Machine learning and real-time applications are gaining attention in software engineering because of their applicability in many fields such as security, networking, and embedded devices says Peter DeCaprio. In this article, we introduce an approach that aims to create a “one-stop-shop” for machine learning and real-time tools by integrating the power of community sites like GitHub with an efficient search algorithm. For example, we can find numerous state-of-the-art deep learning libraries from a single site. Furthermore, our search engine allows you to select multiple features when searching. Tools that have been selected will be listed on a results page which you can drill down into by clicking on the title of each tool. The information from the tool profile is pulled from a number of open source sites, such as GitHub and the Library Link Network.
This article will walk you through the various types of tools that we have included in our site and show you how to use our search engine to find tools relevant to your work.
Types of Tools:
Currently, there are four types of tools that we support. If a tool belongs to more than one category, it will be listed under each category.
1) Machine Learning Tool:
A machine learning tool is an application or program which can be used for machine learning tasks such as clustering and classification.
2) Real-Time Tool:
A real-time tool is an application or program which can encode and decode data in real-time without loss of packets (examples include UDP/TCP/HTTP).
3) Advanced Application:
An advanced application is an application that provides some unique functionality not found in the other types of tools.
4) Data Format Library:
A data format library is a set of routines or programs to handle common file formats such as JPEG, PNG, TIFF, etc.
Search Engine Usage: Searching for Tools on our site is easy and straightforward. To search for a tool type in one or more keywords into the “Keywords” field (separate multiple keywords with commas). Then click on the search button. After clicking on the search button, you will be directed to a results page listing all of the tools that match your criteria. If you want to refine your search further, click on the “Filter” button next to the results page and make your selections from the various fields (note: depending on how many options you choose, this may take some time). After clicking on the filter button, the list of tools will be updated according to your choices. Clicking on a tool title will give you more information about that particular tool explains Peter DeCaprio.
We plan to expand our site by adding additional features such as mechanisms for users to engage in discussions with each other regarding tools and algorithms, a platform for users to create an author profile so they can share their experiences with various tools, and a mechanism for users to rate and review tools.
Here are some FAQs recently asked by our users:
1) How do I add a tool to your database?
To add a tool to our database, simply click on the “Add” button at the top right corner of our site. Fill out the form and submit it. Our team will review your submission and get back to you soon. We ask that you provide a link to a repository where we can retrieve information about your tool/algorithm if possible says Peter DeCaprio. We also strongly recommend that you include a README file with detailed installation instructions, usage examples, etc. The final decision of whether or not to add your tool is up to our editorial board members. So please be patient while they make this decision.
2) If I have multiple tools in my account can I use them in one project?
Yes. In your account, the tools you add become “sub-tools” under a root tool in your account. For example, if you add three tools: A, B, and C; tool A is then its own sub-tool within tool B, while tool B becomes a sub-tool of tool C. You can then use any of these tools in an experiment (ex: using Torch and Tensorflow together).
3) Why don’t I see some common machine learning libraries such as Scikit or PyTorch on your site?
We are currently focusing on state-of-the-art open-source libraries only. We do not list Ruby/Rails libraries or commercial software such as MATLAB unless they are open source.
4) Can I get paid for referring users to your site?
No, not at this time. We are currently focusing on improving the functionality of the site. And don’t have any plans regarding monetization at this juncture. Stay tuned though!.
5) How does one become an editor or reviewer?
We are in need of additional editors and reviewers to improve the quality of our data set.
This is a great compilation of links for various machine learning and real-time applications. Peter DeCaprio says the site host several tools which can be downloaded to build research projects; each tool has its own page with detailed instructions on how to use the software.