Skip to Content.
Sympa Menu

nl-uiuc - [[nl-uiuc] ] Computational Linguistics / Speech Processing talk: Mary Pietrowicz (IBM Watson)

nl-uiuc AT lists.cs.illinois.edu

Subject: Natural language research announcements

List archive

[[nl-uiuc] ] Computational Linguistics / Speech Processing talk: Mary Pietrowicz (IBM Watson)


Chronological Thread 
  • From: "Girju, Corina R" <girju AT illinois.edu>
  • To: "nl-uiuc AT cs.uiuc.edu" <nl-uiuc AT cs.uiuc.edu>
  • Subject: [[nl-uiuc] ] Computational Linguistics / Speech Processing talk: Mary Pietrowicz (IBM Watson)
  • Date: Mon, 28 Jan 2019 18:41:44 +0000
  • Accept-language: en-US
  • Authentication-results: illinois.edu; spf=pass smtp.mailfrom=girju AT illinois.edu; dkim=pass header.d=uillinoisedu.onmicrosoft.com header.s=selector1-illinois-edu; dmarc=pass header.from=illinois.edu

Hi everyone,


Mary Pietrowicz (IBM Watson), will be giving a job talk for the assistant professor position in Computational Linguistics / Speech Processing at 4pm in Lucy Ellis (1080 FLB) today, Monday Jan. 28, 2019.


Her bio, the title and abstract of the talk are below.


Please distribute and attend!


Best,

Roxana

 

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Title: Enabling _expression_‐Driven and _expression_‐Aware Systems

 

Abstract: Human _expression_ is at the heart of communication and creativity. It speaks through our spoken and written language, paralingual _expression_, physical gesture, and creative output. Typical investigations into models of acoustic human _expression_ often focus on low‐level signal processing features and fail to consider human perception. The resulting machine models are frequently out of alignment with what humans perceive and do, and therefore fail to support generalized application development.

 

In this talk, I will introduce and demonstrate a process for developing machine models of human _expression_ which are aligned with perception. I will also show how this process can reveal different dimensions of expressivity organically without requiring use of either predefined vocabulary or mapping onto predefined axes, which is commonly done when mapping emotion onto a 2‐dimensional affect and arousal space. This process enables the discovery of relationships among sub‐dimensional expressive elements which can then be used to improve the performance of machine models.

 

Next, I will present some samples of _expression_‐aware projects which either mirror human _expression_, extend the range of natural human expressivity, or model relationships between spoken _expression_ and wellness. I will conclude by discussing the promise of leveraging models of human _expression_ in the development of _expression_‐driven applications for health, wellness, and creativity.

 

Bio: Mary Pietrowicz is a postdoctoral research scientist at the IBM Thomas J Watson Research Center. Her research focuses on human _expression_, and spans topics across speech and language, human‐computer interaction, crowdsourced creativity, machine learning for signal processing, visualization, music, and interactive art. Prior to studying for her PhD, Mary was an engineer and researcher in industry and academia for many years. Her website is at https://marypietrowicz.com



--
Roxana Girju
Associate Professor of Linguistics,
     Computer Science and Beckman Institute (affiliate/part-time)
Director of Computer Science + Linguistics Joint Major (Linguistics)
University of Illinois


  • [[nl-uiuc] ] Computational Linguistics / Speech Processing talk: Mary Pietrowicz (IBM Watson), Girju, Corina R, 01/28/2019

Archive powered by MHonArc 2.6.19.

Top of Page