Microsoft releases new version of its machine learning framework ML.NET

Microsoft has announced a new version of its machine learning framework “ML.NET (v1.0)” and Model Builder for Visual Studio.

ML.NET is an open-source and cross-platform machine learning framework developed to provide model-based machine learning capabilities to .NET developers across the globe.

“Using ML.NET, developers can leverage their existing tools and skillsets to develop and infuse custom AI into their applications by creating custom machine learning models for common scenarios like Sentiment Analysis, Recommendation, Image Classification and more,” Microsoft writes in a blog post.

The ML.NET v1.1 brings a number of new features and enhancements, including a new algorithm based on Super-Resolution Deep Convolutional Network.

For ML.NET, Microsoft has improved the process of loading images in a model. Now, developers can load in-memory images and process them directly.

There is a new Anomaly Detection algorithm named ‘SrCnnAnomalyDetection’ which is based on a Super-Resolution Deep Convolutional Network. The main benefit of this algorithm is that it doesn’t require prior training.

A new component is added to Time Series NuGet package, that will be useful in predicting uncertain events that are impacted by different situations.

For Model Builder, Microsoft has added support for new scenarios as requested by customers. The updates include a new Issue Classification Template to categorize support issues, information accuracy in the Evaluate step, improvement in Code Generation step and feature that addresses multiple customer feedback.

The new issue classification template will allow users to classify tabular data into multiple classes. For instance, users can use the template to predict the issues on GitHub, route the customer support tickets, and classify emails into different categories.

Customers can get started with ML.NET here.

ALSO READ: Microsoft to integrate Truffle blockchain development platform with Azure

Cloud Cloud News Uncategorized

Microsoft rolls out new AI capabilities in Azure for developers and enterprises

Ahead of the Microsoft Build 2019 Developer Conference, the tech giant is rolling out a number of new tools and services to help developers and enterprises to harness the potential of artificial intelligence (AI).

“AI is fueling the next wave of transformative innovations that will change the world. With Azure AI, our goal is to empower organizations to apply AI across the spectrum of their business to engage customers, empower employees, optimize operations and transform products,” wrote Eric Boyd, Corporate Vice President, Azure AI, in a blog post.

Project Brainwave’s Azure Machine Learning Hardware Accelerated Models are now generally available. Announced for a preview last year, it helps in speeding up the training of AI models. Further, Microsoft has pushed the preview of these models for edge computing, in collaboration with Dell Technologies and HPE.

The company is adding support for ONNX Runtime for NVIDIA TensorRT and Intel nGraph to provide high-speed inferencing on NVIDIA and Intel chipsets.

Azure Machine Learning service is getting new capabilities to allow developers, data scientists, and DevOps professionals to increase productivity, operationalize models at scale, and innovate faster. For instance, there is an automated machine learning UI that will allow customers to train ML models just with a few clicks.

Azure Machine Learning will also have a zero-code visual interface, and notebooks to provide developers and data scientists a code-first ML experience.

The hardware accelerated models are also becoming generally available in Azure Machine Learning. These models run on FPGAs in Azure for low-latency and low-cost inferencing. For Databox Edge, it is currently available in preview.

The Machine Learning service is also getting MLOps or DevOps for ML capabilities. These capabilities include Azure DevOps integration to enable Azure DevOps to be used to manage the entire ML lifecycle.

Also read: Microsoft Teams PowerShell module now up for grabs

Furthermore, Microsoft is also previewing a new service called Azure Open Datasets to help customers improve the accuracy of ML models using rich, curated open data and reduce the time spent on data discovery and preparation.


Palo Alto Networks powers its next-gen firewall with analytics and automation capabilities

Leading cybersecurity firm Palo Alto Networks has unveiled a number of new security capabilities for prediction of malicious attacks and then automatically stop those attacks.

The company is adding software and hardware improvements to its next-generation firewall platform and introduced a new cloud-based DNS security service.

Palo Alto mentioned that over 60 new features and tools are coming to its firewall platform that will help enterprises to boost security and simplify protections across their hybrid cloud environments.

