Welcome to The Long View—where we peruse the news of the week and strip it to the essentials. Let’s work out what really matters.
This week: GitHub touts unbelievable AI stats, and Intel’s beautiful open source typeface. (more…)
Tag: machine learning
-
92% of Devs Use AI, Survey Says ¦ Intel One Mono Font
-
Logz.io Taps AI to Surface Incident Response Recommendations
Logz.io this week added a supervised machine learning capability to its observability platform that reduces mean-time-to-remediation by surfacing recommendations for resolving incidents.
Asaf Yigal, vice president of product for Logz.io, said the Alert Recommendation capability added to the Logz.io Open360 platform uses artificial intelligence (AI) to model the steps a DevOps team needs to complete to resolve an incident.
The goal is to reduce the amount of time required to resolve incidents at a time when IT environments are becoming increasingly complex, he added.
In fact, a recent Logz.io survey found 75% of respondents said it currently takes them hours to resolve production issues, with only 14% satisfied with their current mean-time-to-resolution (MTTR). A total of 41% specifically cited monitoring and observability of Kubernetes environments as a primary challenge.
Alert Recommendation is the latest in a series of investments in AI that Logz.io has made. Previously, Logz.io integrated the ChatGPT generative artificial intelligence (AI) platform to surface links to related information and best practices for resolving IT issues.
In general, AI tools should make it possible to manage IT at a level of scale that eliminates many of the low-level data engineering and analytics tasks that previously required manual effort from a DevOps engineering team. If, for example, the observability platform is generating recommendations to address issues, there may be less of a need to create runbooks that DevOps teams typically create to address a wide range of known issues.
Along with making it simpler to discern the root cause of an issue, AI also makes it possible for less experienced members of a DevOps team to resolve an issue using guidance generated by the observability platform, noted Yidal.
In effect, AI technologies are reducing the cognitive load required to be an effective member of a DevOps team, he added.
One way or another, it’s not so much a question of whether AI will be applied to DevOps as much as it is to the degree. Many of the manual tasks that often conspire to create DevOps bottlenecks should be significantly reduced in the months ahead as more advances are made. The challenge now is determining how best to reallocate DevOps expertise in anticipation of those advances.
Of course, there’s always going to be some sense of trepidation when it comes to AI. However, many of the tasks that are about to become automated tend to be tedious. Many DevOps professionals would just as soon see those tasks become automated in the expectation that more time will become available to take on more complex challenges.
Regardless of the motivation, the way IT is managed is about to change. There may be plenty of instances where AI does not live up to its initial hype, but as AI models are exposed to more data, they will become more accurate. That doesn’t mean, however, there won’t always be a need for a DevOps engineer to ensure those algorithms are behaving as expected.
-
Technological Breakthroughs in Hurricane Forecasting
The 2023 Atlantic hurricane season marks a transformative era in tropical forecasting, driven by the remarkable progress in machine learning and big data analysis. It’s a rapidly evolving field, where technology and meteorology intersect to better prepare the public for the formidable forces of Mother Nature.
NOAA’s 2023 Atlantic Hurricane Season Forecast
The National Oceanic and Atmospheric Administration (NOAA) projects that this year’s Atlantic hurricane season will produce 12 to 17 named storms, with five to nine major (Category 3+) hurricanes. NOAA expresses 70% confidence in these ranges, bolstered by its evolving technology.
A summary infographic showing hurricane season probability and numbers of named storms predicted from NOAA’s 2023 Atlantic Hurricane Season Outlook. (Image credit: NOAA) Upcoming Upgrades: Supercomputing System Expansion
The recent implementation of the Probabilistic Storm Surge model improves storm surge forecasting for the U.S. mainland, Puerto Rico and the U.S. Virgin Islands. The model provides forecasters with a range of potential flooding scenarios, enhancing community preparedness.
NOAA plans to implement several more upgrades this summer, including a 20% expansion of its operational supercomputing system’s capacity. This enhancement will enable NOAA to run more complex forecast models, leading to more accurate predictions.
In late June, the Hurricane Analysis and Forecast System (HAFS) becomes operational. Running in tandem with the current Hurricane Weather Research and Forecast Model System, HAFS is projected to show a 10-15% improvement in track forecasts over existing models.
HAFS represents a significant leap forward in the integration of machine learning algorithms and sophisticated data analysis techniques. By leveraging state-of-the-art computational models and assimilating an extensive array of meteorological data, HAFS has the potential to unlock new insights into the behavior and trajectory of tropical cyclones. This fusion of technology and meteorology is projected to surpass the capabilities of existing models.
