Contact
355 Evans Hall, Berkeley, CA - 94704
Phone : 608-695-7951
email : antonyjoseph[at]lbl[dot]gov
About me
Since the fall of 2012, I have
been employed as a joint postdoctoral scholar with Dr.
Bin Yu, at the Department of
Statistics, UC Berkeley, and Dr. Erwin Frise, who is a biologist at the Department
of Genome Dynamics, Lawrence Berkeley National
Laboratory. Prior to this, I was a PhD student at the
Department of Statistics, Yale University. My thesis
adviser was Dr. Andrew Barron. I obtained my bachelor's
and master's degrees in statistics from the Indian
Statistical Institute, Kolkata.
My research interests are diverse, involving both theoretical and applied statistics. Over the course of my academic career, I have had the pleasure of working on a range of problems in both areas. I am especially interested in problems in high-dimensional statistical inference. Apart from this, I am interested in problems at the intersection of information theory and statistics. I am also involved in interdisciplinary research, applying statistical methods to systems biology.
Papers and Manuscripts
- A. Joseph, B. Yu. Impact of regularization on
spectral clustering. Preprint.
- A. Joseph. Variable selection in high-dimensions
with random designs and orthogonal matching pursuit. Journal
of Machine Learning Research, 2013.
- R. Venkataramanan, A. Joseph, S. Tatikonda.
Gaussian rate-distortion via sparse regression over
compact dictionaries. IEEE Trans. Inform. Theory,
to appear,
2013.
--Appeared in Proceedings of Int. Symp. Inform. Theory, 2012.
- A. Joseph, A. R. Barron. Sparse superposition
codes have near exponentially small error
probability for R < C. IEEE Trans. Inform. Theory, to appear, 2013.
--Parts appeared in Proc. Int. Symp. Inform. Theory, 2011 and 2012.
- D. Campbell, J. Bick, C. M. Yrigollen, M. Lee, A.
Joseph, J. T. Chang, E. L. Grigorenko.
Schooling and variation in the comt gene: the devil is in the details. Journal of Child Psychology and Psychiatry, 2013.
- A. Joseph, A. R. Barron. Least squares
superposition codes of moderate dictionary size
are reliable at rates up to capacity. IEEE Trans. Inform. Theory, 2012. --Appeared in Proc. Int. Symp. Inform. Theory, 2011.