“At Palo Alto Networks, we’re focused on simplifying security by using analytics and automation,” said Lee Klarich, chief product officer.

“Customers choose our next-generation firewall platform because of our commitment to continuous innovation and our focus on reducing the need for standalone products. Today’s announcements include our new DNS Security service, which uses machine learning to stop stealthy attacks aimed at stealing information from legitimate businesses.”

The next-generation firewall will now feature PA-7000 Series, Policy Optimizer, VM-Series, and more capabilities. The company claims that PA-7000 Series will provide faster threat prevention and around 3x faster decryption than before.

Policy Optimizer is aimed to replace legacy rules with intuitive policies for stronger security and easier management. This will help in reducing data breaches.

Whereas, the VM-Series is for the enterprises that need consistent security across their virtualized datacenters and multi-cloud. They can deploy the VM-Series in private and public cloud environments based on VMware, Cisco, KVM, OpenStack, Nutanix, Amazon Web Services, Microsoft Azure, Google Cloud Platform, Oracle Cloud, and Alibaba Cloud.

To get the new capabilities in the firewall, customers will need to upgrade to PAN-OS version 9.0.

Talking about the new DNS Security service, it makes use of machine learning for blocking malicious domains and prevent attacks in the progress.

Apart from the new services, Palo Alto has also made its K2-Series generally available. It is a next-generation firewall for service providers with 5G and internet of things (IoT).

Also read: Sectigo to protect enterprises in Middle East against rising cyberthreats

“We listened and responded to customer feedback and found that what customers want above all is simplicity and control,” said Klarich.

“With this release, we’re not only adding features like the DNS Security service, which eliminates the need for security teams to bolt on yet another standalone tool, we are minimizing manual efforts that are error-prone, so teams can focus on projects aimed at growing their business.”

Cloud Cloud News Datacenter Newss

Oracle to set up its first datacenter in India

Oracle is looking to launch its first datacenter in India this year to provide high availability, better performance, and stronger security to customers.

For Oracle, Indian customers contribute significantly to its annual revenue. The company claims that India is the sixth biggest country for them in terms of revenue.

Several industries in the country, including BFSI (Banking, Financial Services and Insurance), telecom and manufacturing, are rapidly embracing modern technologies like machine learning, artificial intelligence (AI), and IoT. This is driving the demand for datacenter.

“We see huge prospects for India which is currently the sixth biggest country in the world for us in terms of revenue standpoint. That is why we are building up a data centre for the Cloud here. I am sure that is just the first of the many data centres coming as our business grows,” told Andrew Mendelsohn, Executive Vice President, Oracle Database, to IANS.

Currently, Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure are the dominant datacenter providers in India. AWS has two availability zones in Mumbai; Microsoft has datacenters at Mumbai, Pune, and Chennai; whereas, Google launched its datacenter in Mumbai last year. Alibaba Cloud also has two availability zones in the same city.

Oracle’s first datacenter in India is very likely to come up in Mumbai. The cloud major will significantly compete with the leading cloud giants like AWS, Azure, and GCP.

The first datacenter will initially handle Platform-as-a-Service (PaaS) and Infrastructure-as-a-Service (IaaS) workloads. The company is still working on the Software-as-a-Service (SaaS) part on the Oracle Cloud Infrastructure.

At the Oracle OpenWorld 2018, Oracle’s Co-founder and CTO Larry Ellison had outlined the roadmap for company’s next generation datacenters.

Also read: Oracle adds self-driving database cloud service to its Autonomous database portfolio

“We have many customers and partners who want to run their business applications on our Gen 2 Cloud. Oracle will open additional regions in a number of countries, including India, to support our customers and fast-growing cloud business in the country,” said Shailender Kumar, Regional Managing Director, Oracle India.

Cloud Cloud News Web Security

Acronis integrates its anti-malware solution PE Analyzer into Google’s VirusTotal

Backup software and data protection solutions provider— Acronis, is teaming up with VirusTotal, a Google subsidiary that provides service for detection of viruses, worms, trojans, and other malicious content in files and URLs.