Harnessing Machine Learning for Advanced Forecasting
The core strength of HAFS lies in its ability to process vast amounts of atmospheric and oceanic data, utilizing advanced machine learning algorithms to identify patterns and correlations. These algorithms can discern intricate relationships between various meteorological variables, such as sea surface temperatures, atmospheric pressure, and wind patterns, enabling more precise forecasting. By comprehensively analyzing historical storm data and incorporating real-time observations, HAFS can generate forecasts with higher accuracy and longer lead times.
HAFS also harnesses the power of high-performance computing, allowing for faster and more complex simulations. This computational prowess enables the system to simulate a multitude of possible scenarios, accounting for various atmospheric conditions and factors that influence hurricane behavior. The ability to run multiple simulations and generate ensemble forecasts enhances the robustness of predictions and provides forecasters with a comprehensive range of possibilities.
Innovative Technologies: Drones, Saildrones, Underwater Gliders, and Sounding Balloons
Innovative technologies like small aircraft drone systems, additional Saildrones, underwater gliders, and WindBorne global sounding balloons will contribute to advancing knowledge and improving hurricane forecast accuracy.
Small aircraft drone systems offer a unique vantage point for data collection in the lower atmosphere, enabling meteorologists to gather critical information on temperature, humidity, wind patterns, and other atmospheric variables. These drones are equipped with advanced sensors and instruments, allowing them to navigate through hurricane environments and provide valuable real-time data that can significantly improve forecast models.
Saildrones, alternatively, are autonomous surface vehicles equipped with an array of meteorological and oceanographic sensors. They harness wind and solar energy for propulsion, enabling them to traverse vast oceanic regions and collect crucial data on sea surface temperatures, ocean currents, and other parameters relevant to hurricane formation and intensification. By expanding the fleet of Saildrones, scientists can enhance their observational capabilities and obtain a more comprehensive understanding of the complex interactions between the ocean and the atmosphere during hurricane events.
Underwater gliders, equipped with sensors and instruments, offer a unique perspective beneath the ocean’s surface. These autonomous vehicles can collect data on temperature, salinity, and ocean currents, providing valuable insights into the oceanic conditions that influence hurricane development and intensity. By combining data from underwater gliders with other observational platforms, scientists can gain a more holistic understanding of the underlying processes that fuel tropical cyclones.
WindBorne global sounding balloons ascend to the upper levels of the atmosphere, collecting data on temperature, humidity, and wind patterns. This information is vital for initializing and validating numerical models used in hurricane forecasting, enhancing the accuracy of predictions.
NOAA GOES satellite captures Hurricane Ian as it made landfall on the barrier island of Cayo Costa in southwest Florida on September 28, 2022. Improving El Niño & La Niña Forecasts
The Tropical Atmosphere Ocean buoy array is undergoing modernization and upgrades, bolstering the accuracy of El Niño and La Niña forecasts. These climate patterns have significant impacts on hurricane activity, and accurate predictions of their behavior are crucial for understanding the overall climate system.
By modernizing and enhancing the buoy array, scientists can gather more precise data on oceanic and atmospheric conditions associated with El Niño and La Niña, leading to more accurate long-range hurricane forecasts.
New tools and upgrades have also been introduced for more precise hurricane analysis and forecasting. The National Hurricane Center has expanded its Tropical Weather Outlook graphic forecast range to seven days from five. To give communities more time to prepare, the Weather Prediction Center has also extended the Excessive Rainfall Outlook by two days.
A summary graphic showing an alphabetical list of the 2023 Atlantic tropical cyclone names as selected by the World Meteorological Organization. The official start of the Atlantic hurricane season is June 1 and runs through November 30. (Image credit: NOAA) Bolstering Resilience Against Future Hurricanes
The integration of these innovative technologies and the modernization of existing observational systems expands observational capabilities both in the atmosphere and the ocean to track and monitor storms. The evolution of forecasting techniques holds promise for even greater resilience against future hurricanes, showcasing the remarkable impact of technological advancements in safeguarding coastal communities.
-
GitLab Allies With Google to Bring AI to DevOps
GitLab extended its alliance with Google as part of an effort to bring more generative artificial intelligence (AI) capabilities to DevOps workflows.
The GitLab suite of software-as-a-service (SaaS) applications already reside on the Google cloud platform to provide GitLab with the foundation of data required to train those AI models. Over the course of the last two months, GitLab has already added numerous capabilities that rely on multiple types of AI technologies.
For example, there is now an experimental Explain This Vulnerability capability that provides a natural language summary of an issue in a way developers and cybersecurity teams can easily comprehend.
