Frequently Asked Questions
What’s computer vision?The ultimate goal of the computer vision is to enable computers to see and analyze the world through their artificial vision. Just like what humans do. 80% of humans’ brain capacity is directly and or indirectly occupied with tasks related to human vision, from seemingly easy tasks like detecting the differences in color and shape to extremely difficult tasks like finding the closest route on a map made of hundreds of alleys, avenues and boulevards. Your vision helps you taste a food you’ve never eaten before in your life or even smelled, simply by looking at pictures and it really is a marvel of around a billion years of evolutionary choices. While it might seem impossible to mimic what the human brain does with computers, that’s about to change.
What’s the difference between A.I, Machine Learning and Deep Learning?A.I. (Artificial Intelligence) is when computers try to mimic human behaviors, scripted or not. For example if you write a program to say Good Morning everyday on 9:00 a.m, as basic as it is, it’s still considered A.I as it mimics a human behavior. When computers are programmed to generate a totally new, previously unavailable data based on existing data according a mathematical rule, it is called Machine Learning. For example if you give a computer a time-series data on rainy days and an extrapolation function to predict new data based on the existing data and predict wether today is gonna be a rainy day or not and suggest the user to take their umbrella with them, it’s called Machine Learning as the machine (computer) is learning based on the data available to them. Note that in this case the type of data, its structure and the procedures the computer had to take was completely determined beforehand. Every Machine Learning is technically an A.I as it is trying to mimic a human behavior (but not vice versa as the example for A.I is not considered M.L). But what if we program the computer to not only be able to predict new data from the structured data we gave him, but program it to be able to analyze any data, even the data it never saw before? This is where we enter Deep Learning’s domain. In our Machine Learning example, computer was well aware that it was dealing with weather data, it knew what represented a rainy day and what represented a sunny day, moreover it knew the extrapolation algorithm and how to apply it to the data. In Deep Learning computer has no clue what kind of data it will be fed and their characteristics. In deep learning, computer is totally responsible for identifying patterns and matching new data, just like how a human being learns new stuff. For example, in deep learning you can show 100 images of different people to computer, they might have different physical features, different color, different height, different race, and the computer will learn what a human looks like. Computers are able to do this thanks to something called ANNs (Artificial Neural Networks), these are large networks of interwoven mathematical functions just like the neural network of a human body.
What is AIEX?AIEX is a supervised Computer Vision Deep Learning platform that provides you with all the tools necessary for a complete Deep Learning workflow. It helps you create models that can identify the visual features most important to you in images or videos with unprecedented accuracy and speed 24/7. Everything from data management tools to model training and finally deploying the trained models. You can easily transform your visual inspections using the trained models and save on time and money, increase accuracy and speed.
I’m concerned about my sensitive data privacy, what can I do?In order to have a CV (=Computer Vision) model, you have to train an algorithm using the actual images containing the objects that you want your model to identify and predict. These images need to be uploaded to the platform once and you can delete them right after the training is done. No trace of them is kept on our platform whatsoever and we keep them all encrypted for the duration of time their kept on our server (wether you wish to delete them or not). Not only that, but we have implemented numerous safety measures to make sure your data is safe wit us. If you want to deploy your model on AIEX cloud, your model will be kept on a unique GPU, inaccessible to anyone but yourself and we won’t keep any of the data you send to the model, they are deleted the moment the inference process is done and no trace of it will be kept on our systems.
Do I need to know any computer programming language?No! Our platform is built based on ease of use, you can traion your advanced models and deploy them without writing even a single line of code. Though your use case might require additional programming on your behalf (for example you miht want to program your alarm system to stop the machines if a defect is detected in the system) but they have nothing to do with your model training.
I want to get on the A.I train but I know barely anything about it, can I use your platform?Absolutely! Not only our platform is created with ease of use and simplivity in mind, enabling anyone to easily train models with just a few clicks, but we have an extensive knowledge base guiding you and helping you to get to where you aim to go.
I think I don’t have enough data, what can I do about it?We have implemented several tools to help users with less than desired datasets. One of these tools is the itegration of Google image engine which will give you instant access to billions of images hosted on web on the fly, in case you think you need special data not available anywhere on the web (which is very unikely), we have implemented Augmentation functions that can multiply your existing dataset in a meaningful to have a better trained model.
I don’t have the hardware powerful enough to host my trained model or don’t have the technical networking knowledge to deploy them on my own network.You can easily with a single click of a button deploy yuor models on AIEX cloud, a cloud computing service with the most sophisticated hardware to run your trained models and use our simplified APIs for inferrences.
I want to deploy my trained model on my own hardware, can I do it?You can take your models off of AIEX and deploy anytime anywhere. We have implemented services to make sure smooth deployments on Linux and Windows powered PCs as well and nVidia’s Jetson Nano. You can opt to deploy using Triton or DeepStream either on CPU or GPU. (Compatibility may vary based on the framework/model).
I have used other deep learning platforms and the results weren’t satisfactory, can you improve on it?AIEX is fully compatible with different file types so you can easily import your existing data over to AIEX, on the contrary to our competition, we offer dozens of algorithms across 4 major frameworks, namely TensorFlow, PyTorch, nVidia TAO and TensorFlow Lite and the user have vomplete control over the algorithm and framwwork of their choice. Not only that, but we have implemented advanced hyperparameter controls to fine tune your models.
Still have a question?
If you couldn’t find your answer in our FAQ, Fill This form
and send your question, We will answer to you shortly.