Azure Machine Learning Erfahrungen

Azure Machine Learning is Simply Awesome!

Kommentare: It's really amazing. We're able to offer many innovative things to our clients without investing much on engineering.

Vorteile:

If you're new to Machine Learning there could be nothing better than Azure Machine Learning. The best part is without much coding knowledge you may leverage the benefits of Machine Learning. Intuitive UI and help materials make it even easier.

Nachteile:

It's really hard to tell because the way Microsoft is evolving in terms of Azure and Machine Learning is really remarkable. Almost every day there is new addition to make life even easier.

One of the best on cloud ML development Solution!

Kommentare: Azure Machine learning have allowed me to quickly load , transform my data and perform model training with ease without any need to worry about the system configurations or GPUs, etc It had provided all of it on demand.

Vorteile:

1. No worries on resources part as computation power can be easily increased 2. Takes care of end to end Machine learning lifecycle.3. Support Features for monitoring concept and model drifts 4. Allows to easily deploy your trained model through azure devops on cloud.5. Good pretrained model and datasets are available which is useful for experimenting6. Easy to setup and use with great documentations available from Microsoft

Nachteile:

1. Cost is mid-tier, not all features available in trial version2. Not a no code solution, prior knowledge is required to use3. UI can be made more simple as searching features can difficult sometimes.

Machine Learning for GUI users.

Kommentare: Human pattern recognition, image recognition. Automation

Vorteile:

When getting into machine learning topics, if you're not a coder, this is likely the option for you. Most traditional ML frameworks require you to code a lot and know a lot of things like convolutions and setting up certain input and outputs. Azure ML Studio makes a lot of that simpler by creating a lot of drag and drop features that can get you going pretty quick.

Nachteile:

Training isn't as straightforward as it may seem. The tutorial videos they have available are a little outdated and there are a few new features not mentioned in them. Also, the free tier is rather limited and is only intended for personal use or some proof-of-concept dev work. I personally feel that Azure ML Studio takes more time to get used to using than learning how to code something like TensorFlow.

Ease of use from VSCode

Vorteile:

Integrates directly to VSCode and now features Github Copilot, with a subscription as low as $10 a month.

Nachteile:

Does not have a GUI for ML architecture or integration to Visio like Google diagrammatic tool promises.

Best learning site to help advancement of Education

Kommentare: Azure has helped. Us alot for deploying machine Learning Modules ,Database ,file storage and sharing,testing ,quizes and other great features. Offered by Azure with First 12 Months free option.

Vorteile:

It has all the features that can be very useful for learning from machine and contribute to development of knowledge and wisdom with the best Microsoft Technology and features that make Azure Machine learning Model very useful .highly recommended for Educational and training Institutions.

Nachteile:

It should have some android app so that it can be deployed for Mobile learning since smartphones are changing the way of learning specially amid pandemic .

Some great capabilities here, but the number of machine learning models is limited.

Vorteile:

The machine learning models that they provide really work well. It is very easy to set up an account and access the possibilities, There are some services that the first usages are free, so you don't chew up money while learning the basics Very fast. Very detailed and clear billing. Allows you to run problems that require very large computer capacity without having to buy very expensive hardware. You pay just for what you use at reasonable rates. Very good overall web descriptions of the capabilities (especially when compared to Amazon, the largest of the providers of this class of software)

Nachteile:

Few models. There are many machine learning models. Azure, Amazon and each of the companies that make offerings only pick a few to provide. Therefore, you can only use for a problem which matches one of their models, or find a way to twist your problem to fit one of the models provided. At this time, this sees to be true of Azure and the others offering cloud-based machine learning. The level of support that Microsoft provides is much less than what they provide to Windows and Office The Sales people know little other than their talking points to advise whether Azure, in it's present state of development, can solve your problem. Moreover, Azure is growing so fast that they cannot hire & train sales people fast enough. It is difficult to estimate the cost of a particular machine learning operation in advance, and whereas Azure is very transparent with the unit prices, it is very hard to estimate how many of the different price element units you will need to run a problem. There is little or no help available now. Only Amazon does better in this area, and Azure is itself not strong here, but is better than all the other non Amazon competitors. While value for the money is terrific & ML is becoming affordable for almost all of us, Azure still seems to be a little more costly than AMS

Best available tool for machine learning process

Kommentare: Learning machine learning process to implement in AI apps.

Vorteile:

Microsoft Azure Machine Learning Studio is the tool which provides plenty of features which helps us to understand the machine learning process so that we can implement them in real time applications based on python and other programming languages. It helps to create the public demographics and web services that we can use them in various applications. It has tons of machine learning algorithms on the go.

Nachteile:

Azure Machine Learning Studio has vector size and supports limitations. Basic plans give very few space to use and experiments attempts.