As a part of the partnership, Acronis will integrate its machine learning-based malware detection engine called Acronis PE Analyzer into VirusTotal platform.

Execution of malware is rapidly increasing year over year and causing threat to Windows operating systems. As per the leading cybersecurity firm Comodo, over 400 million unique malwares were detected in the top-level domains (TLDs) in the second quarter of 2018 alone. AV-TEST registered nearly 400,000 new malware samples a day, which included trojans, backdoors, ransomware, and cryptojackers.

Acronis PE Analyzer aims to address these threats. It is an effective anti-malware solution that uses machine learning models for detecting any Window PE malware.

The company mentioned that its machine learning model is based on a Gradient Boosting Decision Tree that is integrated with a number of neural network models. This creates a file portrait of the threats on the basis of several static characteristics.

This machine learning model can operate independently without an internet connection, while providing high detection rate.

“Given how quickly data threats are evolving, the nature of data protection is fundamentally changing. Solutions must prevent the malicious attacks that target backups to be effective, which is why Acronis has invested in developing our proactive defensive technologies,” said Oleg Melnikov, Acronis Technology Officer.

“Our mission is to protect all data, however, and incorporating our ML-based engine into VirusTotal is the best way to ensure the entire security industry can benefit from Acronis PE Analyzer’s detection capabilities.”

Also read: Acronis doubles investment in Arizona for AI and blockchain projects

Acronis has built the PE Analyzer as a part of its new cyber protection suite which will be released in 2019. Before launching Acronis PE Analyzer, the company will make several improvements to the solution. These improvements will made on the basis of insights gained by its VirusTotal use.

Last month, Acronis had launched the version 7.8 of its Data Cloud Platform with around 80 new features and advancements.

Cloud Cloud News

Azure Machine Learning service now generally available

Microsoft has announced the general availability of its Azure Machine Learning service, allowing AI developers to easily build, train, and develop machine learning models.

The Azure Machine Learning service comes with several advanced capabilities, support for major frameworks, familiar data science tools, and more.

Azure Machine Learning has been built on a number of design principles to simplify and accelerate machine learning development. These design principles include enabling data scientists to use a familiar and rich set of data science tools, and simplify the use of popular machine learning and deep learning frameworks.

Other than these, Microsoft has also focused on accelerating time to value by providing end-to-end machine learning lifecycle capabilities.

Since the Azure Machine Learning service supports popular open-source frameworks like TensorFlow, PyTorch, and scikit-learn, the developers and data scientists will be able to use the tools of their choice.

To enhance the productivity, there are DevOps capabilities for machine learning that enables experiment tracking and management of models deployed in the cloud or on the edge.

Python software development kit (SDK) of the new service can be accessed from any Python environment, editors and IDEs, and Notebooks. What this means is that users will be able to access all the capabilities of Azure Machine Learning from any Python environment, IDEs like Visual Studio Code or PyCharm, or Notebooks like Jupyter and Azure Databricks.

“We built Azure Machine Learning service working closely with our customers, thousands of whom are using it every day to improve customer service, build better products and optimize their operations,” wrote Venky Veeraraghavan, Group Program Manager, Microsoft Azure, in a blog post.

Data scientists can monitor the progress of training jobs visually in near real-time, as the new service supports both local and hosted notebooks.

When customers use multiple frameworks to build models, they often face various challenges in deploying them to hardware and OS platforms. The challenges arise because the frameworks aren’t designed to be used interchangeably.

On that front, Microsoft has collaborated with industry leaders including Facebook and Amazon Web Services (AWS), to develop Open Neural Network Exchange (ONNX) specification. It will define machine learning model in an open standard format. With ONXX support, Azure Machine Learning allows users to deploy, manage, and monitor ONNX models easily.

Also read: Top 4 AI engines to look out for in 2019

Microsoft is continuously making efforts to make things easier for AI developers. In October, the company open sourced the Infer.NET machine learning engine on GitHub. It is the machine learning engine that Microsoft uses to power its own platforms including Office, Azure, and Xbox.

