no code implementations • ICML 2020 • Yibo Yang, Robert Bamler, Stephan Mandt
Deep Bayesian latent variable models have enabled new approaches to both model and data compression.
no code implementations • 7 Apr 2024 • Andi Zhang, Tim Z. Xiao, Weiyang Liu, Robert Bamler, Damon Wischik
We revisit the likelihood ratio between a pretrained large language model (LLM) and its finetuned variant as a criterion for out-of-distribution (OOD) detection.
1 code implementation • 19 Mar 2024 • Hanqi Zhou, Robert Bamler, Charley M. Wu, Álvaro Tejero-Cantero
This requires estimates of both the learner's progress (''knowledge tracing''; KT), and the prerequisite structure of the learning domain (''knowledge mapping'').
no code implementations • 28 Feb 2024 • Laura Manduchi, Kushagra Pandey, Robert Bamler, Ryan Cotterell, Sina Däubener, Sophie Fellenz, Asja Fischer, Thomas Gärtner, Matthias Kirchler, Marius Kloft, Yingzhen Li, Christoph Lippert, Gerard de Melo, Eric Nalisnick, Björn Ommer, Rajesh Ranganath, Maja Rudolph, Karen Ullrich, Guy Van Den Broeck, Julia E Vogt, Yixin Wang, Florian Wenzel, Frank Wood, Stephan Mandt, Vincent Fortuin
The field of deep generative modeling has grown rapidly and consistently over the years.
1 code implementation • 31 Dec 2023 • Tim Z. Xiao, Weiyang Liu, Robert Bamler
Bayesian neural networks (BNNs) are a principled approach to modeling predictive uncertainties in deep learning, which are important in safety-critical applications.
no code implementations • 30 Oct 2023 • Tim Z. Xiao, Johannes Zenn, Robert Bamler
Variational autoencoders (VAEs) are popular models for representation learning but their encoders are susceptible to overfitting (Cremer et al., 2018) because they are trained on a finite training set instead of the true (continuous) data distribution $p_{\mathrm{data}}(\mathbf{x})$.
no code implementations • 30 Oct 2023 • Tim Z. Xiao, Johannes Zenn, Robert Bamler
However, with this work, we aim to warn the community about an issue of the SVHN dataset as a benchmark for generative modeling tasks: we discover that the official split into training set and test set of the SVHN dataset are not drawn from the same distribution.
1 code implementation • 27 Apr 2023 • Johannes Zenn, Robert Bamler
Annealed Importance Sampling (AIS) moves particles along a Markov chain from a tractable initial distribution to an intractable target distribution.
1 code implementation • 9 Feb 2023 • Tim Z. Xiao, Robert Bamler
Variational Autoencoders (VAEs) were originally motivated (Kingma & Welling, 2014) as probabilistic generative models in which one performs approximate Bayesian inference.
no code implementations • 17 Feb 2022 • Fabian Jirasek, Robert Bamler, Stephan Mandt
We apply the new approach to predict activity coefficients at infinite dilution and obtain significant improvements compared to the data-driven and physical baselines and established ensemble methods from the machine learning literature.
1 code implementation • 5 Jan 2022 • Robert Bamler
Entropy coding is the backbone data compression.
1 code implementation • NeurIPS 2020 • Alex Boyd, Robert Bamler, Stephan Mandt, Padhraic Smyth
Modeling such data can be very challenging, in particular for applications with many different types of events, since it requires a model to predict the event types as well as the time of occurrence.
2 code implementations • NeurIPS 2020 • Yibo Yang, Robert Bamler, Stephan Mandt
We consider the problem of lossy image compression with deep latent variable models.
2 code implementations • ICML 2020 • Yibo Yang, Robert Bamler, Stephan Mandt
Our experimental results demonstrate the importance of taking into account posterior uncertainties, and show that image compression with the proposed algorithm outperforms JPEG over a wide range of bit rates using only a single standard VAE.
1 code implementation • ICLR 2020 • Robert Bamler, Stephan Mandt
Training a classifier over a large number of classes, known as 'extreme classification', has become a topic of major interest with applications in technology, science, and e-commerce.
no code implementations • 29 Jan 2020 • Fabian Jirasek, Rodrigo A. S. Alves, Julie Damay, Robert A. Vandermeulen, Robert Bamler, Michael Bortz, Stephan Mandt, Marius Kloft, Hans Hasse
Activity coefficients, which are a measure of the non-ideality of liquid mixtures, are a key property in chemical engineering with relevance to modeling chemical and phase equilibria as well as transport processes.
no code implementations • 30 Sep 2019 • Robert Bamler, Cheng Zhang, Manfred Opper, Stephan Mandt
In this paper, we revisit perturbation theory as a powerful way of improving the variational approximation.
no code implementations • 4 Jul 2019 • Robert Bamler, Stephan Mandt
Continuous symmetries and their breaking play a prominent role in contemporary physics.
1 code implementation • 1 Jul 2019 • Robert Bamler, Farnood Salehi, Stephan Mandt
Knowledge graph embeddings rank among the most successful methods for link prediction in knowledge graphs, i. e., the task of completing an incomplete collection of relational facts.
Ranked #5 on Link Prediction on FB15k
no code implementations • 27 Sep 2018 • Farnood Salehi, Robert Bamler, Stephan Mandt
We develop a probabilistic extension of state-of-the-art embedding models for link prediction in relational knowledge graphs.
no code implementations • ICML 2018 • Robert Bamler, Stephan Mandt
We show that representation learning models for time series possess an approximate continuous symmetry that leads to slow convergence of gradient descent.
no code implementations • 8 Mar 2018 • Robert Bamler, Stephan Mandt
We show that representation learning models for time series possess an approximate continuous symmetry that leads to slow convergence of gradient descent.
no code implementations • 10 Nov 2017 • Geng Ji, Robert Bamler, Erik B. Sudderth, Stephan Mandt
Word2vec (Mikolov et al., 2013) has proven to be successful in natural language processing by capturing the semantic relationships between different words.
no code implementations • NeurIPS 2017 • Robert Bamler, Cheng Zhang, Manfred Opper, Stephan Mandt
Black box variational inference (BBVI) with reparameterization gradients triggered the exploration of divergence measures other than the Kullback-Leibler (KL) divergence, such as alpha divergences.
no code implementations • 4 Jul 2017 • Robert Bamler, Stephan Mandt
Black box variational inference with reparameterization gradients (BBVI) allows us to explore a rich new class of Bayesian non-conjugate latent time series models; however, a naive application of BBVI to such structured variational models would scale quadratically in the number of time steps.
1 code implementation • ICML 2017 • Robert Bamler, Stephan Mandt
We present a probabilistic language model for time-stamped text data which tracks the semantic evolution of individual words over time.