Jung Yeon Park

ML research @ProfluentBio, PhD @Northeastern

[Jung Yeon Park @park_jungy]

(1/8) Happy to share that our paper LieFlow has been accepted to #ICML2026.

Symmetry is a strong inductive bias in ML, but you have to know the symmetry in advance. Can a model discover it directly from data, for both discrete and continuous groups? 🧵

[Dian Wang @Dian_Wang_]

Long horizon bimanual mobile manipulation requires reasoning in many coordinate frames: base, L/R hands, etc. In which frame would policy work best?
It really depends, so don’t pick. Mixture of Frames Policy: denoise in multiple frame in parallel.
🌐https://t.co/GnKSeNqzHA (1/9)

[Eddie Berman @bermanatorCodes]

I'll be @icmlconf presenting our work Smoothness Errors in Dynamics Models and How to Avoid Them, with Luisa Li, @park_jungy , and @RobinSFWalters !!

Come see our poster #2200 , july 9th, 2:30 PM – 4:15 PM, hall A.

Links to paper, website, and code in thread!

[Jung Yeon Park @park_jungy]

Come check out symmetry discovery with flow matching today 5pm at Hall A #3010!

[Jung Yeon Park @park_jungy]

(1/8) Happy to share that our paper LieFlow has been accepted to #ICML2026.

Symmetry is a strong inductive bias in ML, but you have to know the symmetry in advance. Can a model discover it directly from data, for both discrete and continuous groups? ��

[Jung Yeon Park @park_jungy]

A couple of us @ProfluentBio will be at ICML in Seoul next week! Reach out to me if you want to talk about protein language models, equivariance, and generative models. Coffee chats and dinner signups in the post below.

[Profluent @ProfluentBio]

The Profluent team will be at ICML in Seoul next week. Find us to chat, grab a coffee, or join us for dinner!

We’d love to talk shop (protein language models, protein optimization, sequence-first methods vs structure-first). Basically, anything at the intersection of AI and biology.

Want to grab coffee? Let us know here: https://t.co/dJWt7CjFeS

Or want to hang out and chat over dinner? Reserve a spot at our table: https://t.co/eZNKiXkFVr

[Jaehoon Yoo @wognsfjq96]

Self-conditioning can be distilled into one-step language generation.

We show that self-conditioned flow LMs perform fixed-point iteration, and introduce fixed-point flow maps to compress both the iterations and the entire flow.

At one step, FMLM★ achieves 112.5 gPPL at near-data entropy (5.37), substantially improving over FMLM’s 168.3 gPPL at 5.17 entropy.
FMLM★ also achieves the best 2–4 step results.

📎 Paper: https://t.co/mY0kjEXiSW
⌨️ Code: https://t.co/IRSVNxcpzN

[Jung Yeon Park @park_jungy]

@ruffini Cool work, love the framing that agents are forced to mirror the world's symmetries

[Jung Yeon Park @park_jungy]

(1/8) Happy to share that our paper LieFlow has been accepted to #ICML2026.

Symmetry is a strong inductive bias in ML, but you have to know the symmetry in advance. Can a model discover it directly from data, for both discrete and continuous groups? 🧵

[Jung Yeon Park @park_jungy]

@didacum333 Nice, did not know the connection. I'd love to learn more about this

[Jung Yeon Park @park_jungy]

Clever use of unitary convolutions to overcome graph oversmoothing, but relax them to tailor to the type of dynamics at hand (e.g. diffusive). Excellent work led by @bermanatorCodes and Luisa Li.

Come see us and the poster at #ICML2026 🇰🇷

[Eddie Berman @bermanatorCodes]

I'll be @icmlconf presenting our work Smoothness Errors in Dynamics Models and How to Avoid Them, with Luisa Li, @park_jungy , and @RobinSFWalters !!

Come see our poster #2200 , july 9th, 2:30 PM – 4:15 PM, hall A.

Links to paper, website, and code in thread!

[Jung Yeon Park @park_jungy]

@Nipsuli Thank you!

[Jung Yeon Park @park_jungy]

(8/8) Links:
🌐 https://t.co/LXjMbvq4Ly
📄 https://t.co/ZVibB4ZEkD
💻 https://t.co/QbzV7mE0u0

Joint work with @YuxuanChen23, @FEijkelboom, @jiankeyang, @jwvdm, Lawson L.S. Wong, and @RobinSFWalters.

[Jung Yeon Park @park_jungy]

(7/8) On MI-Motion (real human motion capture), no target symmetry is imposed; it must be learned from data. People move along axis-aligned directions while gravity breaks full SO(3), inducing an approximate C4 about the z-axis, which LieFlow recovers while LieGAN does not.

[Ali Madani @thisismadani]

AI has two modes in drug discovery. Accelerate: moving faster through the existing playbook. Unlock: opening frontiers that weren't possible before.

Excited to announce Profluent is partnering with Eli Lilly, the global pharma powerhouse, to unlock breakthrough medicines for patients.

It's a big deal beyond the numbers ($2.25B + royalties): we’ll get to use our frontier AI models and foundational datasets to design proteins focused on large gene insertion, a therapeutic moonshot.

Proteins govern almost everything in biology. We've built a generalizable AI platform to design all proteins. Onward!

[Boston Symmetry Group @bostonsymmetry]

Boston Symmetry Day is happening TODAY at Northeastern University’s Columbus Place and Alumni Center (716 Columbus Ave, 6th floor)! Breakfast starts at 9 AM, but talks are happening throughout the day, followed by a social. We’ll see you there!

[Boston Symmetry Group @bostonsymmetry]

Registration is now open for Boston Symmetry Day on March 31! Sign up by March 21st at https://t.co/kjpiejs0gB

We have an exciting lineup of speakers (see our website: https://t.co/2V7RQwloCW ) 
Also featuring a poster session so you have a chance to present your awesome work!