Cognitum | Pawel Kaplanski
View Pawel Kaplanski's profile on LinkedIn View Pawel Kaplanski's profile on RSA View Pawel Kaplanski's profile on Google Scholar  

Paweł Kapłański, PhD, CEO at Cognitum

Schedule time with me

In Cognitum I am responsible for investor relations, recruitment, key customer relationship management and collaboration with academia. I am a Fellow of the Royal Society of Arts View Pawel Kaplanski's profile on RSA , Assistant Professor at Gdańsk University of Technology and Former Lead Architect in Samsung R&D.

Research track record

Starting with my PhD research, I devoted my research activities to the field of AI, semantic technologies, ontology engineering and big-data processing. Dealing with difficulties to obtain financial support from the government I founded an AI company and convinced investors to invest private money. It allowed me to build a team of highly skilled programmers and work on our Cognitive Architecture called Cognitum Platform. By now, after several years on the market, among other applications, the biggest implementation of the Cognitum Platform was a Fraud Detection system for the Brazilian government. A stream-based reasoner, developed by Cognitum, allows discovering fraudulent activities in the stream of 5M invoices per day.

Fluent Editor (one of the major components of Cognitum Platform) is nowadays well recognised Ontology Editor [www.w3.org] that enables cognitive programming in Controlled Natural Language supported with R language. Its free version is being used by universities (around 370) from all over the world. Several publications were written about Fluent Editor by us as well as Fluent Editor is cited (according to google scholar) by around 50 independent publications. Fluent Editor was presented at several scientific conferences including OWLED conference.

Ask Data Anything is another component of Cognitive Architecture my team in Cognitum and I have developed. It is an adaptive Data-mining platform and query engine that, using given ontologies, semantically integrates heterogeneous data sources and allows to ask them in Natural Language. Ask Data Anything implemented the idea of Natural Query Language and was presented at several big scientific conferences in CNL Workshop conference.

User interface inferred automatically from a specification written in Controlled Natural Language without requiring programming skills, is the component of Cognitum Platform that allows one to build the software without the need of programming activities. It was a result of collaboration with Medical University during a research devoted to the Clinical Decision Support System for a gist cancer. The results were published in Expert Systems With Applications that is currently ranked number 1 in the Google Scholar h-index listed under the top publications of Artificial Intelligence [www.sciencedirect.com].

The critical part of the system is the Reasoner that is composed of two parts: Left Hemisphere that performs reasoning in a logical manner based on theorem provers for logic expressed in Controlled Natural Language, and Right Hemisphere that performs reasoning in a black-box way, using machine learning, NLP and deep learning for pattern discovery. The connection between these two entirely different kinds of reasoning is realised with R language interface, which allows executing black box reasoning by theorem provers and rule engines.

My current research activities are related to hi-performance reasoning techniques (probabilistic), context management, and security of the reasoning process.

I received PhD in Computer Science from Gdansk University of Technology (GUT) where I am working now as an Assistant Professor. During my employment at the Gdansk University of Technology, I was teaching various topics, directly related to Computer Science. I annually conducted 60 hours of lectures and assisted during 300hours of laboratories. I promoted 4 MsC and 8 BsC theses. I was a reviewer of 20 MsC and 44 BsC theses.