generative AI models. By subjecting these algorithms to rigorous testing, Applause helps clients identify areas for improvement, enhance model performance, and ensure reliable and unbiased outputs. The testing process encompasses real-world scenarios and diverse user interactions, providing valuable insights into the algorithm's capabilities and limitations. The testing scopes are customized to each model's needs, including reviews of functional capabilities, accuracy of responses, checks for bias/inappropriate content, adherence to custom guidelines, and user experience feedback.

Applause provides an enhanced, proven global data collection infrastructure that allows clients to gather diverse and comprehensive datasets for training their LLMs and generative AI models. By leveraging the collective intelligence of a global community of expert testers, Applause ensures the collection of high-quality data covers a wide range of scenarios, use cases and languages. This robust dataset serves as a valuable resource for clients looking to improve the accuracy, performance and overall user experience of their AI algorithms. Applause's data collection and testing services are seamlessly integrated into its crowdtesting platform, allowing clients to easily leverage these capabilities.

"For more than a decade, Applause has partnered with the world’s leading brands to help them deliver the most innovative digital experiences that exceed their customers’ expectations. Our unparalleled experience in training and validating, as well as our established AI/LLM best practices, will continue to enable industry thought leaders to unlock the full potential of generative AI technologies, while helping to improve the integrity and quality of these experiences. Leveraging our testing community, we can provide valuable, nuanced human insights that help algorithms learn and as a result, improve their accuracy,” said Rob Mason, CTO of Applause.

Applause’s model helps to combat the three main risks and challenges in deploying AI:

- Accuracy and Trust: confirming responses, calling out “hallucinations” or factual inaccuracies, and understanding user experience

- Bias and Inappropriate Content: ensuring datasets are broad enough to rule out systemic bias (racial, gender, religious), and do not contain inappropriate or harmful content.

- Ethical and Regulatory Compliance: checking that models are compliant with copyright, IP permissions or government regulations

"As the prevalence of AI technology continues to shape the business landscape, it becomes crucial for companies to possess comprehensive and unbiased data and test their applications with real people. Applause’s global community of over 1.7 million testers has been successfully supporting this capability at scale, offering leading enterprises the necessary training, testing, validation, and user feedback to drive continuous improvement in this rapidly evolving field," said Chris Malone, CEO of Applause. "With the advent of generative AI, we are witnessing an unprecedented expansion of our overall digital quality testing strategy, empowering enterprises to realize significant benefits and reduced risks as they roll out new AI solutions."