Cloud Cloud News

ZeroStack adds AI-as-a-service capability to its platform

The leading self-driving cloud provider ZeroStack is adding AI-as-a-service capability to its platform. With the new capability, the company will enable its customers to provide one-click deployment of GPU resources and deep learning frameworks to users.

Enterprises and MSPs leverage ZeroStack platform to automate cloud infrastructure, applications, and operations. It allows them to focus on services that accelerate their businesses, simplify operations, and reduce costs.

Artificial intelligence (AI) and machine learning solutions are trending today, and reshaping the experiences in computing. With the availability of modern machine learning and deep learning frameworks like TensorFlow, PyTorch, and MXNet, the AI applications have become more viable than ever.

However, enterprises and MSPs often find it difficult to deploy, configure, and execute the AI frameworks and tools. Also, it becomes complicated to manage their inter-dependencies, versioning, and compatibility with servers and GPUs.

With new AI-as-a-service capability, ZeroStack aims to give its customers the power to automatically detect GPUs and make them available for users. The new innovation will also take care of all the operating system (OS) and CUDA library dependencies, allowing users to focus on AI development.

“ZeroStack is offering the next level of cloud by delivering a collection of point-and-click service templates,” said Michael Lin, director of product management at ZeroStack. “Our new AI-as-a-service template automates provisioning of key AI tool sets and GPU resources for DevOps organizations.”

Additionally, the company mentioned that users can enable GPU acceleration with dedicated access to multiple GPU resources for an order-of-magnitude faster inference latency and user responsiveness. GPUs within hosts can be shared across users in a multi-tenant manner.

Also read: Top 4 AI engines to look out for in 2019

To optimize the utilization of new AI-as-a-service capability, admins of ZeroStack self-driving cloud will be able to configure, scale, and allow fine-grained access control of GPU resources to end users.


Microsoft open sources Infer.NET machine learning framework on GitHub

Microsoft has open sourced the Infer.NET, a machine learning engine that the company uses to power its own platforms including Office, Azure, and Xbox.

Infer.NET is now available on GitHub under the permissive MIT license. Enterprises can use it for free for their commercial applications.

Launched in 2004 as a research tool, the Infer.NET has evolved over the years to become a successful machine learning framework.

Microsoft has built the tool to enable a model-based approach, allowing users to integrate domain-based knowledge into the model. The company explained that Infer.NET can then build a bespoke machine learning algorithm directly from the model.

Typically, the conventional machine learning tools provide existing AI algorithms, so that users can push them into the project according to the needs. But, Infer.NET is different. It constructs the machine learning algorithm on the basis of model provided by the users.

“An added advantage of model-based machine learning is interpretability. If you have designed the model yourself and the learning algorithm follows that model, then you can understand why the system behaves in a particular way or makes certain predictions,” wrote Yordan Zaykov, Principal Research Software Engineering Lead, in a blog post.

“As machine learning applications gradually enter our lives, understanding and explaining their behavior becomes increasingly more important.”

The models built using Infer.NET are capable of handling a broad range of data traits, like real-time data, heterogenous data, unlabeled data, insufficient data, etc.

Also read: Orleans 2.1 released with new scheduler, code generator and performance improvements

Microsoft will make Infer.NET a part of ML.NET, the machine learning framework for .NET developers. The tech giant has already taken a number of steps to integrate it with ML.NET, like setting up the repository under .NET Foundation and moving the packages and namespaces to Microsoft.ML.Probabilistic.

Cloud Datacenter Newss

MariaDB acquires Clustrix to advance its database platform

The prominent database provider MariaDB has acquired Clustrix to make its database platform more advanced.

Clustrix is a provider of relational database engineered for cloud and datacenter. Known for the scalability it provides, the ClustrixDB is an ideal database solution for high-transaction, high-value applications. It delivers over 25 trillion transactions every month through its customers, translating to datasets in billions of rows.

Whereas, MariaDB is an adaptable solution because of its architecture that supports pluggable and purpose-built storage engines.

