# Manling Li - Northwestern University - CS PhD/Intern ## Research Interests ### 1. Multimodal Foundation Models - Large Language Models (LLMs) - Vision-Language Foundation Models (VL models, especially for long videos) - Vision-Language-Action Foundation Models (VLA models, LLMs for Embodied AI) ### 2. Planning, Reasoning and Compositionality - From Large Language Model to Large Agent Model - Decision Making GPT, Agent Model Training - Going deeper: Spatial Reasoning, Geometric Structure Understanding, Law of Physics - Going longer: Long Horizon Reasoning, Long Video Encoding and Reasoning ### 3. Mechanistic Interpretability, Knowledge + LLMs - The next generation of knowledge representation (code representation, program induction, symbolic representation, etc) - Model editing, Knowledgeable foundation models (Where does certain knowledge come from? How much do these models know? Is this knowledge correct? If it is wrong, can we fix it?) - Controlling LLMs / foundation models (ACL 2024 Outstanding Paper) ### 4. Factuality, Faithfulness, Trustworthiness - NLP + Psychology (theory of mind, framing, information propagation, debiasing, etc) - Fact checking - AI for Science, Claim feasibility verification, Figure/Chart editing, Paper/Review generation, etc ## Position Details - 3-5 fully-funded Computer Science PhD positions (Fall 2025 or later), 1 postdoctoral position (Fall 2025 or later) - Multiple internship positions, can be in-person or remote (Fall 2024 or later, long-term availability, excellent candidates may be offered in-person RA positions) - Will provide **multiple ideas for interns to choose from**, also very supportive if interns want to pursue their own ideas - Excellent candidates may have opportunities to collaborate with professors from Stanford/UIUC/... - Previously mentored 15 master/undergraduate student, all of whom published papers in top conferences, most of which were their first papers - Can recommend outstanding students for exchange visits to top schools or internships and jobs at top company research institutes - Diversity: I always hope to build a **diverse group**. Students who have worked with me come from diverse backgrounds. I am most proud of a student who transferred from Civil Engineering to Computer Science. - I have received help and guidance from many professors. I had the best doctoral advisor in the world (Prof. Heng Ji, highly recommended!). I saw how to "do research with a relaxed attitude" and gradually **enjoyed the fun of doing research**. Therefore, I hope to do interesting research with students, respect their research interests, and most importantly, help them get to know themselves, find their true passion, do good research, and achieve their goals. - I hope to cultivate students' ability to **publish first-author papers**, rather than working on several projects simultaneously without taking significant responsibility in any. Leadership is an important skill. From my first to fourth year of PhD, I mostly published first-author papers. Currently, most of my highly cited papers on Google Scholar are either my first-author or co-first-author papers. So I hope students can be patient, take their research seriously, and publish quality papers one at a time. Quality always trumps quantity. - I **value collaboration** highly and will focus on developing students' presentation skills, learning how to communicate, pitch to collaborators, explain things as simply as possible, etc. - I conduct **PhD Research Skill Training Lectures** weekly, covering topics such as What is good research, Paper Writing, Figure/Table Tricks, Math/Latex, Paper Title, Coding Tricks, Appendix Writing, Paper Page Limit, Website/Video Making, Rebuttal, Paper Reviewing, Presentation, Best Paper Session, Ethical Issues, Collaborations, Project Management, Energy Management, etc. Students working with me can participate in these sessions. - Most importantly: highly motivated with their ownership on projects. I later learned that the most important thing in research is research passion, followed by effort and talent. With passion, one will gradually form their own research vision, rather than blindly making incremental improvements. We can first discuss your goals together, and then see how we can best help you grow. - Relevant research experience is a plus. ## Advisor Profile - Manling Li is an Assistant Professor at Northwestern University, and currently a postdoc at Stanford University, mainly working with Jiajun Wu and Fei-Fei Li. She obtained the PhD degree in Computer Science at University of Illinois Urbana-Champaign in 2023, where she was grateful to have Prof Heng Ji as her advisor and be mentored by Prof Shih-Fu Chang, Prof Kyunghyun Cho and Prof Jiawei Han. - She mainly publishes at ACL/CVPR/NeurIPS/ICLR/EMNLP/NAACL/... - Her work has been awarded **ACL 2024 Outstanding Paper**, **ACL 2020 Best Demo Paper Award**, **NAACL 2021 Best Demo Paper Award**, etc. - She has rich experience in working with students, leading 19 students to develop UIUC Knowledge Extraction System and ranked first in NIST SM-KBP in 2019 and 2020. - She delivered tutorials at ACL'21, AAAI'21, CVPR'23, AAAI'23, IJCAI24 and lead ACL 2024 Knowledgeable LLMs workshop. - She served as reviewers for top AI journals such as TACL, TKDE, TIST, and served as organizing committee at *CL conferences (Virtual Infrastructure Chairs of ACL 2025, Publication Chairs at NAACL 25, Demo Track Chairs at EMNLP 2024, etc.) - She was awarded Microsoft Research PhD Fellowship, EE CS Rising Star, etc. - Homepage: https://limanling.github.io/ ## School/Program Introduction - Northwestern University is a renowned top private school in the United States, consistently ranking in the top 10 in the US and top 30 worldwide - Located in Evanston, part of Chicago's northern suburbs, it offers a safe and nice living environment - The school is extremely beautiful, with two private beaches for faculty and students - For those who like the convenience of a big city but want less of the hustle and noise and danger, Northwestern is definitely one of the best choices. The campus is dotted with shuttle stations, it's a 30-minute driving or 1-hour Purple Line train to downtown Chicago, and about 30 minutes to Chicago airport. You can order takeouts from Downtown Chicago or Chinatown. ## Application Process - For internship applications, please fill out the form at https://forms.gle/unVqpW97bGwQsrRz5, then send an email to limanling.ai@gmail.com with the subject "Prospective Intern – Your Name", noting that you have filled out the form. - For PhD applications, please apply to the Northwestern CS PhD program at https://www.mccormick.northwestern.edu/computer-science/academics/graduate/admissions by December 1, 2024 [Please list me as one of the interested advisors (I apologize that due to time constraints, I can only review applications that choose me as a potential advisor)]. - **No need to send emails for PhD applications**. If you have time and would like to work together, please feel free to send an email following the internship application process. **I apologize that I cannot reply to individual PhD application emails**, but I will **carefully review every application** at the end of December. After January, please feel free to contact me about your application status and send emails to update your information or papers (Please note that a lack of email response does not indicate a lack of interest; this strategy of not replying to individual emails before all application materials are in place is to ensure a fair evaluation process for all applicants and respect their time).