Taylor McCaslin, a product group manager for data science and AI/machine learning for GitLab, said going forward, most of the AI focus is going to be on leveraging generative AI capabilities. Those capability will be enabled by Google using a large language model (LLM) that GitLab developed for DevOps workflows. That approach enables GitLab to surface more accurate recommendations based on validated data compared to the general purpose LLM that was used to create the ChatGPT service.
In addition, GitLab can continuously update the AI models it is running on the Google Vertex AI cloud service using the data from its SaaS application environment which is continuously monitored and updated, McCaslin noted.
It’s not clear what impact AI may have DevOps workflows, but GitLab is forecasting a 10x improvement. That will be accomplished by, for example, surfacing code that can be used to remediate a vulnerability. Today, many vulnerabilities are not addressed simply because developers don’t have enough time to write a patch.
A recent GitLab survey, however, suggested developers are already embracing AI to improve productivity, with 62% of developers using AI and machine learning algorithms to check code. More than a third (36%) also rely on AI and machine learning algorithms to review code.
At this juncture, the one thing that is certain is AI and other associated technologies are going to make developers more productive. It’s not nearly as apparent what impact increased amounts of code moving simultaneously through DevOps pipelines is going to have on the software engineers that manage those processes. The expectation is that similar types of AI advances will also enable more code to flow through those pipelines without, hopefully, further exacerbating any existing bottlenecks that might exist.
In the meantime, it’s clear the AI genie is out of the bottle. There will soon be more LLMs for all kinds of tasks. DevOps teams should start planning today based on the assumption that many manual tasks that conspire to make software engineering tedious are going to fade away. As such, the roles with a DevOps team are going to change and evolve. The assumption those DevOps teams should make is that these changes will be for the better. After all, the reason organizations embraced DevOps in the first place was to ruthlessly automate IT processes—AI is simply the latest iteration of that commitment.
-
Preparing for Machine Learning Readiness to Unlock its Full Potential
With the rapid advancement of artificial intelligence (AI) and machine learning (ML), companies need to understand and evaluate their readiness to adopt these technologies and drive material business outcomes.
To assess a company’s machine learning readiness, we must consider several key factors, including:
● Vision/strategy
● Data availability
● Expertise, infrastructure and resources
● MLOps and governanceIn this article, I will review these three key factors in assessing ML readiness and how they can help your organization prepare.
1. Vision/Strategy
A company’s vision for its AI/ML strategy is essential to consider when assessing its machine learning readiness. This includes ensuring that the expected AI transformation aligns with the company’s strategic goals and that it will ultimately be meaningful for the business. To evaluate this, organizations should consider their overall strategy for AI adoption and determine how it fits into their broader business objectives. Additionally, companies should identify specific use cases and applications of AI that support their goals and how those can be implemented to drive business outcomes. A company that has already woven an AI/ML strategy into its business planning is in a far better position to assess its ML readiness.
2. Data Availability
I believe data availability is another crucial factor in assessing ML readiness. Machine learning algorithms require vast amounts of data for training and validation to offer real value. It would be challenging for a company to successfully implement machine learning without access to sufficient data, or they’d risk a model that cannot be trained effectively and that would, in turn, perform poorly. Companies must consider the quality and quantity of data at their disposal, as well as any potential future data sources when assessing their readiness for this important element of machine learning. I also think organizations should consider the feasibility of collecting and integrating data from various sources, whether internal or external, to have the most impactful machine learning initiative.
3. Expertise, Infrastructure and Resources
It’s no secret machine learning requires a diverse set of skills, including data science, programming and statistics, and every organization must be sure they have all the necessary talent and resources when building that function. These resources include data preprocessing, model development, model validation, deployment, monitoring, maintenance and the ability to train and onboard new employees to oversee AI/ML programs. Companies should also enable collaboration across these different units and roles. If a company chooses to outsource machine learning projects to third parties, the skills needed internally may need to shift toward project management and collaboration with external teams, as well as a nuanced and sophisticated understanding of the business problem, industry and technical capabilities of the vendors, to ensure proper execution.
A company’s systems and technology stack are key to any operation. This includes selecting the right technologies and tools to enable the end-to-end creation and use of AI/ML-powered analytical applications. Before adopting ML, companies should evaluate their current stack to determine if it’s equipped to handle AI, including the availability of hardware, software and programming languages. Organizations should also consider the costs associated with selecting and implementing the right technology stack for AI/ML, including the costs of any necessary upgrades or new investments. Decision makers should also research and evaluate different technology vendors, their products and services, and find the one that best suits their needs in regards to cost, value and alignment with long-term goals.
