Deepta Rajan

Title: Deepta Rajan: Innovator in Time Series Analysis

Introduction

Deepta Rajan is an accomplished inventor based in Bellevue, WA (US). He has made significant contributions to the field of time series analysis, holding 2 patents that showcase his innovative approach to data analysis.

Latest Patents

One of his latest patents is titled "Systems and methods for time series analysis using attention models." This patent discloses a system designed to capture dependencies across different variables through input embedding. The system maps the order of a sample appearance to a randomized lookup table via positional encoding. It captures dependencies within a single sequence through a self-attention mechanism, determining a range of dependency to consider for each position being analyzed. The system obtains an attention weighting to other positions in the sequence through the computation of an inner product. It utilizes this attention weighting to acquire a vector representation for a position and masks the sequence to enable causality. Additionally, the system employs a dense interpolation technique for encoding partial temporal ordering to obtain a single vector representation and a linear layer to derive logits from this representation. A type-dependent final prediction layer is also utilized in the system.

Career Highlights

Deepta Rajan has worked with notable institutions such as Arizona State University and Lawrence Livermore National Security, LLC. His experience in these organizations has contributed to his expertise in the field of time series analysis.

Collaborations

Some of his coworkers include Andreas Spanias and Huan Song, who have collaborated with him on various projects.

Conclusion

Deepta Rajan's innovative work in time series analysis through attention models highlights his significant contributions to the field. His patents reflect a deep understanding of complex data relationships and showcase his role as a leading inventor in this area.