About VisionWizard
A place to share quality research ideas in AI for everyone
“Life can only be understood backwards; but it must be lived forwards.”
― Søren Kierkegaard
Hello, from VisionWizard. We a team of engineers/researchers working in AI started VisionWizard in one of the worst human adversaries in the history of mankind — Covid19 Pandemic, with a basic need to bridge the gap between research and development by simplifying the explanations of research papers in the field of Computer Vision and Deep Learning.
We ourselves are working every day to solve complex computer vision/deep learning problems as a part of our daily jobs. We know first hand, how much time is wasted behind reading the research papers and not understanding more than half of it due to bloated unwanted explanations written merely to fulfill the minimum length requirement for publishing it. Not all papers do this, but more often than not, this is the case. The readers lose interest and end up wasting their time. This is the sole reason why we started this publication.
By no means, we are professional writers but we are skilled in dissecting complex research papers into chunks of simple ideas that can be easily understood and used for further implementations.
While we can’t cover every research paper, we will try to unravel the most important ones. Your suggestions in the comment sections or mail would be most welcome :)
Pure explanations of complex ideas can help consume less time and help understand things from a larger perspective.
Why Deep Learning/Computer Vision research is important?
After 2012, when Deep Learning actually took off(thanks to AlexNet), research in the field has grown exponentially. Small ventures and MNCs started capitalizing on it to the fullest and a plethora of startups and the labs were founded. It would be fair to say that many recent tech ventures are considering an AI element in their product because of the tremendous value.
Additionally, many universities have added the course of Machine Vision/Artificial Intelligence in undergraduate programs. Right from their academia, students have started building their ideas involving machine/deep learning.
The advent of this motif has made a huge community of upcoming/reigning researchers and developers working on a problem and getting the best out of it. Everyday detailed research papers are published with vast improvements in specific topics of Computer Vision using Deep Learning.
To keep up the pace with these new advancements is an integral part of any R&D team to be ahead in the race for best results. Organizing and finding the best fit for the specific problem at hand is a challenging task for any team. To achieve this, the research team has to equip themselves with the latest and greatest of the research.
Given a huge amount of research in works, you cannot read every paper and thus it becomes important for people like us to bring refined ideas that matter most to the readers like you.
The value proposition to our readers
As mentioned, to ease the process from research idea to implementation to deploying it in production, the vital part is an in-depth study of a research paper’s novel presentations and feasibility of its implementation.
We address following almost all the challenges when we write about a paper or any topic for that matter.
Some challenges faced during a research paper reading
- Readability — It is hard to get through the flow of words presented in paper due to small font and high word density.
- Redundant information — A trend of repetitive information is seen to increase the length of the paper, which makes it hard to connect the dots between the essential points.
- Recursive reading dependencies —There are many concepts and algorithms which influenced the author that results in the final research paper. For a new reader, it becomes very tedious to get the topics in one place and have to do recursive searches from paper to paper wasting a lot of time.
- Less focus on implementation details — Most of the papers are more theoretical based and have rigorous mathematics and theories written to support their premise. Very little information/steps are present on how to implement it.
- Improper flow of information — Transient jumps are seen in the information that makes it more difficult for the reader that results in disinterest in the early stage of reading itself.
- High-level visualizations — Pictorial representation is one of the best ways for a reader to understand concepts. Fewer visualizations and higher bias towards theory make it mundane and time-consuming to understand the actual point.
So why not rewrite a research paper following proper algorithmic flow addressing above issues, visuals, and most important in more simplistic yet effective explanations?
VisionWizard is here to solve these common issues faced by many that can save a lot of time to grasp the information and make it steadfast for implementation purposes.
Consider this publication as your unofficial arxiv journal, where you will get the breakdowns of these exciting research papers full of visualizations, and clear information saving your time and effort.
EDITORIAL TEAM
We are a small team of technology enthusiasts working endlessly to get the best research ideas out there.
Shreejal Trivedi is a young desirous of deep learning and computer vision. His main focus is to leverage upcoming research work along with present establishments, to find the best recipe for a specific problem.
LinkedIn: https://www.linkedin.com/in/shreejalt/
Deval Shah is an AI tech enthusiast. He works on large scale surveillance problems using computer vision/deep learning at a disrupting startup. Loves to read and write about upcoming research trends. Always open to new ideas and discussions.
LinkedIn: https://www.linkedin.com/in/dvlshah
FINAL REMARKS
VisionWizard’s main aim is to bridge the gap between quality research work and its application across different spectrums. As discussed, the community of the researchers is growing at an immense rate day by day. These beautiful research ideas and pattern workflows have helped entrepreneurs and developers immensely to automate things beyond our imagination.
Fellow researchers can contribute and gather at one place and help the readers to excel and pacify the challenges faced.
Let’s get together to make a strong and impacting force in the field of Computer Vision and Deep Learning.
If you are interested to join our team of writers and make your contribution to world of AI, please go through the article on how to contribute.
“Learning to share” is better “to share what we have learned”
— Team VW :)