4. MLOps and Governance
To be machine learning-ready, I find it important for companies to have machine learning ops (MLOps) in place. MLOps is a set of practices that allows companies to manage the end-to-end machine learning life cycle. It covers the entire machine learning process, from data preparation to model deployment and monitoring. Implementing MLOps includes continuous integration and delivery, model versioning and management and monitoring and alerting. Additionally, MLOps practices such as automated testing, model evaluation and deployment can help organizations improve the overall efficiency and effectiveness of their machine-learning efforts over time.
A company’s governance structure for AI/ML adoption is another factor to consider when assessing machine learning readiness. Companies should evaluate their existing governance structures and determine if they are equipped to handle the implementation of AI, including data governance, compliance and risk management. We must also consider how we will manage the deployment and scaling of AI, including the development of policies and procedures to govern the use of AI and the identification of internal stakeholders and decision-makers who will be responsible for its implementation.
Ultimately, I believe the key to machine learning readiness is to assess current capabilities, identify areas for improvement and develop strategies to address them. By taking the time to assess their machine learning readiness, companies can ensure they are in the best position to take full advantage of the benefits that machine learning can provide.
-
Tabnine Extends Generative AI Alliance With Google
Tabnine has extended its alliance with Google Cloud to advance the adoption of generative artificial intelligence (AI) to automate the writing and testing of code.
The generative AI platform provider has already developed its own large language model that is hosted on Google Cloud. Tabnine is now also committing to leveraging the large language model that Google is developing to automate software development life cycle processes.
Brandon Jung, vice president of ecosystems for Tabnine, said it’s apparent that application development and deployment teams will be making use of multiple large language models across SDLC workflows. Not all those large language models will be developed by one single platform provider, he noted.
Rather than being dogmatic about large language models, Tabnine plans to make it possible to invoke multiple models via application programming interfaces (APIs) that are integrated with development environments, he added.
The Tabnine platform supports multiple programming languages, including Python, Java and JavaScript and is designed to be integrated with integrated development environments (IDEs) such as Visual Studio Code and Jetbrains. The company also previously integrated its code completion tool with the GitLab continuous integration/continuous delivery (CI/CD) platform. The overall goal is to make it easier for developers to automatically write code based on custom models using approved source code hosted in a secure private repository.
Generative AI creates a large language model that assesses the probability of what the next line of code will be based on what has preceded it. It’s not likely DevOps teams will be replaced as generative AI is extended across workflows, but the overall size of those teams might shrink as it becomes possible to do more with fewer people. At the same time, the barrier to DevOps adoption will also fall as AI platforms make it simpler for more organizations to embrace DevOps best practices.
The challenge will be making sure the data collected to train AI models is of a high enough quality to ensure the desired outcome. Organizations will need to find a way to apply generative AI to data models that have been validated by DevOps teams.
One way or another, it’s only a matter of time before generative AI capabilities are applied more broadly. Platforms such as OpenAI’s ChatGPT are only the tip of an iceberg that impacts almost every manual process, including software development and deployment. The issue will be determining how quickly those innovations will become practical enough to employ.
In the meantime, it is already apparent generative AI platforms are having a significant impact on the rate at which code can be developed. Inevitably, that means the amount of code moving through DevOps pipelines at any one time should increase significantly. DevOps teams should expect generative AI technologies to be applied both before and after application code is built and deployed, noted Jung.
In the meantime, the biggest issue is, arguably, keeping up with a rate of generative AI innovation that is occurring faster than many organizations can absorb.
-
How Open Source Can Benefit AI Development
Enterprises are increasingly reliant upon open source software. A full 95% of IT leaders say open source tools are key to their enterprise infrastructure. Simultaneously, we’ve witnessed a sharp increase in the adoption of artificial intelligence (AI) and machine learning (ML). Most IT divisions want to take advantage of AI in some form, and open source technologies will likely be vital to these efforts.
I recently met with Moses Guttmann, co-founder and CEO of ClearML, to understand the benefits of adopting open source software within the sphere of artificial intelligence. To Guttmann, open source has many benefits—the visibility it grants can equate to helpful feedback from the developer community, and transparency can inform new feature development and identify potential security gaps. There are also countless open source tools to help organizations develop and deploy ML models.
Below, we’ll consider the benefits and potential drawbacks of incorporating OSS in AI development and explore ways to optimize the use of AI within an organization.
Benefits of Using Open Source for AI
AI has evolved in leaps and bounds. What was pure sci-fi a few years ago has become a reality through ChatGPT and other AI-powered platforms. The capabilities of AI are awe-inspiring, but Guttmann believes it still requires fine-tuning to move from entertainment purposes into a valid business driver.
