AWS announced the launch of a series of artificial intelligence application services at the re:Invent 2023 conference, and enhanced the user experience through the underlying foundation model.
The services announced this time include the Amazon Transcribe automatic speech recognition service, which uses a foundational model to enhance language recognition and artificial intelligence computing, and the Amazon Personalize machine learning service, which also uses a foundational model to enhance automatic generation capabilities. It also includes the Amazon Lex chatbot service, which can improve the accuracy of conversational responses through large-scale natural language models.
Among them, the accuracy of the Amazon Transcribe service, after being enhanced by the basic model, is approximately between 20% and 50%. The speech recognition function now supports more than 100 languages and provides ease of use, customization, and customization for security and privacy to meet different usage needs.
The Amazon Personalize service uses auto-generative artificial intelligence technology to enhance more advanced personalized learning capabilities, thereby generating email content that is more in line with personal writing style, generating titles that increase user viewing, and even making online product descriptions more clickable.
AWS also provides the Amazon Personalize service on the open source framework LangChain, allowing users to build various applications based on basic models to make application operations more personalized, such as search functions that can find content that meets personal preferences more quickly, or making e-commerce service recommendations more attractive to users.
In addition, after combining the Amazon Lex chatbot service with a large natural language model, it can now answer common questions more intelligently and quickly, and can provide correct answers by connecting database content, and even significantly reduce the difficulty of chatbot development.
