Oscar Davis
@osclsd
Followers
326
Following
3K
Media
9
Statuses
48
PhD ML @UniofOxford; generative modelling; previously at @MSFTResearch, @EPFL, @imperialcollege
Oxford, UK
Joined May 2024
Introducing Generalised Flow Maps π A stable, few-step generative model on Riemannian manifolds πͺ© π Read it at: https://t.co/iCTHedwCxf πΎ Code: https://t.co/MeukcthFN2
@msalbergo @nmboffi @mmbronstein @bose_joey
3
22
112
Come do a PhD with me π! Promise of fun science and great coffee β
31
70
735
@msalbergo @nmboffi @mmbronstein @bose_joey Animation made with manim: https://t.co/Hq8C2MwK1c Earth texture from:
ftp.solarsystemscope.com
Online 3D simulation of the Solar System and night sky in real time - the Sun, planets, dwarf planets, comets, stars and constellations
0
0
4
As the IMM paper came out in March, I implemented it myself for some project, before the true source code was made available. I am releasing my version now: https://t.co/VXOBw91GXk It contains most/all features, and should be easy to (re-)use! Hope someone finds it helpful π
github.com
Non-official Inductive Moment Matching implementation in PyTorch with Lightning. Clean and simple. - olsdavis/imm
1
1
9
Very excited to share this! We introduce a new approach to knowledge graph foundation models built on probabilistic equivariance. The model is simple, expressive, and probabilistically equivariant β and it works remarkably well! Collaboration led by @jw9730 and @hxyscott.
New preprint: Flock, a foundation model for link predictions on knowledge graphs that zero-shot generalizes to novel entities and relations. Instead of message passing, Flock operates on anonymized random walks, processed by sequence neural nets. Paper: https://t.co/bKmKwmh7Fa
0
6
42
π’Interested in doing a PhD in generative models π€, AI4Science π§¬, Sampling π§βπ¬, and beyond? I am hiring PhD students at Imperial College London @ICComputing for the next application cycle. πSee the call below: https://t.co/kAG4qdTHXt And a light expression of interest:
5
61
225
Apply for the AITHYRA-CeMM International PhD Program! 15-20 fully funded PhD fellowships available in Vienna in AI/ML and Life Sciences Deadline for applications: 10 September 2025 https://t.co/0gQFlg9sQX
1
33
82
Wrapping up #ICML2025 on a high note β thrilled (and pleasantly surprised!) to win the Best Paper Award at @genbio_workshop π Big shoutout to the team that made this happen! Paper: Forward-Only Regression Training of Normalizing Flows ( https://t.co/2dMjkvF4qX)
@Mila_Quebec
6
12
108
GenBio Workshop ORAL Presentation π Title: FORT: Forward-Only Regression Training of Normalizing Flows π When: Fri 18 Jul πΊοΈ Where: East Exhibition Hall A π arXiv: https://t.co/Sgx2A2xwXN w/ @danyalrehman17 @osclsd @jiarlu @tangjianpku @mmbronstein @Yoshua_Bengio
1
5
9
π¨ Excited to announce that "How Expressive are Knowledge Graph Foundation Models?" is coming to ICML 2025! π π
Wednesday, July 16th π 4:30 PM π Booth #E-3011 Come by to chat about motifs, expressiveness, and the future of graph foundation models! πππ
Knowledge Graph Foundation Models (KGFMs) are at the frontier of graph learning - but we didnβt have a principled understanding of what we can (or canβt) do with them. Now we do! π‘π 𧡠with Pablo Barcelo, @ismaililkanc, @mmbronstein, @michael_galkin, @JuanLReutter, @OrthMiguel
2
18
125
π¨ ICML 2025 Paper π¨ "On Measuring Long-Range Interactions in Graph Neural Networks" We formalize the long-range problem in GNNs: π‘Derive a principled range measure π§ Tools to assess models & benchmarks π¬Critically assess LRGB 𧡠Thread below π #ICML2025
3
16
52
πPersonal update: I'm thrilled to announce that I'm joining Imperial College London @imperialcollege as an Assistant Professor of Computing @ICComputing starting January 2026. My future lab and I will continue to work on building better Generative Models π€, the hardest
98
34
627
Excited to release FORT, a new regression-based approach for training normalizing flows π₯! π Paper available here: https://t.co/2dMjkvEwBp New paper w/ @osclsd @jiarlu @tangjianpku @mmbronstein @Yoshua_Bengio @AlexanderTong7 @bose_joey π§΅1/6
2
23
89
Great final lecture in @mmbronstein GDL course given by @bose_joey πfeat famous smiley gif by @osclsd π
1
5
36
NeurIPS'24 has over 4k papers! Below is my selection of 5 papers which considers information geometry: 1/ https://t.co/6iLldc9hOa 2/ https://t.co/0gyl6nD0Q4 3/ https://t.co/ZwIsBGxvIv 4/ https://t.co/mmAOvrTpWm 5/
1
21
152
New preprint! π¨ We scale equilibrium sampling to hexapeptide (in cartesian coordinates!) with Sequential Boltzmann generators!Β π π€― Work with @bose_joey, @WillLin1028, @leonklein26, @mmbronstein and @AlexanderTong7 Thread 𧡠1/11
3
23
76
Knowledge Graph Foundation Models (KGFMs) are at the frontier of graph learning - but we didnβt have a principled understanding of what we can (or canβt) do with them. Now we do! π‘π 𧡠with Pablo Barcelo, @ismaililkanc, @mmbronstein, @michael_galkin, @JuanLReutter, @OrthMiguel
5
88
334
π© Topological blindspots is coming to ICLR as an oral presentation! π© We prove that message-passing based topological deep learning (TDL) architectures are unable capture basic topological invariants including homology, orientability, planarity and more.
4
25
153