For example, custom ML models are difficult to construct from scratch, and generic APIs aren’t always that customized to the use case at hand to be valuable. Instead, Guttman foresees more and more organizations downloading open pre-trained models and retraining or customizing them to meet their specific needs.
For software providers, there are many benefits to adopting an open source strategy. The visibility it creates helps evolve a project with feedback from the community, said Guttmann. This, in turn, helps organizations build better, more stable products. For example, ClearML adopts plenty of OSS to power their user interface, backend and other components, in addition to providing a popular open source CI/CD framework for ML workflows. “We’re true believers in the power of community and like to work with open source and contribute back to these projects,” said Guttman.
Use Cases for Open Source in AI Development
Open source software is powering innovation in many ways. And when it comes to AI, there are numerous open source projects to help accelerate AI development. Such projects encompass pre-trained models, data-structuring tools, frameworks to train models and AI deployment pipelines. For example, many open source machine learning frameworks can help engineers build and train their models from scratch. These free-to-use tools include TensorFlow, Keras, PyTorch, H2O, XGBoost and others.
However, training models from scratch is time-intensive and complicated, said Guttmann. Instead, another option is to use pre-trained models accessible on GitHub or other repositories. Some popular examples are PaddleNLP and Natural Language Toolkit for natural language processing, OpenCV for machine vision and DeepFaceLab for deepfakes. There are also many open source Python packages that can structure and clean data for ML model training purposes.
Open source software can also be applied to streamline the ML DevOps pipeline itself. AI requires intense processing and is typically deployed to its own computing environments. Infrastructure management around running and training AI is a promising area where open source could assist, said Guttmann. “Automating this process with an abstraction layer to allow data science and ML engineers to access product-grade hardware is a must,” said Guttmann.
Potential Downsides of Open Source Usage
Yet, of course, there are some potential downsides to using open source software within the field of AI. Supply chain attacks are still a persistent threat, and numerous open source software risks can present themselves. These include bad actors compromising a legitimate package, name confusion attacks, unmaintained software and untracked dependencies. License restrictions could also potentially block the inclusion of OSS in certain projects.
Thankfully, the benefits of OSS appear to outweigh these concerns. Automated vulnerability scanning techniques can help detect known issues. It also helps that 89% of technology leaders see enterprise open source as more secure or as secure as proprietary software, according to a Red Hat report.
Tips For Getting the Most From AI
Open source is powering digital innovation on many fronts, including AI. But, simply looping in OSS isn’t the only aspect of fully leveraging AI—another will be the operationalization of ML management. The right automation infrastructure in place could enable the agility necessary to continuously retrain models. And since there are potential gaps in the CI/CD process, Guttmann recommended separating AI/ML from your normal CI/CD for security reasons.
Guttmann also believes that, in general, companies should be more transparent around ML models, including sharing details behind their creation and visibility regarding the datasets fed into them. Overlooking this could lead to ethical concerns. Furthermore, limiting the “hallucinations” generative AI produces is another huge gap that must be bridged, said Guttmann. This could be solved by simply allowing the system to respond with “I don’t know,” instead of confidently spouting nonsense.
In summary, open source could bring many benefits to bespoke AI/ML adoption. Many mature open source tools exist to enable engineers to create advanced AI and oversee the management and deployment of custom ML models. But the real test will be retaining this open creed by upholding visibility and transparency into the end creations.
-
Six Mainframe DevOps Predictions for 2023
DevOps, or the automation of application development and hand-off to operations, is more prevalent than ever on the mainframe. As we begin 2023, below are six predictions for what the coming year will bring. The composite themes will undoubtedly be speed, alignment and experience.
1. Enterprise DevOps Will Break Down Silos Across Core Systems
DevOps practices will continue to expand beyond cloud and distributed systems to include more core systems than ever before. Core platforms, including the mainframe, are the functional backbone of mission-critical applications that drive the world economy. In a recent IDC survey, 82% of organizations reported taking steps to prioritize using the same application development tools across distributed and mainframe teams. 2023 will continue to see a rise in enterprise DevOps that will bring down the silos between platforms. As core system workloads increase, organizations will see the benefit of aligning platforms such as mainframe and distributed to realize the full benefit of their DevOps initiatives.
2. Growth in AI/ML as Key Components for Measuring DevOps Performance Metrics
The value of analytics will increase as organizations emphasize the use of key performance indicator (KPI) data to measure the performance and maturity of DevOps pipelines. In this way, organizations will not only gain more insights into their pipelines, but will also be able to proactively manage the data in real-time. They will be able to adjust loads as needed and create smart pipelines with advanced error handling. Organizations will leverage benchmarks within the analytics to see how they compare with their peers.
