Paul Liang (@pliang279) / X

Paul Liang

3,753 posts

Paul Liang

@pliang279

Assistant Professor MIT @medialab @MITEECS @nlp_mit || Foundations of self-evolving multisensory AI to enhance the human experience.

Joined July 2012

Pinned
Paul Liang
@pliang279
Aug 27, 2025
A bit late, but finally got around to posting the recorded and edited lecture videos for the **How to AI (Almost) Anything** course I taught at MIT in spring 2025. Youtube playlist: youtube.com/watch?v=0MYt0u… Course website and materials: mit-mi.github.io/how2ai-course/… Today's AI can be
106K
Paul Liang
@pliang279
Feb 17, 2025
This spring I am teaching a new class at MIT called **How to AI (Almost) Anything** Its name is a play on 2 seminal @medialab courses: how to make almost anything (on design & fabrication) and how to grow almost anything (on synthetic biology) We are now in the AI age, and
217K
Paul Liang
@pliang279
Jul 31, 2024
📣 I'm thrilled to share that I’ll be joining MIT as an assistant professor this fall, joint between @medialab & @MITEECS. My group will advance the foundations of multisensory AI to enhance the human experience. I look forward to tackling exciting challenges in multimodal AI
92K
Paul Liang
@pliang279
Oct 23, 2024
📣 Announcing the name and theme of my new research group at MIT @medialab @MITEECS: ***Multisensory Intelligence*** media.mit.edu/groups/multise… Our group studies the foundations of multisensory AI to create human-AI symbiosis across scales and sensory mediums. We are hiring at
44K
Paul Liang
@pliang279
Nov 14, 2023
Despite the successes of contrastive learning (eg CLIP), they have a fundamental limitation - it can only capture *shared* info between modalities, and ignores *unique* info To fix it, a thread for our #NeurIPS2023 paper w Zihao Martin @james_y_zou @lpmorency @rsalakhu:
105K
Paul Liang
@pliang279
Jun 16, 2022
If recent models like DALL.E, Imagen, CLIP, and Flamingo have you excited, check out our upcoming #CVPR2022 tutorial on Multimodal Machine Learning - next monday 6/20 9am-1230pm cmu-multicomp-lab.github.io/mmml-tutorial/… slides, videos & a new survey paper will be posted soon after the tutorial!
Paul Liang
@pliang279
Feb 28, 2023
As PhD visit days are coming up, I'd like to share this collated resource for prospective & current PhDs, covering how to choose advisors & schools, advice for research, teaching, fellowships, networking & more github.com/pliang279/awes… Credit to the original authors of each link!
32K
Paul Liang
@pliang279
Jul 7, 2023
I'm on the faculty market in fall 2023! I work on foundations of multimodal ML applied to NLP, socially intelligent AI & health. My research & teaching cs.cmu.edu/~pliang/ If I could be a good fit for your department, I'd love to chat at #ACL2023NLP & #ICML2023 DM/email me!
00:00
49K
Paul Liang
@pliang279
Aug 19, 2022
Check out slides & video recordings of our recent tutorials on multimodal machine learning at CVPR 2022 and NAACL 2022: video: youtube.com/playlist?list=… slides and readings: cmu-multicomp-lab.github.io/mmml-tutorial/…
Paul Liang
@pliang279
Apr 11, 2022
Really excited to release the video of my guest lecture on Multimodal Deep Learning for CMU's Deep Learning class @rsalakhu @mldcmu It covers 5 fundamental concepts in multimodal ML: representation, alignment, reasoning, translation & co-learning Youtube:
Paul Liang
@pliang279
Nov 3, 2023
Multimodal AI studies the info in each modality & how it relates or combines with other modalities. This past year, we've been working towards a **foundation** for multimodal AI: I'm excited to share our progress at #NeurIPS2023 & #ICMI2023: arxiv.org/abs/2302.12247 see long 🧵:
81K
Paul Liang
@pliang279
Mar 3, 2022
[11877 Advanced Topics in Multimodal ML] In week 5’s session, the class aimed to define a taxonomy of multimodal reasoning: the (hierarchical) composition of unimodal and multimodal evidences into higher-level abstract concepts for prediction. Notes here: cmu-multicomp-lab.github.io/adv-mmml-cours…
Paul Liang
@pliang279
Feb 25, 2025
Conventional wisdom has assumed that aligning representations (eg. contrastive learning) from different views or modalities is critical for better performance. In our work, we show a fundamental limit to how much we should align representations and scenarios where more alignment
Megan Tjandrasuwita
@mmtjandrasuwita
Feb 25, 2025
Interested in how and when neural networks trained on different modalities learn similar representations without explicit training objectives? Check out our work: "Understanding the Emergence of Multimodal Representation Alignment."
29K
Paul Liang
@pliang279
Feb 10, 2023
This semester, @lpmorency and I are teaching 2 new graduate seminars @LTIatCMU @mldcmu. The first, 11-877 Advanced Topics in Multimodal Machine Learning, focuses on open research questions and recent theoretical & empirical advances in multimodal ML: cmu-multicomp-lab.github.io/adv-mmml-cours…
25K