PTPT in a Nutshell

How can I converge knowledge sources? Is there Artificial Intelligence to automatically make sense out of that? How can I apply that new knowledge?

The answer to these questions is PTPT. Three steps:

1) Connect your data sources
These maybe on-device resources (on your PC or phone), database connections, connections to services. They can be personal data sources (like email or your LinkedIn account), company resources (like Microsoft Sharepoint) or public resources (like patents on USPTO.gov). We have developed a couple of demo importers that you can readily use or you can write your own importers. You can define their update behaviour and to some degree map the sources’ structures to PTPT’s paradigm. PTPT will then automatically import the data and break it up into knowledge molecules and build associations across them. Since PTPT is source agnostic, knowledge about a topic from different sources will converge automatically.

2) Make the machine think
Next you can select modules or write code that runs periodically on PTPT’s knowledge graph and “thinks” about the data. PTPT’s knowledge graph is like a web of nodes (entities) and edges (associations), a Thinker can travel along this web like a spider and work on the graph. Thinkers look at entities, analyse their connections or contents and may decide to add new entities, clean up data, modify contents or add & remove associations or entities. Thus, incoming information is retained in the graph and is “alive”. PTPT is not just a linguistic processing pipeline nor is it a passive CMS, it actively can digest and produce new knowledge.

3) Use its output
Certainly, it is nice to build a graph that knows more than its single sources. But the real value arises from actually using that knowledge. Again, PTPT’s outputs are versatile:

  • use our interfaces to navigate / explore the data…
  • … or build your own user interface
  • write software that connects to the database and uses the insights in your own application
  • define output channels such as email alerts and news providers
  • develop a webservice using the data…
  • …or an Android app that works on-device as PTPT is small enough to run on a mobile phone.
  • use PTPT to pre-fill forms

One special feature is the Copilot: Let the user do stuff (like write an email or browse a website) and analyse what information he’s currently concerned with to show relevant contextual information.

You see, the usage of PTPT’s output ranges from connective over informative to assistive, just as your usage scenario requires it.

Have a look at Inspirations what could be build using PTPT, browse the Modules to see what capabilities PTPT is offering, review Develop to deep-dive into our SDKs. Or keep reading to familiarise yourself a bit more with PTPT:

Intro to PTPT

PTPT is a natural language knowledge management application and development framework. It is designed to support users in performing activities, accomplishing tasks and organizing the information and resources they use every day. PTPT uses the personal data people generate as they send and receive emails and texts, read websites, and interact over social media, finding structure and regularity in their everyday activities. It makes efficient use of the limited resources available on mobile devices to support the constant addition of new information. Features: Continue reading

PTPT’s Graph Database

The heart of PTPT is its innovative Graph Database, a flexible, typed, semi-structured database that supports complex and intelligent interaction while remaining small enough to run on mobile platforms. Graph databases have a number of advantages for applications, and a long history in artificial intelligence and natural language processing. They reflect more naturally the structure of human linguistic information than more traditional relational and record-oriented databases. Continue reading

Graph nodes: Entities

PTPT uses a system of four abstract, fundamental types of graph elements to characterize user data. At this basic level of analysis, an element of user data represents some record of user activity like an email, an SMS, a contact list entry, a social media posting, a calendar entry, an instance of using an app, or any other kind of user-generated record. Information extracted from these records is stored as an instance of one of four basic information types: Continue reading

Entity Relations

The PTPT Graph Database contains Entities that are connected by Relations. The four basic types entail some kinds of relations intuitively: An email account has a type Place and can intuitively been seen to contain a repository of emails called “Inbox”, also of type Place which contains a specific email of type Thing. An email is naturally associated with a send-time of type Time and a sender or recipient of type Person.

Unlike entities, relations are not always typed in PTPT, but typed entity relations help the database to better perform searches and to do intelligent reasoning about the entities it contains.

Time references support a special, more explicitly schematic system of relations. Each time reference is presumed to correspond to an event of some kind. Borrowing its terminology from semantic theme analysis in linguistics, PTPT dotes each entity of type Time with three relations relevant to the semantics of events: Continue reading

Named Entities in PTPT

The PTPT database goes beyond just storing meta-information along the four main dimensions described earlier. Using natural language processing technologies, PTPT extracts from the unstructured text itself additional references to times and people, as well as geographic names like cities, company names, and potentially many other categories. These too are stored in PTPT, with relations to the documents from which they are extracted and to each other if they appear together in the text. Continue reading