Manling Li Experiences EECS Rising Stars 2022
11/14/2022 11:15:00 AM
From the first year she started the Illinois Computer Science PhD program, now four years ago, Manling Li knew about and wanted to be a part of the EECS Rising Stars program. As a vessel for career progression and outreach, Li felt there are few better opportunities than becoming a Rising Star.
At the end of October, she made that moment come true.
Li participated in the two-day workshop hosted this year by The University of Texas at Austin. Its focus is as an intensive workshop for graduate students and postdocs with historically marginalized or underrepresented genders who are interested in pursuing academic careers in electrical engineering, computer science, and artificial intelligence and decision-making.
“The most valuable takeaway for me was improving my job search skills and preparing myself better for the job search,” Li said. “One of the most valuable things I have learned is to think and act like a professor during the job search process. As we approach the end of our graduate studies, we should not view ourselves as students, but, rather, as leaders in our fields, prepared to guide cutting-edge research.”
Li has already finished her thesis proposal and is in the final stage of completing her thesis research.
From a doctoral experience spent studying with Illinois CS professor and advisor Heng Ji, while focused on Natural Language Processing (NLP) and Computer Vision (CV), Li was able to present on three different instances at Rising Stars.
First, she had a poster presentation on “Event-Centric Multimodal Data Understanding.” Then, Li presented a 12-minute talk for the NLP group on “CLIP-Event to connect vision and language via event structures.” Finally, Li also had an 8-minute talk in front of peers, which served as a rehearsal of her introduction to a job talk.
“It was a special experience for me to participate in the job talk practice. I have presented my work at multiple universities, but this presentation is in front of a group of peers who are in the job market at the same time and will share insights from their perspectives,” Li said. “Additionally, it was a great opportunity to meet with folks from the NLP group at UT Austin, which allowed me to exchange ideas with researchers and students there. We had a great discussion about what values to embrace in the era of large-scale pretraining.”
As for her research interest, Li said that her primary focus is to “transform traditional information access from Entity-Centric to Event-Centric as well as from Text-Only to Multi-Modal.”
While she has achieved success thus far, Li believes there is still space to further explore the understanding of the “deep semantics behind the scenes” – such as, what happened, to whom, when, where, why, and what happens next.
And it was valuable for her to open her findings and discourse up to the participants and viewers at Rising Stars 2022.
“It was a pleasure to meet with many brilliant peers from very different fields,” Li said. “I believe this is a good starting point for me to explore more interdisciplinary research. Being an NLP and CV researcher, I have greatly benefited from interdisciplinary research, as well as the comparison of different understandings of knowledge and structures.
“During this event, I talked more with people from Human-Computer Interaction, and Robotics. I am looking forward to integrating NLP with the physical world and evaluating how machines can work more effectively with humans, and ultimately improving the general performance of NLP that is collaborative between humans and machines.”
In an already decorated academic career, Li considered the Rising Stars inclusion as an important milestone.
Prior to this, Li received the ACL 2020 Best Demo Paper, the NAACL 2021 Best Demo Paper, and the Microsoft Research PhD Fellowship last year.
“I see these acknowledgments as a natural progression in my development as an independent researcher,” Li said. “With these recognitions, I am committed to leading our community to greater achievements on event-centric multimodal data understanding, helping humans to better comprehend the deep semantics behind the data and the scene, and to enhance the quality of multimedia content consumption.”