Last updated on July 24, 2018.
Published on July 24, 2018.
Choosing the right cloud platform for your business can be tricky and ultimately depends on what specific service-based capabilities you are looking for. To help identify platforms and services to evaluate for your Natural Language Processing (NLP) needs, we have compiled and categorized the related services from several leading cloud platform providers.
The cloud platform providers we evaluated are Amazon, Google, IBM, Microsoft, Oracle, Salesforce, and SAP. We looked at the services offered by each and matched up those that provide similar capabilities. Please see the table below.
NLP Cloud Platform Comparison
| Service | AWS | IBM | Microsoft Azure | Oracle | Salesforce | SAP | |
|---|---|---|---|---|---|---|---|
| Text to Speech | Amazon Polly | Cloud Text-to-Speech | Watson Text to Speech | Bing Speech API | Java Speech API | ||
| Speech to Text | Amazon Transcribe | Cloud Speech-to-Text | Watson Speech to Text | Bing Speech API, Speaker Recognition API (Preview), Custom Speech Service (Preview) | Java Speech API | ||
| Language Translation | Amazon Translate | Google Cloud Translation API | Watson Language Translator | Translator Speech API, Translator Text API | SAP Leonardo Machine Learning Foundation | ||
| Text Analysis | Amazon Comprehend | Google Natural Language API | Watson Natural Language Understanding | Text Analytics API, Language Understanding (LUIS) | Einstein Sentiment | ||
| Conversation Bots | Amazon Lex | Dialogflow Enterprise Edition | Watson Assistant (Formerly Watson Conversation) | Azure Bot Service | Oracle Intelligent Bots Service | Einstein Bot | SAP Conversational AI |
| Language Classifier | Watson Natural Language Classifier | ||||||
| Linguistic Analysis | Watson Tone Analyzer | Linguistic Analysis API | Einstein Sentiment | ||||
| Spell Check | Bing Spell Check API | SAP Leonardo Conversational AI Foundation | |||||
| Web Language Model | Web Language Model API |
Two common services provided are Text-to-Speech and Speech-to-Text. Typical capabilities for Text-to-Speech services include natural sounding voices, pitch tuning, speed control, and support for multiple languages. Speech-to-Text capabilities include easy-to-read, multiple language support, custom vocabulary, and ability to recognize multiple speakers.
Other common services are language translators, text analyzers, and conversation bot builders. The language translation services are used for secure, batch and real-time translations. Common capabilities for text analysis services are keyphrase extraction, sentiment analysis, entity recognition, language detection, and support for multiple languages. Conversation bots are conversational interfaces, also known as “chatbots.” We have seen significant growth in interest in chatbots over recent months.
IBM, Microsoft, and Salesforce offer linguistic analysis services. Linguistic analysis includes simplifying complex language concepts, parsing text, sentence separation and tokenization, part-of-speech tagging, and constituency parsing. IBM also offers a language classifier for classifying text passages, evaluating results, and building conversational applications.
Microsoft and SAP both offer spell check services, which not only help correct spelling errors but also help recognize brand names and slang, as well as understand homophones. Microsoft also offers a web language model, for automating tasks, joint and conditional probabilities, and next word completions.
Click here to contact us if you would like to discuss how these NLP services can be effectively put to use to meet your project needs.