March 2022 Summaries

Keyword extraction, a text analysis technique that utilizes machine learning and natural language processing, automatically identifies key terms and expressions from text, aiding in summarizing content and recognizing main topics. This technique has evolved since 1999, gaining importance with the growth of the internet and is now supported by various APIs, including AWS, IBM Watson, Microsoft Azure, and OpenAI, which offer diverse functionalities like sentiment analysis and entity recognition. Eden AI provides a platform for integrating multiple keyword extraction APIs, allowing users to optimize performance, cost, and accuracy across different data processing needs. This platform supports industries like marketing, SEO, e-commerce, and healthcare by offering a unified API for accessing various AI engines, ensuring data protection and ease of integration. Eden AI's flexible architecture enables seamless switching between providers, facilitating rapid experimentation and adaptation to specific project requirements.

Eden AI, a French company focused on democratizing artificial intelligence usage, has developed an AI engine aggregator that provides a single API to connect developers and businesses with the most suitable AI solutions for their data processing needs. Supported by the 50 Partners accelerator and backed by successful entrepreneurs from companies like Datadog and Algolia, Eden AI has raised €1.5 million to enhance its platform, which already includes over 20 AI engine providers and more than 500 users globally. The platform offers various AI technologies, such as image recognition and machine learning, enabling companies to adopt AI-as-a-Service without needing extensive internal resources. Eden AI aims to expand its supplier network and integrate new features, including dynamic recommendations, while positioning itself as a leader in the rapidly growing AI market.

Speech-to-Text (STT) technology, also known as Automatic Speech Recognition (ASR), converts spoken language into written text and has evolved significantly since its inception in 1952 with Bell Laboratories' first speech recognition system. Over the decades, advancements like Hidden Markov Models and neural networks have enhanced STT's accuracy and applications, enabling its use in various fields such as healthcare, call centers, and media. Companies like Assembly AI, AWS, Deepgram, Google Cloud, IBM Watson, Microsoft Azure, and others offer diverse STT APIs, each with unique features like real-time transcription, speaker diarization, and multi-language support. Eden AI provides a unified platform for accessing multiple STT engines, offering benefits like performance optimization, cost-effectiveness, and the ability to switch providers easily, without storing user data. By centralizing API management, Eden AI ensures a standardized response format and facilitates seamless integration across different AI services, empowering businesses to leverage STT technology efficiently.

Optical Character Recognition (OCR) is an AI technology that transforms images of text into machine-encoded text, enabling automated extraction from various document types, such as images, PDFs, and scans, while recognizing diverse fonts, sizes, and languages. Initially developed for telegraphy in the early 20th century, OCR technology has significantly advanced, evolving from simple single-font systems to sophisticated deep learning models that provide near-human accuracy. Prominent OCR APIs such as AWS Textract, Google Cloud Vision, and Microsoft Azure Form Recognizer offer specialized capabilities, including structured data extraction and multilingual support. Eden AI facilitates access to multiple OCR APIs through a unified platform that simplifies integration and allows for seamless switching between providers, ensuring flexibility and efficiency in choosing the best service for specific needs. The platform also provides standardized JSON responses and ensures data protection, offering GDPR-compliant options without storing user data.

Named Entity Recognition (NER) is a process used to identify entities such as names, organizations, and locations within text, and various APIs offer this capability with differing features. Key criteria for selecting a suitable NER API include accuracy, pricing, language support, response speed, and integration ease, with Eden AI providing a unified platform to access multiple providers through a single API key, allowing easy switching between them. The platform offers a free version for initial testing, although some features are exclusive to the Enterprise version, and emphasizes the importance of benchmarking APIs on specific data before making a commitment. Integration typically involves REST APIs with standardized JSON responses, and many providers offer free tiers or trial credits to facilitate testing and comparison before moving to production. Language and format support can vary, so it's essential to review each provider's documentation for specifics.

Deepgram, a San Francisco-based speech AI startup founded in 2015, has integrated its advanced speech recognition engine into the Eden AI platform and API, expanding its reach to enterprises by offering reliable and accurate speech-to-text capabilities. With over 60 customers and having processed more than 100 billion spoken words, Deepgram has significantly grown its workforce and secured $25 million in Series B funding led by Tiger Global. The integration allows users to access Deepgram's services in multiple languages through the Eden AI platform or API, enabling developers to enhance their applications with intelligent automation. Eden AI provides a user-friendly experience with options to compare different AI providers based on accuracy, speed, and cost, and supports the creation of composable AI workflows by combining features like OCR, NLP, and vision. The platform offers a free tier for testing purposes and a pay-as-you-go pricing model for production workloads, with centralized cost tracking available on the dashboard.

Assembly AI's Speech-to-Text API, known for its high accuracy and advanced AI models, has been integrated into the Eden AI platform and API, allowing users to accurately transcribe audio and video files with features such as speaker diarization, automatic punctuation, and word timings. The integration supports English speakers and will expand to 35 more languages by early 2022. Developers can access this service by creating a free Eden AI account, generating an API key, and testing the feature before full integration, with the platform offering composable AI workflows that can chain multiple AI capabilities like OCR and NLP. Eden AI provides a free tier for testing and pay-as-you-go pricing for production, with the ability to compare multiple AI providers based on accuracy, speed, and cost.

Sentiment Analysis is a natural language processing (NLP) technique used to identify and extract subjective information, such as opinions, emotions, and attitudes from text, classifying it as positive, negative, or neutral, with potential for more detailed emotion detection. The field has evolved from early rule-based lexicon approaches to sophisticated machine learning and transformer-based models like BERT and GPT, offering high accuracy across languages and domains. Various leading APIs, including those from AWS, Google Cloud, Microsoft Azure, IBM Watson, and OpenAI, provide sentiment analysis services with features like sentiment scoring, opinion mining, and context-aware detection. Eden AI serves as a unified platform to access multiple sentiment analysis APIs, offering streamlined integration, centralized billing, and data protection, while allowing users to easily switch between providers. Key factors for selecting a sentiment analysis API include task-specific accuracy, cost, language support, response time, and integration ease, with most providers offering free trials to facilitate benchmarking.