NeoML, launched by ABBYY, the digital intelligence company is an open-source library ideal to build, train, and deploy machine learning (ML) models.
The open-source library which was earlier launched in June 2020 is now available on GitHub as well. NeoML provides huge support to both deep learning and traditional machine learning algorithms.
Best features of NeoML:
Offers 15 – 20 percent faster performance for a pre-trained image processing model that runs on any device.
The cross-platform framework works well with applications that run on mobiles, desktop, and cloud environments.
Best suit for mobile solutions that need both on-device data processing and seamless customer experience.
Around 95 percent of the IT leaders mentioned open source to be a staple while developing mission-critical software. ABBYY also aims at providing support advancements in artificial intelligence (AI). This will only be possible by open sourcing the framework.
Developers can now use NeoML for building, training, and deploying models for predictive modeling, semantic segmentation, object identification, classification, and verification to achieve multiple business goals.
Let’s quote an example here, a bank develops a model to predict customer attrition rate and manage credit risk. These models are further used –
For building remote client identification using face recognition and data verification.
For analyzing retail and fast-moving consumer goods (FMCG) and performance of every marketing campaign.
One of the major advantages NeoML offers is the efficient use of cloud resources. This means that the framework is highly useful when working in a cloud environment. As an AI professional, you need to be on a constant lookout for the latest innovation in the AI realm. As technology keeps getting updated, it is important to keep yourself equipped with the latest news or updates.
How NeoML works?
NeoML is a universal tool that helps in processing and analyzing data of all sorts – text, video, and image. The framework also highly supports programming languages such as Java, Objective-C, and C++. However, Python is yet to be added.
The framework supports more than 100 layer types and offers more than 20 traditional machine learning algorithms – clustering frameworks, regression, and classification. It also contains an abundance of cross-platform single-base code that can run on multiple operating systems – Linux, iOS, Windows, macOS along with android which is further optimized for both the GPU and CPU processors.
Ivan Yamshchikov, AI Evangelist at ABBYY says, “the launch of NeoML reflects our commitment to contribute to industry-wide AI innovation."
Not to mention, the company (ABBYY) is said to have a track record of technological innovation with over 400 patents and patents application. Sharing the framework is critical for developers since it can control cross-platform capabilities, the inference speed, and offers high potential on mobile devices. On the other hand, the feedback received will eventually help improve the open-source library (NeoML).
The company is thrilled to share the advancements made on AI and support machine learning which when applied has been proven efficient in many cases. ABBYY is also planning to add newer algorithms and architectures to further increase the speed which can be achievable with the help of the framework.
The company now invites every tech professional such as business analysts and data scientist to contribute and use NeoML on GitHub where the code is licensed under Apache License 2.0.
Besides this, the open-source library is highly supportive of the Open Neural Network Exchange (ONNX). ONNX is a global ecosystem for interoperable machine learning models. This ML model boosts the compatibility of tools by making it easy for AI professionals to select the right combination for their projects.
The ONNX is being highly supported by Facebook, Microsoft, and other partners of the open-source library.
From: https://www.datasciencecentral.com/profiles/blogs/unveil-the-significance-of-neo-machine-learning-neoml Posted by Aileen Scott
