Error message

  • Warning: Parameter 1 to contemplate_node_view() expected to be a reference, value given in module_invoke_all() (line 817 of /home/hcomp/
  • Warning: Creating default object from empty value in Sweaver->get_current_style() (line 175 of /home/hcomp/


The Human and the Machine

Information drives innovation, however it comes in diverse forms, can be used in diverse ways, and requires equally diverse methods of analysis. Tomnod focuses on three interconnected areas of technology development: signal processing; machine learning; and massive scale human parallel computing (crowdsourcing) to enable massive scale data analysis. Our approach combines the scalability of machine systems with the adaptability and accuracy of human perception to process a broad range of multimedia data including images, video, audio, and text. Some key Tomnod technologies:
  • SatEnhance™ Technology: A pansharpening approach to increasing the spectral resolution of satellite images. The method not only produces state-of-the-art results but is also designed to be computationally feasible for large data sets.
  • CrowdRank™ Technology A proprietary algorithm for ranking the reliability and accuracy of individual users. The method allows optimal crowd training as well as higher quality crowd-based analysis.
  • Augmented Machine Learning™ Technology A hybrid system that combines the cognitive power of the human mind with the speed and scale of computer machine learning algorithms.
  • Data Avalanche Management (DAM) System A system that leverages existing commercial cloud infrastructure to manage massive datasets and corresponding metadata with structured, searchable databases.
  • Tomnod ROTGRS (Rapid On-The-Ground Response System) A mobile platform solution that integrates unstructured data with processed analytics to provide real-time contextual information to researchers and personnel on the ground.
Key platform components include: Front End The interface with the crowd. A web-based platform to distribute tasks, engage and educate the crowd. This platform can integrate or reside within existing social networks, such as Facebook. Back End The wisdom of the crowd is stored and processed. Huge datasets are divided and distributed, data analytics are performed, and each individual's contribution to the overall solution is optimized. Top Level Overview Seeing the solution. Tomnod provides statistics, results and visualization derived from the crowdsourced knowledge in a top-down overview of the data analysis. Google Earth, ArcGIS, and custom information displays are available. On the Ground Crowdsourced knowledge impacts the real world. Tomnod offers hardware/software solutions to provide clients with actionable results. Machine learning models trained on crowdsourced input automate further data analysis. Location-aware tablets deliver real-time analysis to clients on the ground. For example, for its most recent National Geographic engagement, Tomnod created a custom iPad application that integrated the current position of explorers in the field with real-time, crowdsourced analysis of geographic imagery generated by a crowd of thousands from around the world. Data Flow Tomnod manages the storage and flow of data between all components of the system, transforming massive data problems into actionable knowledge. Additionally, Tomnod has contracts in place for specialized data acquisition (e.g., aerial and satellite imagery) and is available for consulting and custom system design. Examples