Great ML as a Service option

Kommentare: We could focus on our application and the customers needs, istead of all the theory and security issues of the Machine Learning world.

Vorteile:

The Azure Machine Learning as a service platform is a great option because it provides all the security and tools necessary that allows your team to focus on the application, instead of the technicalities and the math behind all the machine learning algorithms.

Nachteile:

The set of features is large and very complete, but the price is really not accessible for avery company.

Machine Learning for Developers

Kommentare: This is a wonderful service for straightforward machine learning applications such as regression, that don't require deep learning frameworks. It also provides a convenient way of deploying the models thus less time is spent setting up the infrastructure and more time training and tuning the model.

Vorteile:

- Drag and drop configuration so no need to write actual code. As long as someone has a basic understanding of machine learning concepts, they can configure a working solution. - Ability to organize projects into separate distinct units, each with its own datasets, experiments and assets - One-click deployment of the machine learning solution as a service provided one has an Azure account. - Comprehensive documentation - You can write custom scripts in R or Python if you need to manipulate the data in some way not provided by the pre-built modules. There is also built in support for Azure Notebooks, so you can attach a notebook anywhere along the workflow, and visualize or manipulate the data in the notebook.

Nachteile:

- The free tier offers a limited number of training hours and functionality so you need to pay a small fee for the service through Azure - Lacks some specialized deep learning algorithms

Ease of use for non technical person

Kommentare: It was a very good experience in getting hands-on Azure Machine Learning. It was very simple to use & check how the output comes; later on, experimenting it I was able to get more technical understanding of concepts.

Vorteile:

The overall use of the software is pretty much simple and straight forward even a non-technical person can use Machine learning through drag and drop features. Most of them I like how the data life cycle is handled You can directly download data from multiple sources like links uploaded files.

Nachteile:

In Machine learning Studio designer if there are mode modules then handling those designer get bit clumsy. (Need to zoom in & out in complex models). Also Deploying a model takes much more time than expected

A Simplified Solution to Leverage Advanced Technologies.

Kommentare: As an Instructional Designer and data analyst, I use the Azure Machine learning service to generate insights and predictions based on students' learning records and grades. The platform recommends components and algorithms based on my goals. It helps me create data pipelines, build and train machine learning models, and finally deploy an end-to-end solution without writing a single line of code.

Vorteile:

Azure Machine learning service makes it easy for almost anyone to build (Basic) end-to-end artificial intelligence solutions on the cloud without writing a single line of code. The Azure Machine Learning Designer has a drag-and-drop user interface that allows users to connect data sets, choose among recommended algorithms, and build pipelines on a visual canvas. It's a No-code/Low-code cloud-based solution that empowers non-technical business users to leverage the benefits of the world's big data, intuitive machine learning algorithms, and cutting-edge artificial intelligence technologies. The service is secured, scalable, and well-documented.

Nachteile:

Although Azure Machine learning service facilitates the deployment of end-to-end solutions, it's still relatively complex and requires a solid computer science background and intermediate-level programming skills.

Huge improvements in the last 2 years

Kommentare: Overall, it has been positive. It's nice having a workspace where I can make VMs and not have to worry about breaking the Azure environment because it's compartmentalized now in Machine Learning.

Vorteile:

2 years ago, Azure's Machine Learning (known as Machine Learning Studio) tool option was a drag and drop tool that was slow and limited. Luckily, they got a lot of feedback from data scientists and people using their tools, and they modified their platform a lot so that a data scientist can provision Virtual Machines, connect to databases, and deploy trained machine learning models without needing infrastructure support. It's a great workspace for building models and deploying models now.

Nachteile:

The hardest parts, as usual, is learning specifically how to do something in Azure. If you're coming from AWS, you'll have to relearn how to prototype models and how to deploy them. It's also a little more expensive than standard Azure services, as seems to be the case with machine learning on any platform.

With an new machine learning platforms rolling out Azure gave us a head start

Kommentare: Azure's Machine Learning Studio is a welcome addition to the world of emerging technologies. As a research and development firm we have to experiment of these applications. Azure is a respected product and their foray into machine learning is a welcome one.

Vorteile:

Microsoft has a solid, reliable and well supported cloud environment that we've been using internally for years. Azure has been a part of what we do for as long as I've been at the company. The price is somewhat okay and the training materials are accessible enough. It's nice to enter the studio to your customized taste and continue to tweak it for maximum efficiency. Database creation/management is where this really shines.

Nachteile:

Pricing isn't the most competitive but we're lucky to be able to ignore that for the most part. Actual support isn't the best but the community does a good job supporting itself.

