Hannah Lawrence

81 posts

Hannah Lawrence banner
Hannah Lawrence

Hannah Lawrence

@HLawrenceCS

PhD @ MIT CSAIL. Finding/exploiting patterns in data, especially symmetries. AI4Science. https://t.co/0XcSE5V8S2. Cofounder @bostonsymmetry.

Katılım Nisan 2021
721 Takip Edilen899 Takipçiler
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
@DimitrisPapail Really neat head-to-head comparison! The "pair tokens" are reminiscent of the "index hints" that arxiv.org/pdf/2310.16028 uses (Section 5.1) to enable length generalization in arithmetic. Maybe Codex's solution will generalize to >10 digit addition!
English
0
0
0
114
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
@VictorKnox99 @GRaM_org_ Speaking for both myself and the workshop: yes, very much! Work on the geometry of activations would be a perfect fit.
English
0
0
2
58
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
Have a new result to showcase? Want to explain a concept people often misunderstand? Write a blogpost for the @GRaM_org_ workshop! Opinions, open problems, and reproducibility results are all welcome. Not due until April 6, so you can even start now. 😉 gram-workshop.github.io
English
1
9
31
3.7K
Hannah Lawrence retweetledi
Samuel Stanton
Samuel Stanton@samuel_stanton_·
My cofounders and I have been building Coefficient Bio for the past 5 months. We're ushering biopharma into the Intelligence Age. It will change everything about how the industry learns and makes decisions. If that's a future you want to build, get in touch!
Nathan C. Frey@nc_frey

Join the early team @CoefficientBio. We have an AI team that I’m incredibly proud of, and truly think is the most effective and exceptional technical team building in AI x bio. We're looking for people to: * Build and maintain robust AI systems to an exacting standard * Design and run experiments that let us iterate fast * Work on challenging scientific and engineering problems for human flourishing NYC-based (or willing to relocate). If you’re curious about what we’re building, reach out directly. If we don’t know each other yet, a great cold DM (or email: join@coefficientbio.com), or warm intro goes a long way.

English
4
9
107
13.7K
Hannah Lawrence retweetledi
GRaM Workshop at ICLR 2026
GRaM Workshop at ICLR 2026@GRaM_org_·
📢The second edition of ✨GRaM workshop✨ is here this time at #ICLR26. 🌟Submit your exciting works in Geometry-grounded representations. We welcome submissions in multiple tracks i.e. 📄 Proceedings 📝extended abstract 👩‍🏫Tutorial/blogpost as well as an exciting challenge!
GRaM Workshop at ICLR 2026 tweet media
English
1
22
41
11.8K
Hannah Lawrence retweetledi
Teresa Huang
Teresa Huang@TeresaNHuang·
Want a flexible equivariant model that can process many symmetries? Come to our spotlight talk at @neur_reps tomorrow 9:10am (upper ballroom 6A), to learn about ✨Any-Subgroup Equivariant Networks✨. w/ the great team: Abhinav Goel, @dereklim_lzh, @HLawrenceCS, Stefanie Jegelka.
Teresa Huang tweet media
English
0
4
23
1.7K
Hannah Lawrence retweetledi
Oumar Kaba
Oumar Kaba@sekoumarkaba·
Happy to have presented this work with @HLawrenceCS @vportilheiro and Yan! Thanks to those who came! Check out the paper to learn about the link between symmetry breaking, equivariant distributions and positional encodings (+experiments on Ising models) arxiv.org/abs/2503.21985
Oumar Kaba tweet media
Hannah Lawrence@HLawrenceCS

Equivariant functions (e.g. GNNs) can't break symmetries, which can be problematic for generative models and beyond. Come to poster #207 Saturday at 10AM to hear about our solution: SymPE, or symmetry-breaking positional encodings! w/Vasco Portilheiro, Yan Zhang, @sekoumarkaba

English
1
5
50
5.2K
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
Equivariant functions (e.g. GNNs) can't break symmetries, which can be problematic for generative models and beyond. Come to poster #207 Saturday at 10AM to hear about our solution: SymPE, or symmetry-breaking positional encodings! w/Vasco Portilheiro, Yan Zhang, @sekoumarkaba
Hannah Lawrence tweet media
English
2
15
96
10.3K
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
@HordanSnir @sekoumarkaba E.g. even if the graph has no automorphisms, the eigenvectors have a sign ambiguity, so there is unnecessary "randomness" (and the ambiguity may not be broken by svd in the same way when you permute). It's still a decent method in practice. Happy to discuss more @ the poster :)
English
1
0
3
217
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
@HordanSnir @sekoumarkaba Yes, certainly this is one method of breaking symmetries! It fits within our more general framework. That said, the randomness you are injecting to break symmetries is "higher entropy" than strictly necessary, because of the ambiguities for eigenvectors.
English
1
0
3
236
Hannah Lawrence retweetledi
Boston Symmetry Group
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 tweet media
English
0
3
12
1.5K
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
@bostonsymmetry Reminder to register, and sign up to bring your most recent geometric deep learning poster(s)! Should be a great, low-pressure opportunity to share research ideas and meet other local researchers :)
English
0
0
3
69
Hannah Lawrence
Hannah Lawrence@HLawrenceCS·
Join us at Northeastern on March 31st for a great day of speakers, posters, and networking with the Boston-area equivariant learning community! (Plus: there will be free food!)
Boston Symmetry Group@bostonsymmetry

Registration is now open for Boston Symmetry Day on March 31! Sign up by March 21st at docs.google.com/forms/d/e/1FAI… We have an exciting lineup of speakers (see our website: bostonsymmetry.github.io )  Also featuring a poster session so you have a chance to present your awesome work!

English
0
2
9
1.1K

Keşfet

@DimitrisPapail @VictorKnox99 @GRaM_org_ @iclr_conf @neur_reps @dereklim_lzh @chaitjo @ADuvalinho