The acquisition will help MariaDB to add Clustrix’s database scalability to its own platform and better compete in the market. The joint solution will come with higher availability and greater scalability as compared to other traditional distributed options.

“Today, the choices for a scale-out database option are limited – go with a traditional solution like Oracle with high cost and bloat or choose a NoSQL solution that has limited capabilities for data integrity,” said Michael Howard, CEO, MariaDB Corporation.

“With Clustrix, MariaDB can provide a better solution for our customers that have challenging scale-out enterprise environments. Our distributed solution will satisfy the most extreme requirements of our largest customers and gives them the freedom to break from Oracle’s lock-in.”

Recently, MariaDB made additional investment in its platform for further innovation. The company believes that the acquisition of Clustrix will allow MariaDB Labs to tackle the challenges in database field. These challenges are extreme in distributed computing, machine learning, next-generation chips, and memory and storage environments.

“At Nielsen Marketing Cloud, we needed a high-performance solution to process over 100,000 transactions per second,” said Brent Keator, VP of Infrastructure for Nielsen Marketing Cloud (formally eXelate).

“Distributed technology from Clustrix combined with their world-class support team gave us the scale, performance and reliability that we could trust. With Clustrix becoming part of MariaDB, we’re excited to continue to use the technology that handles our high transaction workloads with ease while using the popular, modern MariaDB platform with its proven enterprise functionality.”

Also read: MariaDB open sources MariaDB TX 3.0 compatible with Oracle database

Clustrix is the second company to be acquired by MariaDB in 2018. Earlier this year, MariaDB also acquired MmmothDB to strengthen its analytics solution called MariaDB AX.

The financial terms of the deal were not disclosed.

Articles Web Security

Top 6 emerging cybersecurity and risk management trends: Gartner

One of the main security objective of all the organizations is to protect information confidentiality. The organizations must consider IT security and IT risk management as a part of the executive business planning. According to Gartner, the IT security objectives must be defined for the organization as a whole.

The analyst firm identified the emerging trends in cybersecurity and risk management that security leaders should harness to enhance the resilience of organization while uplifting their own position.

Top cybersecurity and risk management trends:

1. Business leaders realizing importance of cybersecurity for successful business

The senior business executives rarely considered IT security a board-level topic or a key part of digital business strategy. But the recent major cyberattacks like WannaCry and NotPetya that caused financial/brand damage and customer churn for organizations, have changed the sentiment of business leaders.

Finally, they are becoming aware of the impact of cybersecurity to achieve business goals and protect the reputation of organization.

2. Mandatory data protection practices impacting digital business plans

Personal information of customers is the lifeblood of all digital businesses. But, in the US alone, the number of companies that faced data breaches grew from nearly 100 in 2008 to over 600 in 2016.

With the rise in number of data breach incidents like Cambridge Analytics scandal or Equifax breach, the governments are issuing regulatory and legal data protection practices like Europe’s GDPR. These practices impact the digital business plans and demand more emphasis on data liabilities.

3. Cloud-first services becoming norm with advent of modern technologies

The modern technologies that require large amount of data are driving the adoption of cloud-delivered security products. These products provide more agile and adaptive solutions and can use the data in near real-time.

4. Machine learning to solve security issues

As per Gartner, machine learning will become a normal part of security solutions by 2025.

ML can efficiently address a number of security issues like adaptive authentication, insider threats, malware and advanced attackers.

5. New geopolitical risks in software and infrastructure buying decisions

Gartner identified that decisions of buying software and infrastructure are based on the geopolitical considerations of partners, suppliers, and jurisdictions. The trend is driven by rise in levels of cyber political interference, cyber warfare and government demands for backdoor access to software and services.

6. Centralized networks increasing the security risks

While there are numerous benefits of centralized networks, however, it is seriously threatening the organizational goals. Gartner said that if centralized ecosystem significantly affects the organization, then the decentralized architecture should be considered.

Suggested reading: Public cloud services revenue in India will reach $2.5 billion in 2018: Gartner

Gartner will discuss these trends at the Gartner Security & Risk Management Summit.

Page 1 of 3
1 2 3