Hi, I'm

Oana Ignat

Passionate about research & applying science to help create a better world.

About me

CS Postdoctoral researcher at UMich, LIT Lab, advised by Rada Mihalcea.
I am on the job market for tenure-track Professor positions.
My research is at the intersection of Natural Language Processing & Computer Vision.
Current focus on AI for social good, building models for diverse location and socio-economic backgrounds.
For my thesis, I worked with social media videos such as YouTube vlogs, to develop Datasets & Models for Human Action Understanding.

  • Current Work: Postdoctoral researcher, University of Michigan
  • Education: PhD Computer Science, University of Michigan
  • Graduation Date: August, 2022
  • Location: Ann Arbor, MI, USA
  • Contact: oignat@umich.edu
Download Resume

Work Experience

  • Academia

  • Postdoctoral researcher - University of Michigan

    2023 - Present

    Mentor for Master and Ph.D. students, Co-PI Global Health Equity Grant, co-organizer of ACL Year Round Mentorship and many other events and workshops centered on improving diversity in CS.

  • Graduate Research Assistant - University of Michigan

    2017 - 2022

    Volunteer for AI4ALL, I introduced high school students to AI by teaching them how to implement and present an NLP project related to social good.

  • Primary Instructor - University of Michigan

    Fall 2021

    Primary Instructor for Dicover CS (EECS 198), a class primarily designed for underrepresented minorities in CS

  • Engineering Teaching Consultant - University of Michigan

    2020 - 2021

    As a Teaching Consultant, I helped other GSIs/IAs to improve their teaching through one on-one consultations where we discuss challenges they face during teaching, collected feedback from students of the GSIs/IAs and discussed it with the student instructor.

  • Graduate Student Instructor - University of Michigan

    2019 - 2020

    EECS 281: Data Structures and Algorithms and EECS 476: Data Mining

  • Industry

  • Research Intern - Amazon Prime Video

    September 2022 - December 2022

    Mentored by Jingru Yi and Burak Uzkent, I implemented and presented research on image-phrase grounding using contrastive learning and image augmentation.

  • Research Intern - Facebook FAIR

    May 2021 - August 2021

    Mentored by Paco Guzman, Vishrav Chaudhary and Jean Maillard, I implemented and presented research on OCR impact in Machine Translation for low resource languages.

  • Research Intern - Facebook FAIAR

    May 2020 - August 2020

    Mentored by Alon Halevy, Jane Yu and Y-Lan Boureau, I implemented and presented research on Identifying and Analyzing Inspiration in social media data.

  • Image Processing Developer - Telenav

    August 2016 - August 2017

    Research and implement efficient methods for road detection using raw images from mobile phones.

  • Image Processing Intern - Robert Bosch

    July 2015 - August 2016

    My undergraduate thesis was done as part of the internship and was about implementing a real-time road recognition algorithm having as input a disparity (depth) image taken from a stereo camera.

Publications (Paper, Poster & Code)

Scalable Performance Analysis for Vision-Language Models

  Santiago Castro*, Oana Ignat*, Rada Mihalcea
  12th Joint Conference on Lexical and Computational Semantics (*SEM 2023)

Fill-in-the-Blank: A Challenging Video Understanding Evaluation Framework

  Santiago Castro*, Ruoyao Wang, Pingxuan Huang, Ian Stewart, Oana Ignat, Nan Liu, Jonathan C. Stroud, Rada Mihalcea
  Annual Meeting of the Association for Computational Linguistics (ACL 2022)

OCR Improves Machine Translation for Low-Resource Languages

 Oana Ignat*, Jean Maillard, Vishrav Chaudhary, Francisco Guzm├ín
  Annual Meeting of the Association for Computational Linguistics (Findings ACL 2022)

When did it happen? Duration-informed Temporal Localization of Narrated Actions in Vlogs

 Oana Ignat*, Santiago Castro, Yuhang Zhou, Jiajun Bao, Dandan Shan, Rada Mihalcea
  ACM Transactions on Multimedia Computing, Communications, Applications (TOMM 2022)

WhyAct: Identifying Action Reasons in Lifestyle Vlogs

 Oana Ignat*, Santiago Castro, Hanwen Miao, Weiji Li, Rada Mihalcea
  The 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP 2021)

Detecting Inspiring Content on Social Media

 Oana Ignat*, Y-Lan Boureau, Jane A. Yu, Alon Halevy
  9th International Conference on Affective Computing & Intelligent Interaction (ACII 2021)

Identifying Visible actions in Lifestyle Vlogs

 Oana Ignat*, Laura Burdick, Jia Deng, Rada Mihalcea
  Annual Meeting of the Association for Computational Linguistics (ACL 2019)

Disparity image segmentation for free-space detection

 Oana Ignat*, Cosmin Lazar
  IEEE International Conference on Intelligent Computer Communication and Processing (ICCP 2016)

Ph.D. Thesis

Here you find my thesis, Towards Human Action Understanding in Social Media Videos using Multimodal Models, together with all resources (Datasets & ML models).

My Hobbies