Gissella M. Bejarano

Picture Gissella Bejarano

I am a PhD candidate at the Computer Science Department of Binghamton University (SUNY) working with Professor Arti Ramesh . I graduated as a Master in 2017 at Binghamton too. I earned my bachelor degree in Computer Science Engineering from Pontifical Catholic University of Peru. I have published in conferences such as AAAI and Buildsys.

News

Interests

My field of research is in Machine Learning for Smart City problems. More specifically, I have woked with probabilistic graphical and deep learning models for problems such as water consumption prediction and energy disaggregation. Most of my work focuses in structured prediction for sequential data. I have also explored Natural Language Processing models using variational recurrent neural networks and I am collecting a dataset to construct multimodal models for Sign Language Recognition.

Research

  • Emergency Response Time Prediction (2020): I explored different preprocessing approaches for unequidistant datapoints in the NYC Open Emergency Dataset. I constructed a deep learning based sequence-to-sequence model for structured prediction.

  • Water Consumption Prediction (2019): I designed a probabilistic graphicl model (SGCRF) and an seq2seq encoder-decoder for water consumption prediction.

  • Language Generation (2018): I explored Word-embedding and adaptation of VRNN for language generation for COCO dataset.

  • Energy Disaggregation (2017-2018): I developed a model based on VRNNs (Variational Recurrent Neural Networks) to address the Energy Disaggregation Problem.

  • Sign Language Recognition (2016-2017): Application of CNN and RNN for the recognition of the Australian Sign Language (UCI) as part of Project Termination for the Master degree.

  • Public Transit Routes optimization (2013-2016):This project was funded by FINCYT (PerĂº) in 2013. More information can be found in this link.

Publications

  • G. Bejarano, A. Kulkarni, X. Luo, A. Seetharam, and A. Ramesh. DeepER: A Deep Learning based Emergency Resolution Time Prediction System. The 2020 IEEE International Conference on Cyber, Physical and Social Computing, 2020.

  • G. Bejarano, A. Kulkarni, R. Raushan, A. Seetharam, and A. Ramesh. SWaP: Probabilistic Graphical and Deep Learning Models for Water Consumption Prediction. Proceedings of the 6th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, BuildSys 2019.

  • G. Bejarano , D. Defazio, and A. Ramesh. Deep Latent Generative Models for Energy Disaggregation. AAAI Conference on Artificial Intelligence, 2019

  • G. Bejarano , Deep Learning and Probabilistic Models Applied to Sequential Data. PhD Forum, SMARTCOMP 2018.

  • G. Bejarano , M. Jain , A. Ramesh , A. Seetharam , A. Mishra. A Predictive Analytics for Smart Water Management in Developing Regions. SMARTCOMP 2018

  • G. Bejarano, J. Astuvilca and P. Vega. Automation of process to load database from OSM for the design of public routes. 2nd Annual International Symposium on Information Management and Big Data, SIMBIG 2015

Teaching Assitant

  • CS575 Design & Analysis Computational Algorithms
  • CS580 Introduction to Machine Learning
  • CS535 Introduction to Data Mining

Presentations

  • October 2020 Why to do research in Artificial Intelligence? Keynote speaker at CINCIT 2020
  • October 2020 Why do we need an Artificial Intelligence Strategy at TECHSUYO 2020
  • March 2020 How do we bridge the social gap with Artificial Intelligence at WIDS Lima, 2020

Mentoring

  • Jose Esparza (REPUcs/Binghamton - 2021)
  • Anusha Kamath (Binghamton 2018)
  • Alexander Van Roijen (Binghamton 2017)
  • Walter Barrantes (PUCP 2013)
  • Julio Candela (PUCP 2013)

Contact

E-mail: gbejara1[at]binghamton.edu