Issue link: https://read.uberflip.com/i/1444470
Document #US43529118TM ©2020 IDC. www.idc.com | Page 3 IDC White Paper | Accelerate Machine Learning Development to Build Intelligent Applications Faster Cloud-based tools such as managed machine learning services like Amazon SageMaker and preconfigured images like the AWS Deep Learning AMIs (Amazon Machine Images) provide capabilities that handle many of these factors and help developers and their organizations speed deep learning applications to market. There are numerous deep learning tools and frameworks such as TensorFlow, PyTorch, and Apache MXNet — all have valuable attributes that make them useful in developing the algorithms to build the model. However, there are many factors involved that inhibit the development of deep learning models: » Choosing the right deep learning framework for the job at hand » Choosing the right deep learning algorithm » Adjusting and tuning the deep learning algorithm and data for the most accurate predictions » Identifying, locating, and curating training data for deep learning models » Having the right amount of compute resources for both model training and generating predictions in production (inferences) » Integrating deep learning models into existing enterprise applications » Operationalizing models to perform at scale in production Cloud-based tools such as managed machine learning services like Amazon SageMaker and preconfigured images like the AWS Deep Learning AMIs (Amazon Machine Images) provide capabilities that handle many of these factors and help developers and their organizations speed deep learning applications to market. For organizations that prefer the ease and convenience of using pre-trained AI services via APIs and services, Amazon Rekognition for images and video, Amazon Lex for chatbot integration, Amazon Polly for text to speech, Amazon Kendra for intelligent search, Amazon Textract for OCR text extraction, Amazon Translate for natural language translation, Amazon Transcribe for speech recognition, and Amazon Comprehend to find relationships in text can accelerate the addition of intelligent capabilities to applications. AWS offers fully managed services such as Amazon Personalize, Amazon Forecast, Amazon CodeGuru, and Amazon Fraud Detector that make the process of building models very easy. These services automatically inspect user data, extract features, select the appropriate algorithms, and then build and deploy the models into business processes via a single API call. The advantage to developers is that without any machine learning expertise, they can just use these APIs simply and easily without having to go through the entire process of creating their own custom deep learning models. Most developers would be well served to check whether a preexisting API can solve their problem before beginning the process of creating a custom deep learning model.