3. Cross-Functional Pipelines Will Create Hybrid Automated Everything
This will enable the cloud, core systems and data to deliver a more holistic view of “platforms, processes and people.” Silos will finally disappear. Automation will become a key priority in this effort as organizations seek to free up time for developers to innovate and achieve greater velocity, quality and efficiency. A single integrated and connected pipeline will emerge that automates the creation and testing of applications in a single flow. The rationale of prioritizing this alignment between distributed and core systems includes the reuse of best practices, workflow simplification, cost-effectiveness, application integration and streamlined training.
4. DevOps Will Embrace Platform Engineering and Site Reliability Engineering (SRE)
Platform engineering is the next evolution of DevOps, taking a DevOps platform and turning it into a self-service product to improve the developer experience. It treats DevOps as a product and allows organizations to create an end-to-end offering that serves developers, operations and platform teams. SRE focuses on the “Ops” side of DevOps, specifically how to automate operations capabilities. This is vital to the DevOps mission since it provides the data that allows organizations to manage what they can measure.
5. DevSecOps Will Play an Increasingly Central Role
Security must become a central part of the toolchain and a “first-class citizen” in the pipeline. Static application security testing (SAST), also known as white box testing, along with dynamic application security testing (DAST), also known as black box testing, will be further integrated into processes to enable security teams to scan large loads at rest or in real-time quickly and efficiently.
6. Git-Centric Enterprise Source Code Management (SCM) Strategy Will Grow
Git has become the de facto standard for SCM for distributed application workloads. In a recent IDC survey of industry professionals across 11 countries and multiple industries, 87% said they will migrate all or some of their mainframe source code to Git over the next three years. This means Git and Git-native development will become important capabilities for mainframe DevOps companies to support. Organizations will increasingly prefer implementation approaches and tools that offer flexibility in both scope and pace.
The techniques highlighted in these predictions are, naturally, related. They point to an ever-increasing capacity for efficiency and quality that will enable core and distributed technologies to work better together.
-
Graduating From DevOps to MLOps? 5 Tools to Help
DevOps is a set of practices that aims to bridge the gap between software development and operations. It aims to improve collaboration and communication between these two teams and to automate the process of software delivery so that changes can be made and deployed more quickly and easily. This can include continuous integration, continuous delivery and infrastructure-as-code. The goal is to enable faster delivery of new features and improvements to customers while also improving the reliability and stability of the software.
What is MLOps?
Machine learning operations (MLOps) is the practice of applying DevOps principles and practices to machine learning projects. It involves the collaboration of data scientists, machine learning engineers and operations teams to streamline the process of building, testing, deploying and maintaining machine learning models in a production environment.
The goal of MLOps is to increase the speed and reliability of deploying ML models while also improving the overall quality of the models. This can be done by using tools such as version control, continuous integration and automation of the model deployment process.
Why DevOps is Not Enough: The Benefits of MLOps
Machine learning models have unique requirements, such as data privacy, version control and model deployment that are not addressed by traditional DevOps practices. MLOps is specifically designed to address these unique requirements and to ensure that machine learning models are deployed and managed in a production environment in a way that is consistent with other software components.
There are several benefits to implementing MLOps in a machine learning project, above and beyond the benefits of DevOps:
- Improved collaboration: MLOps promotes collaboration between data scientists, machine learning engineers and operations teams, which helps to ensure that models are deployed quickly and efficiently.
- Faster deployment: By automating the model deployment process and using continuous integration and continuous delivery (CI/CD) practices, MLOps can help to speed up the time it takes to deploy a model to production.
- Improved model quality: By using version control and automated testing, MLOps can help to ensure that models are thoroughly tested and of high quality before they are deployed to production.
- Better monitoring and maintenance: MLOps can help to improve the monitoring and maintenance of deployed models by providing better visibility into the performance of models and enabling easy rollbacks if necessary.
- Better model governance: MLOps can help to improve model governance by providing a clear process for the management of models, including versioning, testing, and deployment.
5 MLOps Tools to Help Transition From DevOps to MLOps
Making the move from DevOps to MLOps typically requires dedicated tools that can help manage the complexity of machine learning models and datasets. Here are a few tools, some open source and some commercial, that can help.
MLflow
MLflow is an open source platform for managing the machine learning life cycle. It aims to make it easier for data scientists and machine learning engineers to develop, deploy and track machine learning models. MLflow provides several capabilities that can help streamline the machine learning development process, such as:
- Experiment tracking: MLflow allows data scientists to track and compare different versions of their models, including the code, parameters and metrics associated with each run.
