A Beginner’s Guide to Document Understanding
While some of you are familiar with the topic of document understanding and the software applications related to it, many of you are new in this field. To help you to start before going through all the posts and the archives I have compiled this overview as an entry point and introduction.
Document Understanding is the general term that I use to combine a variety of concepts and technologies that are typically named with technical words: Recognition, classification, categorization, extraction, text analytics, semantics, document analysis and machine learning. All these terms describe the application of software algorithms to mimic the human abilities to read a document, understand its meaning and draw conclusions. We also call these abilities the human cognitive capabilities. There are many approaches today to build some of these capabilities into software and hardware products to make humans more productive and to automate repetitive tasks. Robotics, driver automation, speech recognition are examples where deep learning is used to adapt software to cognitive tasks. We at Skilja are using deep learning algorithms for the creation of knowledge and intelligence on the task reading and understanding documents and text.
For a long time this has only been a research topic in computer sciences. But in recent times practical applications have become available that provide a basic functionality in understanding documents. Typically these applications are used to manage large amounts of incoming documents (especially paper) and to offer some kind of automatic recognition and distribution of documents. These early and simple solutions have proven successful and will now lead to a new generation of more powerful applications.
Skilja will closely follow this evolution through this web site and blog, but we will also actively work on new solutions and technologies through our company Skilja GmbH.
My name is Alexander Goerke and I am the author of these pages. I have pioneered this area since twenty years and created several products successful in the market but in the end only covered simple problems as mentioned above. If you want to learn more about my professional history please read here.
The Skilja web site is organized as a blog with posts and static pages.
I use the following categories to label the blog articles:
- Essentials: Basic articles on the fundamentals of document understanding. To get an overview over the field you should filter for this term to see only the more general posts.
- Technology: Focuses on technological achievements and methods as well as on algorithms used for document recognition. But not only basic technology topics are subsumed here. Also the application of these technologies in projects and software products falls into this section.
- Cognition: Category for thoughts about cognitive sciences and how they are related to concrete applications in document understanding.
- Market: Any news and observations on the market of document understanding.
- News: General category for short posts on news from the market, technological advances as well as social and political events that are relevant for us.
This section contains links to fundamental posts that deal with basics of document understanding and how it is related to cognitive sciences. Makes it easy for you to read and go back to them.
Browse the archives
As the archive is filling up you can browse it here and also search by category, tags and by full text search. We might even add some “document understanding” in the sense of smart context search in the future, but we need to develop that first