Overview of Azure AI Services

Azure AI offers next gen pre-trained AI models, conversational AI to enterprise AI ranging in the areas of images, face, spatial analysis for the fields of Vision along with Speech, Text, Video, Anomaly detection and Open AI services. Microsoft Azure conceptualized the democratization of AI platform incorporating its well versatile Machine Learning (Azure ML) & Deep Learning services.

In this blog post, we’ll define the various categories of Azure AI services.

• Azure AI Vision (Computer Vision) — It’s the API helps to image classification, object detection model build and deployment. Computer vision provides Image analysis API which in turn enables the detection of custom objects, tagging of images, adding custom captions to the images, optical character recognition (OCR) etc.

a) Azure AI Computer Vision API is built into four major catalogs — Face API, OCR API, Spatial analysis and image analysis

b) Apart from building the model using Python/Azure CLI, Azure Vision Studio can be used for building quick development / test PoC. https://portal.vision.cognitive.azure.com/gallery/featured

• Custom Vision — Azure AI Custom Vision API allows to create customized model to classify images and detection of objects. According to the visual characteristics, it enables to identify objects present in an image by labelling the images.
• Speech Service — Azure AI Speech service API enables to build models with speech to text and text to speech capabilities. This Speech API can synthesize custom voice recognition, transcribe speech to high level of accuracy, natural language understanding, translation and speech captioning.

Some of use cases on Azure AI Speech API are as follows —

a) Audio captioning service to create anecdotes, applying profanity filters, customization of audio scripts, translation etc.

b) Create audio input generation by using neural interactions with chatbots and voice assistants.

c) Utilize call center/virtual assistants to transcribe or summarize audio conversation with real customers identifying information and providing insights.

d) Language learning from human feedback using RLHF which is the capability integrated to allow to learn from human feedback from each conversation audio input to train the speech model, support real time transcription etc.

• Language Service — Azure AI Language API service provides natural language processing (NLP) service to provide insights from text, categorization of entity in unstructured text through Named Entity Recognition (NER), detection of sensitive data (PII & PCI), language detection, text summarization and sentiment analysis etc.

— This Language AI service today unifies the classic APIs:

a) Text Analytics

b) QnA maker

c) Language Understanding or LUIS API

• Translator Service- Azure AI Translator service allows via cloud based ML models to translate text and documents through CLI and client library SDK. We can create custom translator model with domain, language and culture neutral capability.
• Document Intelligence- Accelerate your business processes by automating information extraction. Document Intelligence applies advanced machine learning to accurately extract text, key/value pairs, and tables from documents. Document Intelligence tailors its understanding to your documents, both on-premises and in the cloud. Turn forms into usable data at a fraction of the time and cost, so you can focus more time acting on the information rather than compiling it.
• Bot Services — Azure AI Bot services enables to create virtualized chatbots and intelligent virtual assistants through APIs/SDKs.

Using Azure AI Bot services we can build power virtual agents, health bot, skill bot etc. using Bot builder SDK.

• Anomaly Detector — Azure AI Anomaly Detector API can detect anomalies in time series data using ML models including both batch and real time inference data. Anomaly Detector API can detect both unvariate and multivarite anomaly detection derived from single to multi variable model collected from complex systems.
• Content Moderator (Safety) — Azure AI Content Moderator (Safety) AI offers harmful content detection capability embedded in image and texts developed by user and AI generated applications and services.
• Personalizer API- Azure AI Personalizer API creates rich, personalized experiences for every user of your app. Prioritize relevant content and user experiences, improving app satisfaction, usability, and engagement, with Azure AI services Personalizer. Azure delivers this powerful reinforcement learning-based capability through an easy-to-use API. Reinforcement learning is the set of techniques that allow artificial intelligence (AI) to optimize for a goal based on your configuration.
• Azure AI Metrics Advisor — Azure AI Metrics Advisor allows embedding AI-powered monitoring features to stay one step ahead of incidents no machine-learning expertise required. Metrics Advisor monitors the performance of your organization’s growth engines, from sales revenue to manufacturing operations. It helps you quickly identify and fix problems through a powerful combination of monitoring in near-real time, adapting models to your scenario, offering granular analysis with diagnostics, and alerting.
• Immersive Reader- Azure AI Immersive Reader API allows intelligent reading capability to learners by isolating specific word/ content for better readability, display pictures for words, highlight specific part of speech, transcription and translation of speech, reading content aloud etc.
• Azure AI multi-service account- It allows to access multiple Azure AI APIs like vision, speech, search, language using a single key and endpoint. It also enables the billing consolidation of the Azure AI services in usage.
• Azure Cognitive Search — Azure Cognitive Search API provides intelligent search services to web and mobile apps. This search API is empowered by search engine for text and vector based search based on search index with user owned content. It offers rich indexing, lexical analysis and AI enrichment with content extraction and transformation. It also facilitates the vector search capability, text search, fuzzy search required for vector tokenization often utilized in OpenAI apps.
• Azure OpenAI Service- Enables new business solutions with OpenAI’s language generation capabilities powered by GPT-3 models. These models have been pretrained with trillions of words and can easily adapt to your scenario with a few short examples provided at inference. Apply them to numerous scenarios, from summarization to content and code generation.

# Happy transformation with AI