- Model packaging: MLflow allows data scientists to package their models in a format that can be easily deployed to a variety of environments, such as a local machine or a cloud platform.
- Model deployment: MLflow provides tools for deploying models to a variety of environments, such as a REST API or a container.
- Project management: MLflow allows data scientists to organize their work into projects, including multiple models and associated files.
- Model registry: MLflow provides a centralized model registry that allows data scientists to store and manage models.
Pachyderm
Pachyderm is an open source platform that provides data versioning and pipeline management for machine learning and big data workloads. It allows data scientists and engineers to manage, version and collaborate on large datasets and create reproducible and reusable data pipelines.
Pachyderm uses a version control system (VCS) like git to version data, which allows data scientists to track and collaborate on data sets and enables reproducibility and rollbacks. This allows tracking the changes in the data, seeing who made the changes and when they were made and rollback to previous versions if needed.
It also provides a powerful pipeline management system, which allows engineers to create, test and deploy data processing pipelines. The pipeline management system is designed to work with containerized workloads, which makes it easy to deploy and manage pipelines.
Pachyderm is built on top of Kubernetes, which allows for easy scalability and can be run on-premises or on any cloud provider. It also supports data processing frameworks like Apache Spark and Apache Flink, as well as machine learning frameworks like TensorFlow and scikit-learn.
Amazon SageMaker
Amazon SageMaker is a fully-managed cloud-based machine learning platform. It provides a suite of tools for building, deploying and managing machine learning models. It is designed to make it easy for data scientists and developers to quickly and efficiently build, train and deploy machine learning models in a production environment.
Some of the key features of Amazon SageMaker include:
- Model building: Amazon SageMaker provides a variety of tools for building machine learning models, including pre-built algorithms and frameworks for popular libraries such as TensorFlow, Apache MXNet and scikit-learn.
- Model training: Amazon SageMaker allows data scientists and developers to train machine learning models on large-scale datasets, using distributed training and automatic scaling to improve performance and reduce training time.
- Model deployment: Amazon SageMaker makes it easy to deploy machine learning models to production, with a variety of options for hosting models, including Amazon SageMaker hosted endpoints and Amazon Elastic Container Service (ECS) and Kubernetes.
- Model management: Amazon SageMaker provides tools for monitoring, updating, and maintaining deployed machine learning models, including automatic A/B testing, canary deployments and automatic model rollbacks.
- Experiment management: SageMaker experiment management is a feature that allows data scientists to keep track of the different versions of their models, the parameters and the metrics associated with each run.
- AutoML: SageMaker AutoML is a feature that allows data scientists to automatically explore different algorithms and hyperparameters to find the best model for a given dataset.
H2O MLOps
H2O.ai’s MLOps is a platform that allows data scientists and engineers to collaborate and deploy machine learning models in production. It provides a set of tools for model development, management, and deployment, as well as monitoring and governance capabilities.
It aims to streamline the process of taking machine learning models from development to production while maintaining high standards of model quality and performance. Some of the main features include:
- Model Development: It provides a set of tools for developing and testing machine learning models, including support for popular frameworks such as TensorFlow, PyTorch and H2O.
- Model Management: It allows users to organize and manage models in a central repository, making it easy to track versions and collaborate with other team members.
- Deployment: It provides an easy way to deploy machine learning models in a variety of environments, including on-premises, in the cloud or at the edge.
- Monitoring: It offers monitoring capabilities that allow users to track the performance and quality of deployed models in real-time and automatically detect and diagnose issues.
- Governance: It provides governance capabilities to help ensure compliance with regulatory requirements, such as data privacy, and to help manage the risks associated with deploying machine learning models in production.
Neptune.AI
Neptune.ai is an open source platform that provides a suite of tools for organizing, tracking and reproducing machine learning experiments. It is designed to help data scientists and machine learning engineers to keep track of their work, collaborate more effectively and improve reproducibility.
Some of the key features of Neptune.AI include:
- Experiment tracking: Neptune allows data scientists to track and compare different versions of their models, including the code, parameters and metrics associated with each run.
- Experiment management: Neptune provides a user-friendly web interface for organizing and managing experiments, which makes it easy to keep track of progress and share results with others.
- Reproducibility: Neptune provides a way to reproduce experiments by tracking the code, data, environment and dependencies used in each run.
- Collaboration: Neptune allows data scientists to share and collaborate on experiments with others and provides a way to track who made changes and when they were made.
- Model management: Neptune provides a centralized model management system that allows data scientists to store and manage their models, including versioning, testing and deployment.
