ZITA MARINHO

About

I am currently a Data Scientist Researcher at Sacoor Brothers, and I did my Ph.D at the Robotics Institute at CMU as part of the dual degree Carnegie Mellon Portugal PhD program with the Institute for Systems and Robotics at Instituto Superior Técnico, University of Lisbon. My research interestes lie in the application of Machine Learning algorithms for sequence prediction and planning problems with focus on Natural Language Processing and Robotics applications. My PhD advisors were Geoffrey J. Gordon (MLD), Siddhartha Srinivasa (RI) and André F. T. Martins (Unbabel). Prior to graduate school i did my Master studies in Physics Engineering at IST in Portugal.


REASEARCH INTERESTS

My research interests lie in general machine learning problems. My prior research work focused on learning methods using semi-supervision. Typically working with large unlabeled datasets and small supervised information. I was also interested in developing algorithms for planning with reinforcement learning.

PUBLICATIONS

Conference Papers

Recurrent Predictive State Policy (RPSP) Networks - Zita Marinho, Ahmed Hefny, Wen Sun, Siddhartha Srinivasa, Geoffrey J. Gordon, In Proceedings of the 35th International Conference on Machine Learning, Stockholm, Sweden, July 2018 [pdf] [bib] [slides] [code]

Predictive State Models for Prediction and Control in Partially Observable Environments - Ahmed Hefny, Carlton Downey, Zita Marinho, Wen Sun, Siddhartha Srinivasa, Geoffrey J. Gordon, In Conference on Robot Learning (CoRL), Mountain View California, November 2017 [pdf] [bib]

Unsupervised Learning for Nonlinear PieceWise Smooth Hybrid Systems - Gilwoo Lee, Zita Marinho, Aaron Johnson, Geoffrey J. Gordon, Siddhartha Srinivasa, Matthew Mason, In , ArXiv preprint, 2016 [pdf] [bib]

Semi-Supervised Learning of Sequence Models with the Method of Moments - Zita Marinho, Shay B. Cohen, André F. T. Martins, Noah A. Smith, In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, Austin, Texas, USA, November 2016 [pdf] [bib] [slides]

Functional Gradient Motion Planning in Reproducing Kernel Hilbert Spaces - Zita Marinho, Anca Dragan, Arunkumar Byravan, Byron Boots, Geoffrey J. Gordon, Siddhartha Srinivasa, In Proceedings of Robotics: Science and Systems, (RSS-2016), Ann Arbor, Michigan, USA, July 2016 [pdf] [bib] [slides] [poster]

Journal Papers

Three dimensional magnetic flux concentrators with improved efficiency for magnetoresistive sensors - Zita Marinho, Susana Cardoso, Rui Chaves, Ricardo Ferreira, Luis V. Melo, Paulo P. Freitas, In Journal of Applied Physics, 109, 2011 [pdf] [bib]

Dissertations

Kernel and Method of Moments for Prediction and Planning. Applications to Robotics and Natural Language Processing. - Zita Marinho, In PhD thesis dissertation, Carnegie Mellon University, Pittsburgh, USA and Instituto Superior T'{e}cnico, Universidade de Lisboa, February 2018 [pdf] [bib]

3D Magnetic Flux Concentrators With Improved Efficiency for Magnetoresistive Sensors - Zita Marinho, In Master Thesis dissertation, Instituto Superior Técnico, Lisbon, Portugal, November 2010 [pdf] [bib]

Contact Me


zmarinho [at] cmu [dot] edu

Carnegie Mellon University
5000 Forbes Av.
Pittsburgh, PA 15213,USA

Institute for Systems and Robotics
Instituto Superior Técnico
Av. Rovisco Pais, Lisbon, Portugal