- The latest FM-enhanced Amazon Transcribe, the automatic speech recognition service by Amazon, achieves a notable accuracy improvement ranging from 20% to 50% across most languages, as AWS stated.
- Fueled by FMs sourced from Amazon Bedrock and authorized knowledge sources, CFAQ is touted to empower companies to deliver precise, automated responses to prevalent customer inquiries naturally and engagingly.
Amazon Web Services Inc. recently unveiled various new capabilities for artificial intelligence services bolstered by foundation models.
Unveiled at the annual re:Invent 2023 conference, the upgraded capabilities encompass Amazon Transcribe, now featuring FM-powered language support and AI-enhanced call analytics. Additionally, Amazon Personalize utilizes foundation models to generate more comprehensive content, while Amazon Lex employs large language models to deliver precise and conversational responses.
The latest FM-enhanced Amazon Transcribe, the automatic speech recognition service by Amazon, achieves a notable accuracy improvement ranging from 20% to 50% across most languages, as AWS stated. The updated Automatic Speech Recognition (ASR) system introduces distinctive features across its support for over 100 languages. These features focus on aspects such as ease of use, customization, user safety, and privacy.
Illustrative features include automatic punctuation, custom vocabulary, automatic language identification, speaker diarization (identifying and separating different speakers in an audio recording), word-level confidence scores, and custom vocabulary filters. The claimed support for diverse languages and incorporating value-added features are asserted to empower enterprises. This empowerment extends to unlocking valuable insights from their audio content and enhancing the accessibility and discoverability of their audio and video content across various domains.
Amazon Personalize, the machine learning service crafted to assist developers in creating personalized recommendations for their customers, has introduced hyper-personalization using FMs through a feature known as Content Generator.
This innovative feature utilizes natural language to craft simple and engaging text, elucidating the thematic connections between the recommended items. As per AWS, this functionality empowers companies to automatically generate compelling titles or email subject lines, enticing customers to click on videos or make purchases.
AWS has expanded its offerings by providing Personalize on the open-source LangChain framework. This enables customers to construct their applications based on FM. Through this integration, users can invoke Amazon Personalize, fetch recommendations for a campaign or recommender, and seamlessly incorporate them into their FM-powered applications within the LangChain ecosystem.
Lastly, Amazon Lex, the fully managed AI service from Amazon for integrating conversational interfaces into applications using voice and text, is also receiving FM-powered capabilities. This enhancement aims to expedite bot development and enhance containment.
Amazon Lex has introduced Conversation FAQ, a novel capability designed to intelligently and efficiently address frequently asked customer questions at scale. Fueled by FMs sourced from Amazon Bedrock and authorized knowledge sources, CFAQ is touted to empower companies to deliver precise, automated responses to prevalent customer inquiries naturally and engagingly.
CFAQ streamlines bot development by eliminating the necessity to manually create intents, sample utterances, slots, and prompts. This simplification is particularly beneficial for handling a broad spectrum of frequently asked questions. It accomplishes this through a novel intent type known as QnAIntent. This intent securely connects to knowledge sources such as Amazon Bedrock, Amazon OpenSearch Service, and Amazon Kendra knowledge bases, retrieving the most pertinent information to respond to a question effectively.