Conclusion
In conclusion, MLOps is a practice that applies DevOps principles and practices to machine learning projects to streamline the process of building, testing, deploying and maintaining machine learning models in a production environment. The goal of MLOps is to increase the speed and reliability of deploying ML models while also improving the overall quality of the models.
There are several tools available that can help organizations to implement MLOps, including MLflow, Pachyderm, Amazon SageMaker, H2O MLOps and Neptune.AI. Each of these tools provides a different set of features and capabilities, but all aim to make it easier for data scientists and machine learning engineers to develop, deploy and track machine learning models.
-
ClearML Releases New Reports Feature to Share Real-Time Results of Machine Learning Projects and Ignite ML Collaboration Across the Enterprise
Tel Aviv, IL, January 19, 2023 – ClearML, the leading open source, end-to-end MLOps platform, today announced it has released a new Reports feature that is now generally available. This new feature makes it easy to create and share real-time reports within ClearML as well as connect to third-party editors such as Confluence, Monday, Notion, Colab (Jupyter), and others using embed code. A report can contain charts that auto-update continuously throughout an experiment lifecycle.
ClearML users can now easily create, collaborate on, and share reports, as well as graphs and charts, in order to summarize and explain experiments and how model versions improve, analyze results, show experiment comparisons, discuss bugs, demonstrate progress towards milestones, and publish their work.
“Reporting and showcasing your findings to colleagues, managers, or even your future self is a core component of any modern collaborative workflow,” said Moses Guttmann, Co-founder and CEO of ClearML. “Having one central place where you can seamlessly report, analyze, plan, and collaborate on your work in real time makes it that much easier. ClearML-Reports connects seamlessly with existing ClearML functionality and is based on markdown, so it’s very easy to export or connect to an external reporting tool.”
Features and Benefits of ClearML Reports
The new ClearML Reports feature was designed with a multitude of functionality in mind, including:
- Creation: It’s now easy to write reports using markdown language, and there is a handy markdown cheatsheet to get started. Reports can be filtered by tags or archived like datasets or experiments.
- Visualization: Users can embed scalars, plots, or debug samples from the ClearML experiment manager right into the markdown editor with preview updates. They can even embed descriptive statistics of their datasets or comparisons graphs from multiple experiments — what’s important is that these are not HTML exports, but are still connected to the back-end such that when new data comes in, it will be visible after a report refresh.
- Team Collaboration: Teammates with ClearML logins can edit reports that are not published as read-only. Reports are also searchable based on their descriptions.
- Sharing of Reports: Users can share on-platform with others who have a ClearML login and can export to PDF for a time-bound report to share with anyone else. Publishing the report will make it read-only, so it can be used as a single source of truth among colleagues. Users can easily export your reports out of ClearML as markdown or PDF.
- External Tooling Integration: Many users and their teams each have their own workflows with their own toolchains. By not limiting the reporting functionality to only ClearML reports, users can easily integrate ClearML elements into their favorite third-party reporting tools, like Notion, Confluence, Monday, and others.
“Reporting our findings on the weekly standups was quite manual,” said Omri Bar, AI Research Team Leader at Theator, the creator of Surgical Intelligence. “I’d have to jump around between experiments, moving from one plot to another, narrating what was interesting about the graph in question. That’s not very easy for my colleagues to follow along with. Now I can prepare a neat report ahead of time, with all the interesting parts embedded inside, important notes right next to them, cleanly organized by topic. It has made our communication much smoother.”
See the power of ClearML Reports for yourself by watching the overview video now.
Get started with ClearML by using our free tier servers or by hosting your own. Read our documentation here. You’ll find more in-depth tutorials about ClearML on our YouTube channel and we also have a very active Slack channel for anyone that needs help. If you need to scale your ML pipelines and data abstraction or need unmatched performance and control, please request a demo. To learn more about ClearML, please visit: https://clear.ml/.
About ClearML
ClearML is a unified, open source platform for continuous machine learning (ML), trusted by forward-thinking Data Scientists, ML Engineers, DevOps, and decision makers at leading Fortune 500 companies, enterprises, academia, and innovative start-ups worldwide. We enable customers to build continuous ML workflows — from experiment management and orchestration through data management and scheduling, followed by provisioning and serving — to achieve the fastest time to ML production, fastest time to value, and increased performance. In this way, ClearML accelerates ML adoption across business units, helping companies reach their revenue potential and materialize their ML investments. With thousands of deployments and a vibrant, engaged community, ClearML is transforming the ML space — bridging software, machine learning, and automation. To learn more, visit the company’s website at https://clear.ml.