<img style="float: right" src="./pablo.png" height="150" width="150"> [Group Page](http://habla.dc.uba.ar/english.html) | [Lab Page](http://liaa.dc.uba.ar) # Pablo Brusco * PHD Student (final year) Advisor: [Agustín Gravano](http://habla.dc.uba.ar/gravano). * Computer Science Department, Exact and Natural Sciences School, University of Buenos Aires ## Contact Information Departamento de Computación, FCEyN, Universidad de Buenos Aires. Pabellón I, Ciudad Universitaria. (C1428EGA) Buenos Aires, ARGENTINA Email: first letter of my name followed by my last name, at dc.uba.ar (pbr…@dc.uba.ar) ## Education * Currently pursuing the fifth year of my Ph.D. in Computer Science at the <br/> University of Buenos Aires. * Computer Science Major <br/> Licenciatura in Computer Science at the University of Buenos Aires (June 2014). <br/> Roughly equivalent to a B.S. + M.S. in C.S. <br/> Degree Thesis: *Keyword spotting in languages without previous training data.* <br/> Advisors: [Luciana Ferrer](http://habla.dc.uba.ar/lferrer) and [Agustín Gravano](http://habla.dc.uba.ar/gravano). ## Interests * Machine Learning * Interpretability * Variable Importance * Spoken Language processing * Spoken Dialogue Systems * Dimensionality Reduction Techniques * Topic modeling * Turn-taking * Programming Languages * Programming Paradigms * Teaching ## Research My research topic primarily consists of understanding different natural language phenomena through the application of **machine learning as an inference tool**. In particular, I am interested in how different statistical methods behave when computing the importance of variables seeking for reproducibility and trustworthiness of results. I am currently working on understanding how people behave when exchanging turns in dialogue through the application of machine learning as a tool for analyzing data collected from different languages. Our ultimate goal consists of including this knowledge into spoken dialogue systems and therefore, improve their naturalness. ### Refereed Publications [Google Scholar](https://scholar.google.com.ar/citations?user=OucNeLgAAAAJ) * **Pablo Brusco**, Juan Manuel Pérez, Agustín Gravano. ``Cross-linguistic study of the production of turn-taking cues in American English and Argentine Spanish". Interspeech 2017, presented in Stockholm, Sweden in August 2017 * Agustín Gravano, **Pablo Brusco**, Stefan Benus, ``Who do you think will speak next? Perception of turn-taking cues in Slovak and Argentine Spanish'', Interspeech 2016, San Francisco, CA. September 2016 * Edgar Altszyler, **Pablo Brusco**, ``Análisis de la dinámica del contenido semántico", Jornadas Argentinas de Informática (JAIIO), 16th Argentine Symposium on Artificial Intelligence (ASAI 2015), Rosario, Argentina. September 2015. * **Pablo Brusco**, Luciana Ferrer, Agustín Gravano, ``Detección de palabras claves en lenguajes sin datos de entrenamiento'', Argentine Congress on Computer Sciences (CACIC 2014). Universidad Nacional de la Matanza, Buenos Aires, Argentina. July 2014. ### Research Group * Member of the [Applied Artificial Intelligence Lab](http://liaa.dc.uba.ar/) *June 2014 - Present* * Collaborator in **``Neurophysiological correlates of interaction between humans''**. We are currently investigating the neurophysiological correlates of coordination between humans during dialogue. Advisors: Agustín Gravano and Juan Kamienkowski. * Collaborator in **``Cross-linguistic study of prosodic cues in English, Spanish and Slovak''**. Advisors: Agustín Gravano and Stefan Benus. ### PhD Scholarship The National Scientific and Technical Research Council (CONICET) PhD Scholarship. April 2015 - March 2020 Topic: ``Pattern recognition over temporal series''. Instituto de Investigación en Ciencias de la Computación, CONICET-UBA, Buenos Aires. Advisor: Agustín Gravano ### Talks and intensive courses - Introduction to Machine Learning. (July - August 2018 at 10Pines). - [Cross-linguistic study of prosodic cues in English, Spanish and Slovak](interspeech_2017.pdf) (Interspeech 2017) - [Machine Learning, the end of classical programming? (Spanish)](charla_ml_pbrusco.pdf) Universidad General Sarmiento (Nov 2018) - INNOVA UNGS 2017, Machine Learning as an inference tool. Universidad General Sarmiento. Buenos Aires (May 2017). - Machine Learning as an inference tool. Universidad Tecnológica Nacional (2016) - [Variable Importance under correlation in Random Forest](charla_variable_importance.pdf) (LIAA) ### SRI International Internships - STAR Lab - United States *Dec 2018 -- March 2019* (4 months) - STAR Lab - United States *Sept 2016 -- Dec 2016* (3 months) Internship at the Speech Technology & Research Laboratory (STAR Lab) of SRI International as a Student Associate. We developed state-of-the-art tools applied to Acoustics Events recognition using Machine Learning techniques. ## Academic Experience #### At the [Department of Computer Science](http://dc.uba.ar/), School of Exact and Natural Sciences, University of Buenos Aires. [See student surveys](http://encuestas_finales.exactas.uba.ar/doc/d8127.html) ##### 8 semesters as **Head of TAs** - *2018-2*, Machine Learning. - *2019-2C*, *2019-1C*, *2018-1*, *2017-2*L: Algorithms and Data Structures I. - *2017-1*: Algebra I. - *2016-1, 2014-2*: Programming Languages Paradigms. ##### 6 semesters as **First-Class Teaching Assistant** - *2015-2, 2015-summer, 2014-2, 2013-2*: Algebra I. - *2015-1, 2014-1*: Programming Languages Paradigms. ##### 7 semesters as **Teaching Assistant** - *2013-2, 2013-1, 2012-2, 2012-1*: Programming Languages Paradigms. - *2011-2*: Numerical Methods - *2011-2, 2011-1*: Algorithms and Data Structures I ### Student Supervision #### Degree Thesis Advisor * Sebastián Sicardi. *``Turn-exchanges classification using Convolutional Neural Networks''*. 2019. In progress. **Degree thesis advisor** (co-advisor: Pablo Riera) * Martín Caravario. *``Machine Learning based on Small Data in EEG signal processing''*. August 2018. **Degree thesis advisor** (co-advisor: Laura Kaczer) * Gabriel Defagot. *``Topic modeling for acoustic scene recognition: combining T-SNE \& Latent Dirichlet Allocation''*. March 2018. **Degree thesis advisor** (co-advisor: Pablo Riera) * Alejandro Daniel Masseroli. *``Information extraction through subtitles classification''*. December 2016. **Degree thesis advisor** (co-advisor: Edgar Altszyler)