Great way to start Machine Learning using the Microsoft Data Stack

Vorteile:

Who knew that Machine Learning could be so easy! The drag and drop interface makes creating a repeatable data model and algorithm super easy. In addition, there are tons of practice models, data sets, and help items to get you up in running. The cost of this is pretty inexpensive, once you get thru training the model.

Nachteile:

The initial cost outlay can be a bit high as you train the model. In addition, not sure if this is really a con, but you do need to know the basics of machine learning. You still need to be trained as a data scientist (or at least starting to become one).

very good software to machine learning.

Vorteile:

easy to use for beginners in machine learning. no need lot of knowledge about machine learning.

Nachteile:

It will reduce new researchers in machine learning.

The future now

Kommentare: I use it at work

Vorteile:

I really like the usability of the platform, it permits to implement machine learning algorithms without write code, the UI permits you to move blocks and create a workflow, train a model with data (the import is really easy) and public a webservice with your machine learning KB. Very good.

Nachteile:

One of the cons that I coud see is that you cannot download a trained model, you can only use the webservice, I can understand this choice from MS, but I think it could be helpful, at least in a free/limited version.

The Best Application for Training a Machine Learning Model Easily

Kommentare: I have used Azure Machine Learning Studio nearly one year for training the data sets and got more reliable results.

Vorteile:

With the growing technology, Machine Learning is sustaining an important role. For training the machine learning model, Azure Machine Learning Studio provides an easy construction. Azure Machine Learning Studio has a proper guidance for the beginners. Features in the experiment are categorized in a proper way and using drag and drop method can be helped to make the experiment very quickly. Users are able to work easily because many experiment templates are provided in the Azure Machine Learning Studio. Worldwide people can use this application since it supports for few languages.

Nachteile:

This application requires to follows all the guidance for get a complete idea about training machine learning model. Since it provides documentations, reading the whole guidance can be a tedious work to user. Hence I suggest that providing video tutorials will be attractive rather than the documentations.

Easy to Use

Kommentare: The experience was fun because it has everything I need. The pricing is quite reasonable and does not add much of a financial strain.

Vorteile:

I love that this allows me to explore models and see what I can do with them. It was easy to analyze the data and finalize what was to be done. If you are not so good at ML, this is the service for you as it makes it easier for you to understand without having to code extensively. They have a plethora of options for all your needs.

Nachteile:

Sadly there were some issues with the UX which made it a bit difficult to navigate. Plus, there were not so many models as I expected and the fact that it was a bit difficult to use Python as the interface wasn't as intuitive. There were some hiccups when it came to complex models.

Great for beginners

Kommentare: I have tried to handle other data consolidation and programming tools and so far this has been the easiest for me. Furthermore, it can be integrated with other programming languages. However, it still needs some development.

Vorteile:

I am not an expert in technology and computer science issues and it seems to me that with azure it is easy to consolidate data and import information from the PC. You only need to watch a couple of videos to understand how it works and follow the instructions.

Nachteile:

The price is a bit high for independent consultants who do not have the support of large companies. It should also have a more comprehensive guide to using its features

So much better stuff out there

Vorteile:

It is easy to integrate with other Microsoft products

Nachteile:

It is not as robust or as supported as matlab’s machine learning toolkit. Also doesn’t work well with tensorflow, a python library, that almost 50% of machine learning developers use.

Data Learning Tool

Vorteile:

Useful software for learning of machine languages. It has great features and functionalities. User's friendly.

Nachteile:

The app is not available in IOS. Lacks some features.

Machine Learning Studio

Vorteile:

Easy to use. Loading data and deploying a model into a webservice is straight-forward Customization of the features concerning data transformation and algorithms are easy to do via fields.

Nachteile:

The gui is often difficult to see. Many of the features won't work without a specific order being followed which limits what's needed for the machine learning model.

Machine Learning with Azure

Kommentare: Great experience with initial hiccups but must application for ML

Vorteile:

Predictive analysis model is easy to build, test and deploy to generate value out of data. It is heavily demanded to save millions in maintenance cost. Very powerful tool and collaborative in nature.

Nachteile:

Needs understanding of the underlying algorithm to fully generate value out of it.

Faster machine learning insights

Vorteile:

If you are looking to get some insights from your data without much code, this is probably a great option for you. It has a graphic interface that allows you to use different tool without writing little to no code at all

Nachteile:

It is a bit of a learning curve to get up to speed and unleash the power it has

Azure Machine Learning Studio is a great to for doing predictive analytics but not intuitive to use

Kommentare: Not a lot of benefit as we didn't get accurate results from our predictive model.

Vorteile:

It's drag and drop functionality is great and the visuals are very helpful for finding anomalies in your data.

Nachteile:

The tool isn't very intuitive if you are not a data scientist. It would be great if there were longer descriptions on each of the types of algorithms