Help us shape the future of deep learning in .NET

Brigit Murtaugh

Deep learning is a subset of machine learning used for tasks such as image classification, object detection, and natural language processing. It uses algorithms known as neural networks to learn and make predictions on image, sound, or text data.

Neural networks learn from experience, just like we do as humans. Similar to how we may try an activity and adjust our actions based on the outcome, these algorithms perform a task repeatedly and tweak various actions or variables each time to improve their results, making them very powerful and intelligent.

We’d love to learn more about your current or prospective usage of deep learning in .NET through the quick 10-minute survey below. Whether you’re already implementing deep learning or just starting to learn about it now, we want to hear from you. We’ll use your feedback to drive the direction of deep learning support in .NET!

5 comments

Discussion is closed. Login to edit/delete existing comments.

  • Max Mustermueller 0

    Biggest issue, in my opinion, is the way you get into machine learning. There are tools, there are examples but everything is based on “We got this. We do this. And this are our results”. Great but what if I have something slightly different? Something else? Expect something else?

    As a developer and I think I speak for many developers, it’s a dream to have a machine understanding what we want and do what we want to do for our specific scenario. The only question is, how do we get there?

    Someone wants the “machine” to suggest a text (search entry?) based on previous entered texts. Someone else wants the machine to filter out values based on “we always filtered out these values matching a criteria”. Another one wants to recommend something based on what a user liked. The usage scenarios are very unique and different and I am missing a guide how to rewrite an existing example to my criteria. Because all those existing examples are very static and the comments in their codes explain what they do but never how to get there and how it should be modified for another scenario.

    So as of now ML seems to be a topic for very advanced users. All I want is that ML is more user friendly. Because that’s why I started with coding at all and why I have chosen C#. Because tons of tutorials and simple steps and fast progress and for every question you will find an answer within a second. I really want to get into ML but as of now, I feel like I am stuck, because I understand all examples but I have no clue to use ML for my own things.

    • Mahmudul Hasan 0

      You made a point. Everything seems like static.

    • Christian Klassen 0

      In my experience it’s not advisable to try to cut corners when trying to get into deep learning.

      I would suggest to bite the bullet and buy a book like “Deep Learning” by Ian Goodfellow et al. and work through it.

  • Akif Faruk Nane 0

    Hi, I have been developing a deep learning library in .NET 5 as well. In my opinion, users and researchers want full access of the library and the library should be easy to use. My library is based on a dynamic graph and I try to make it as flexible as possible. Here is the link: https://github.com/faruknane/DeepLearningFramework

    I’d be glad if I get support from community.

  • Cihan Yakar 0

    I did not encounter CNTK in the survey. Isn’t that a bit strange?

Feedback usabilla icon