Web · 2018-05-11

Build deep learning models with a custom interface – #Startups

Artificial Intelligence (AI) is a very complicated field. As such, only the crème of computer science professionals get to understand & use it. Even more complex is the subset of Machine Learning (ML) as it requires 1 to get through a whole mesh of frameworks & jargon in the field before they can effectively use it. This can make it quite difficult for any Tom, Dick, & Harry to execute ideas that they might have involving its use. As such the uses of such AI were previously very limited. Lobe is a Web-based visual tool looking to make this field easier to get into.

Mike Matas, one of the creators & co-founders of the Web app has experienced firsthand how difficult it can be to break through into the field. This was 1 of the reasons that he & his team felt this a worthy endeavor. The initial goal for them was to make ML more understandable, & as a result more accessible to a larger population of people. Mike has worked with co-founders Markus Beissinger & Adam Menges to bring lobe to fruition.

Using this technology 1 can easily make groundbreaking apps without needing to learn code. Users are required to drag & drop their training data onto the platform, prompting the app to build a custom deep learning model. The app then uses Cloud-based computing to train itself to produce desired results. The fact that this is all done on the Cloud makes it a lot less taxing on personal computers and hence no reason for a PC to slow down its performance.

Once the learning part is done the user can then export a fully trained model to CoreML & TensorFlow. This should allow the model to be used on various platforms inclusive of iOS & Android. Another way for a user to integrate their model in an app is by using Lobe’s developer application programming interface. This avails code snippets in several different languages that one can copy & paste into any app they want.

The custom deep learning models mentioned earlier consist of smart building blocks currently known as lobes. For any individual seeking to customize their models, they can add or remove these lobes as they see fit. These lobes also consist of several layers underneath which users are at liberty to edit & fine tune to their individual needs.

With its operating base in California, USA lobe was launched in early 2018. The current version of the tool is in beta mode. While the use of the tool is free the team running lobe is only allowing in a few people for now.

Here’s how lobe works:

 

– This is a startup profile based on publicly available material & not a review –

Don’t leave without reading this…

… we seek a favor. Online advertising revenues are falling, especially for media Sites. Also, as you may have realized, unlike many media & publishing organizations, we haven’t put up a paywall – we firmly believe that information, including our content must be accessible by all across the globe. But content takes a time & money to produce. If you’ve read the above piece & liked it, we urge you to contribute in our support.

 

 


 

Click here to opt-out of Google Analytics