DeepStack — World first self-managed AI Server
The field of Artificial Intelligence has progressed incredibly with the emergence of powerful algorithms like Deep Learning, High Performance hardware like the NVIDIA GPUs as well the massive amount of textual, image,sound and numerical data generated on the internet and embedded devices. This progress has led tremendous possibilities in Computer Vision, Natural Language processing, Speech and Data Analytics.
Like every other technological transformation, today’s modern AI started at a very low-abstract level in which only a few with the technical expertise as well as the resources can actually harness its potentials and deploy in an applicable use case. Today, most application of modern AI is done by companies and government agencies with the finance to hire the few available experts that can build an entire AI solution stack.
To fix the technical challenges faced in deploying modern AI, a number of cognitive cloud services emerged and started offering AI resources via APIs on a pay-as-you-use basis. As convenient as these services may be when compared with the task of setting up the entire AI solution stack in-house, there are a number of strings attached to these cognitive cloud solutions; a number of them are:
- High Cost: It is expensive to deploy in real-time and long-term basis use cases, considering API costs and bandwidth costs when large-sized input data like images and videos are involved
- Network Dependency: There is need to maintain a constant supply of reliable connection to the internet which means places with little or no internet resources are alienated from these solutions.
- Privacy: Every single input data generated will have to be sent to third party servers irrespective of the confidentiality of the data.
At DeepQuest AI, we recognize the impact today’s state-of-the-art AI will have on every field of study and practice. We decided to built an ultimate AI solution that will ensure you can easily setup, integrate and deploy state-of-the-art AI resources in-house. Therefore, we started building DeepStack.
DeepStack is a dockerized AI server that provides all the simplicity cognitive cloud services provides through the use of API endpoints but with the fact that it can installed on your own computer system or private cloud-based virtual machine and it works completely offline and without connecting to a third party server. In short, DeepStack once installed and running on your computer system will provide a central AI engine with an IP address like http://10.0.221.267 that you can send data to it via endpoints like /detection, /face, /scene from any computer, smartphone or IoT device within a Local Area Network (e.g Wifi). This ensures the AI Server can work with any software and application built with most programming languages like NodeJS, Python, Java, Ruby, PHP, C# and more.
The first version of DeepStack will be available starting from 21st of December, 2018 for Insider Preview Program where a few selected individuals will get a copy of DeepStack for free with one-on-one support provided along with sample codes, sample client applications and their source code, documentations and more via a dedicated slack group. These version will support a number of computer vision features such as object detection, face detection, scene recognition, action recognition and more.
The major release of DeepStack (version 2.0)will be available before the end of 1st Quarter of 2019 . You can join the waiting list to have first hand experience of DeepStack 2.0 via this link.
At DeepQuest AI, Our Mission which we choose to accept is to advance and democratize Artificial Intelligence and make it accessible to every individual and corporate entity of all sizes on the planet.
Visit our website via https://www.deepquestai.com to learn more about our open source projects, tutorials, free machine learning and deep learning e-book, datasets and solutions.