@oxcsml

Computational Statistics, Statistical Methodology, and Machine Learning Research Group. Based at @OxfordStats.

 Oxford, England
csml.stats.ox.ac.uk/
Joined November 2018
  • Tweets318
  • Following165
  • Followers3.3K
12Media
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
Bayes for continual learning: it just works, right? Turns out the details matter for *approx* Bayes in deep nets. Variational inference over predictive functions works pretty well! #ICML2022 paper with @timrudner @qixuan_feng @yeewhye @yaringal proceedings.mlr.press/v162/r… 1/6

Continual Learning via Sequential Function-Space Variational Inference

Sequential Bayesian inference over predictive functions is a natural framework for continual learning from streams of data. However, applying it to neural ne...

proceedings.mlr.press
1
22
79
In our #ICML paper, we show that *Sequential Function-Space Variational Inference* allows for effective Bayesian updating in NNs and enables state-of-the-art continual learning!🚀 ✍️Paper: timrudner.com/sfsvi 💻Code+Notebooks: timrudner.com/sfsvi-code Details in 🧵! [1/N]
1
23
1
114
Congratulations @desirivanova . Well done!
Isaac Newton Institute@NewtonInstitute7 Mar 2022
Congratulations to @desirivanova 🥉 (@OxUniMaths), Matthew Tointon 🥇 (@BristolUniMaths) and Teresa Bautista 🥈 (@KingsCollegeLon) on their @STEM4Brit awards this afternoon! 👏🎉🧮 #mathematics #stem @UKParliament
6
Applications due this Friday. Please retweet everyone, and please apply if you are a UK undergrad from under-represented backgrounds interested in what research in artificial intelligence and machine learning is like! 🙏
Yee Whye Teh@yeewhye7 Feb 2022
Oxford's Uniq+ graduate access program has an AI/ML strand! UK undergrads from under-represented backgrounds interested in AI/ML pls apply. Scholarship, free lodging, 6 or 10 weeks summer research project w/ faculty @CompSciOxford @OxfordStats @OxfordEng. stats.ox.ac.uk/archive-news/…
1
25
1
44
If you're an undergraduate looking for a summer research project in AI or ML - check this out! Up to £4,200 stipend for 10-week project, free accommodation, great projects. Apply by 18 Feb. stats.ox.ac.uk/archive-news/…
13
13
Oxford's Uniq+ graduate access program has an AI/ML strand! UK undergrads from under-represented backgrounds interested in AI/ML pls apply. Scholarship, free lodging, 6 or 10 weeks summer research project w/ faculty @CompSciOxford @OxfordStats @OxfordEng. stats.ox.ac.uk/archive-news/…
1
32
2
69
Uncertainty Quantification in End-to-End Implicit Neural Representations for Medical Imaging Francisca Vasconcelos, @bobby_he, @yeewhye Oral at MedNeurIPS: neurips.cc/virtual/2021/work… Poster at Bayesian Deep Learning Workshop: neurips.cc/virtual/2021/work… Paper: cse.cuhk.edu.hk/~qdou/public…
1
2
2
🤖 Uncertainty Baselines: Benchmarks for Uncertainty & Robustness in Deep Learning 🤖 📄Paper: arxiv.org/abs/2106.04015 📡Blog: ai.googleblog.com/2021/10/ba… 🖥️Code: github.com/google/uncertaint… @zacharynado @neilbband @timrudner @sirbayes @balajiln @yaringal @dustinvtran
1
5
8
We also have three workshop papers: 🤖 PCA Subspaces Are Not Always Optimal for Bayesian Learning 🤖 📄Paper: openreview.net/forum?id=iPYH… 🎙️Talk: nips.cc/Conferences/2021/Sch… Alexandre Bense @amirjoudaki @timrudner @vincefort
1
3
4
Multi-Facet Clustering Variational Autoencoders Poster session: Fri 10 Dec 8:30 a.m. PST — 10 a.m. PST neurips.cc/virtual/2021/post… Paper: arxiv.org/abs/2106.05241
1
1
On Contrastive Representations of Stochastic Processes @MathieuEmile, @AdamEFoster, @yeewhye Poster session 1: neurips.cc/virtual/2021/post… Paper: arxiv.org/abs/2106.10052 Code: github.com/ae-foster/contras…
1
2
On Locality of Local Explanation Models: We investigate why popular model-agnostic local explanation models like SHAP are not locally faithful and how their locality can be increased with neighbourhood sampling. With @SGhalebikesabi, @TerLucile,@karlado,@cholmesuk
1
3
Uniform Sampling over Episode Difficulty (Spotlight) Authors: Sébastien M. R. Arnold*, Guneet S. Dhillon*, Avinash Ravichandran, Stefano Soatto Sampling uniformly over episode difficulty outperforms curriculum learning, easy / hard mining, and the standard baseline.
1
2
@ValentinDeBort1,@JamesTThorn,@jeremyhengjm, @ArnaudDoucet1 Paper: //arxiv.org/abs/2106.01357 Code: github.com/JTT94/diffusion_s…

GitHub - JTT94/diffusion_schrodinger_bridge: PyTorch Implementation of DSB for Score Based Genera...

PyTorch Implementation of DSB for Score Based Generative Modeling. Experiments managed using Hydra. - JTT94/diffusion_schrodinger_bridge

github.com
1
1
2
Diffusion Schrödinger Bridge with Applications to Score-Based Generative Modeling (spotlight) A principled generalization of score/diffusion based generative modeling using iterative reverse diffusions, with connections to optimal transport.
1
3
🤖 Outcome-Driven Reinforcement Learning via Variational Inference 🤖 🎙️Talk: neurips.cc/virtual/2021/post… 📄Paper: openreview.net/forum?id=4bza… 🖥️Results: sites.google.com/view/od-ac @timrudner @vitchyr @rowantmc @yaringal @svlevine
1
1
2
12
🤖 On Pathologies in KL-Regularized Reinforcement Learning from Expert Demonstrations 🤖 🎙️Talk: neurips.cc/virtual/2021/post… 📄Paper: openreview.net/forum?id=sS8r… 🖥️Results: sites.google.com/view/nppac @timrudner @cong_ml @maosbot @yaringal @yeewhye
1
3
1
6
Self-Attention Between Datapoints: Going Beyond Individual Input-Output Pairs in Deep Learning: @janundnik @neilbband* @clarelyle, @AidanNGomez, @tom_rainforth, @yaringal, @OATML_Oxford 🚀 Non-Parametric Transformers 🚀 papers.nips.cc/paper/2021/ha…
1
2
5
BayesIMP: Uncertainty Quantification for Causal Data Fusion We propose a Bayesian Conditional Mean Embedding to embed the interventional distribution in the RKHS to estimate the average treatment effect with uncertainty quantification under the causal data fusion setting.
1
3
4
Further info: Poster session 1: neurips.cc/virtual/2021/sess… Paper: arxiv.org/abs/2105.12909 Code: github.com/shahineb/decondit…
1
3
Load more