a group-wide, corporate memory ecosystem
About

CoMem is an ecosystem based on evolving knowledge graphs. It integrates information silos and legacy systems as well as (personal) information sources of knowledge workers and teams. Embedded in their office environment, CoMem supports daily work with knowledge services to form a bottom-up Corporate Memory.

As both a long-term, multi-project spanning research endeavor, as well as, an operating software prototype, CoMem transfers state-of-the-art research to the real world. Research focuses on high innovation, exploration and empirical experimentation, while at the same time, our continuously running prototype system has high demands on reliability and performance.

See showcase for CoMem in industrial practice at enviaM (a German energy provider)

Philosophy

The philosophy behind the CoMem approach has a decade-long tradition of research in our DFKI department Smart Data & Knowledge Services (SDS) of embedding knowledge management into the corporate environment, work processes and human activities by using a broad spectrum of AI technologies. Below are a few of the key aspects:

Organizational Memory

CoMem acquires, stores, structures and combines information from various organizational data sources and maintains an Organizational Memory (OM). The knowledge representation for the OM is based on semantic web standards and allows semantic querying and intuitive access.

Bottom-up Knowledge Mgt.

CoMem's bottom-up approach to Knowledge Management (KM) aims at instant benefits for individual users. This lowers barriers for KM activities and fosters group KM. User-private knowledge as well as user-specific annotation must be allowed to benefit individual users.

Office Work Integration

KM works best when additional work is avoided. CoMem is embedded into the office environment by our Semantic Desktop ecosystem. Application plug-ins allow to interact with the application's data and CoMem's knowledge graph. Innovative UIs (e.g. sidebar) parallel to office applications offer pro-active information delivery and direct OM access.

Continuous Integration

We use our own CoMem instance since 2011. New features from researchers are integrated and tested with with real data in real-life settings. The CoMem service runs 24/7 and offers ubiquitous access from any device.
We live rapid prototyping and love to test new ideas in new domains as research as well as in industry projects.
Research

Our current and future research and development is aiming at the following highly innovative features to build a next-generation KM assistance:

Context

Contexts are first order citizens in CoMem, allowing to build contexts from tasks, meetings, notes, etc. Knowledge-services use contexts to derive information need, relevancy, or to enrich information. Our most recent development here are cSpaces: self-organizing context-spaces.

Managed Forgetting

CoMem is extended with strategies (similar to human forgetting) to realize hiding of information not relevant for the current context. The goal is to avoid distractions and find the right information faster, thus, helping the user to better focus on the actual work.

Domain Adaptation

CoMem adapts to new domains by learning new domain ontologies and vocabularies from native structures such as folder hierarchies and by extending user interfaces. CoMem was applied in scenarios such as research & consulting, 3rd level support, property management, personal photo management, or meeting support.

Connecting to Legacy Data

Digitalization does not necessarily mean exchanging all legacy systems. Via Semantic Bridges, CoMem connects distributed information sources to leverage corporate data, aggregate, index it and make the whole enterprise knowledge graph accessible at one place and serving as glue between legacy systems.
Team

Research and development is done by the DFKI SDS research department. Throughout several research projects our approach to organizational memories and the semantic desktop has been done as indicated by the publications list. The department is lead by Prof. Dr. Andreas Dengel.

Prof. Dr. Andreas Dengel
Head of Department SDS
andreas.dengel@dfki.de

CoMem is built on decade-long research and development. As such, a great deal of researchers and engineers were involved in its lifetime. Currently, this is the core CoMem team:

Dr. Heiko Maus
Team Lead, Chief Engineer
heiko.maus@dfki.de
Dr. Sven Schwarz
Senior Researcher
sven.schwarz@dfki.de
Dr. Christian Jilek
Senior Researcher
christian.jilek@dfki.de
Dr. Markus Schröder
Senior Researcher
markus.schroeder@dfki.de
Michael Schulze
PhD Researcher
michael.schulze@dfki.de
Andreas Lauer
Senior Software Engineer
andreas.lauer@dfki.de
Rudolf Koch
Software Engineer
rudolf.koch@dfki.de
Emil Baitemirov
Software Engineer
emil.baitemirov@dfki.de
Desiree Heim
PhD Researcher
desiree.heim@dfki.de
See also