Gerard (Rod) Rinkus, President and Chief Scientist

    Research Statement

Visiting Scientist, Brandeis Biology, Lisman Lab 2006-present.       Brandeis Website

Curriculum Vitae

 Rod's ResearchGate


Google Citations

 ORCID: 0000-0003-1725-910X     Loop ID: 7601

Rod Rinkus is seeking to understand the essential algorithm/circuit underlying biological intelligence. After receiving his Ph.D. in Cognitive and Neural Systems from Boston University in 1996, he developed AI and neuromorphic systems at several Boston-area research firms. In 2010, he founded Neurithmic Systems, which has been supported by ONR, DARPA, and commercial contracts and continues to develop sparse distributed representation (SDR)-based neuromorphic models for application to spatiotemporal problems, e.g., learning to recognize events in video. Computationally, Rod's most important contribution thus far has been the invention of a generic, hierarchical model, Sparsey® (originally called Temecor), which has fixed time for both storing (learning) a sequence and for retrieving the best-matching sequence, a capability that has not been demonstrated for any other algorithm, including any hashing method (e.g., LSH, semantic hashing). In addition to developing practical applications of Sparsey, e.g., for indexing, search, mining, of massive multivariate time series databases (of any modality), as well as compression, Rod is deeply interested in mapping Sparsey's algorithmic and structural details ever more closely to the brain, in particular, to neocortical macrocolumns and hippocampus.

Rod's primary professional goals are to:

  • Understand the information processing principles used in biological brains
  • Develop algorithms and software embodying these principles
  • Apply them to real-world problems such as event recognition in video, and discovering structure in all manner of multivariate time series data, as well as text/linguistic data
  • Use the developed models to explain increasing gamut of neural and cognitive phenomena and provide guidance for ongoing neuroscientific research (e.g., informing experiments using calcium imaging or other methods designed to see whole circuits, e.g., whole cell assemblies and phase sequences of cell assemblies in operation).