Melanie Weber
@mweber_PU
Followers
2K
Following
497
Statuses
75
Assistant Professor at Harvard @hseas. Previously Hooke Research Fellow @OxUniMaths and PhD @Princeton. Studying Geometry and Machine Learning.
Cambridge, MA
Joined November 2016
#NeurIPS2024 Hardness of Learning Neural Networks under the Manifold Hypothesis. Today 11-2. East Exhibit Hall A-C #2308.
New paper w/ @bobak_kiani, Jason Wang #NeurIPS2024: We study the complexity of learning neural networks under the manifold hypothesis, showing that learning on input manifolds of bounded curvature is hard, but additional volume assumptions ensure learnability.
0
3
31
#NeurIPS2024 Unitary Convolutions for Learning on Graphs and Groups. 11-2 today in East Exhibit Hall A-C #3108.
#NeurIPS2024 spotlight w/ @bobak_kiani, Lukas Fesser: Group-convolutional networks have shown great success on symmetric domains, but can suffer from instabilities as their depths increases, e.g., over-smoothing in GNNs. We propose unitary convolutions for learning on graphs and groups that allow for deeper networks with less instability.
0
2
33
RT @KempnerInst: There are two #KempnerInstitute affiliated spotlights at #NeurIPS2024 today! Be sure to check out “Unitary Convolutions fo…
0
2
0
RT @KempnerInst: College grads interested in intelligence research: the application for #KempnerInstitute's post-bac program w/ the @Harvar…
0
5
0
A postdoc position is available in my group at Harvard @hseas to perform research at the intersection of Geometry & Machine Learning. Research interests include Representation Learning, Learning on Graphs & Manifolds, and applications in the Sciences. Details here:
0
23
157
RT @neur_reps: 📢 Last North American NeurReps Speaker Series seminar before Europe! @Harvard 🇺🇸 🎙️@mweber_pu 🏷️ Discrete Curvature a…
0
6
0
#NeurIPS2024 spotlight w/ @bobak_kiani, Lukas Fesser: Group-convolutional networks have shown great success on symmetric domains, but can suffer from instabilities as their depths increases, e.g., over-smoothing in GNNs. We propose unitary convolutions for learning on graphs and groups that allow for deeper networks with less instability.
0
4
79
RT @Zakobian: In our preprint, for the 1st time we make neural operators equivariant with respect to PDE symmetry groups. These can be very…
0
42
0
New paper w/ @bobak_kiani, Jason Wang #NeurIPS2024: We study the complexity of learning neural networks under the manifold hypothesis, showing that learning on input manifolds of bounded curvature is hard, but additional volume assumptions ensure learnability.
1
28
239
RT @networkspapers: J. Phys. Complex.: Augmentations of Forman’s Ricci curvature and their applications in community detection https://t.co…
0
1
0
Harvard Applied Mathematics @hseas is looking for the next Carrier Fellow! Applicants from all areas of Applied Mathematics, including Geometric Machine Learning, are welcome.
0
17
53
RT @iaifi_news: A familiar face! @mweber_PU, who lectured at last week’s Summer School, is now presenting at the #IAIFISummerWorkshop on ge…
0
1
0
RT @KempnerInst: The application is now open for our #KempnerInstitute Research #Fellowship! Postdocs studying the foundations of #intellig…
0
30
0
On Sat, 11.40am, visit our poster @MLGenX to learn about contrastive hyperbolic embeddings for single-cell data analysis. with @nithyabhasker, Hattie Chung, Louis Boucherie, Vladislav Kim, and Stefanie Speidel
0
1
6
On Wed, 10.45am, Halle B #288, stop by to learn about Local Curvature Profiles, structural encodings based on discrete Ricci curvature that can enhance the performance of Graph Neural Networks. with Lukas Fesser
1
0
1