Research Challenges

From ngfci
Revision as of 18:47, 21 July 2012 by Louiqa (talk | contribs)
Jump to navigation Jump to search

There will be 2 90 minute Research Challenge sessions on Thursday and Friday afternoon. We have identified 5 themes for multi-disciplinary / computational research. Participants must indicate their interest in participation in at most 2 themes - primary and secondary. Depending on the response we may choose to conduct all 5 themes in parallel over both sessions or design an alternative scenario. Participants are also free to propose additional multi-disciplinary themes or propose merging or splitting themes.

  • Network Analysis and Visual Analytics and Machine Learning and Prediction
Coordinators: Bill Ribarsky and  Akhtarur Siddique and Amitabh Varshney
- Network analysis and clustering and prediction
- Latent variables and hidden networks and hypergraphs
- Network evolution
- Information visualization pertaining to systemic risk
  • Contractual Reasoning and Semantics and Taxonomies and Metadata
Coordinators: Benjamin Grosof and Leora Morgenstern and Andreas Cali and Frank Olken
- Parsimonious/machine representation of a financial contract
- Contract evolution
- Data models and schema.
- Metadata
- Taxonomies and ontologies and poly-hierarchies
- Validation and reasoning
  • Information Integration and Entity Resolution and Information Quality
Coordinators: Lucian Popa and Joe Langsam and Rachel Pottinger
- Human language technologies and document collections
- Information extraction
- LEI and post LEI challenges
- Entity resolution
  • Social Media and Crowdsourcing and Markets
Coordinators: Johannes Gehrke and Louiqa Raschid and Michael Wellman
- Social media modeling and prediction
- Crowdsourcing
- Market mechanisms
- Prediction markets
- Agent based models
- Media:Session4.txt
  • Model Representation and Model Management
Coordinators: Phil Bernstein and H.V. Jagadish and Amol Deshpande and Pete Kyle
- Data models and schema and metadata
- Representing financial models as first-class data objects
- Reconciling different perspectives in representation: financial, accounting, legal, …
- Error correction, and propagation of corrections through derived data
- Privacy

We have created a shared google doc. You need to email Michelle (mlui@rhsmith.umd.edu) an account that will allow access to docs.google.com . [1]