Instrumentation for Multimodal Data Analysis for Representation Learning and Visual Recognition

Abstract

We request funding to set up a high-end GPU computing system to support research in the area of multimodal data analysis for representation learning and visual recognition. The proposed system consists of a small number of high-end GPU servers and GPU workstations that will be useful for the design, development and evaluation of novel algorithms for multimodal data analysis, representation learning and visual recognition. In particular, the requested equipment will allow us to develop novel approaches for characterizing information content in multimodal data such as images, videos, accelerometers and depth sensors, and support multiple data analysis tasks such as dimensionality reduction, representation learning, object recognition and action recognition. Outreach efforts to Army laboratories and other universities that are partners in the upcoming ARO MURI ÒSemantic Information Pursuit for Multimodal Data AnalysisÓ and the upcoming IARPA DIVA ÒMulti-camera Based Detection of Objects, Persons and ActivitiesÓ are envisioned through cooperative efforts. The availability of the proposed equipment will significantly enhance the research of graduate students, postdocs and research scientists to be supported by these upcoming projects.

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 14, 2019
Source ID
W911NF1810220

Entities

People

  • Rene Vidal

Organizations

  • Army Contracting Command
  • Johns Hopkins University
  • United States Army

Tags

Readers

  • Neural Network Machine Learning.
  • Research Science/Academic Research