Search Results for author:
Found 62 papers, 31 papers with code
Coherent energy and force uncertainty in deep learning force fields
In machine learning energy potentials for atomic systems, forces are commonly obtained as the negative derivative of the energy function with respect to atomic positions.
BEND: Benchmarking DNA Language Models on biologically meaningful tasks
The genome sequence contains the blueprint for governing cellular processes.
DiffEnc: Variational Diffusion with a Learned Encoder
Diffusion models may be viewed as hierarchical variational autoencoders (VAEs) with two improvements: parameter sharing for the conditional distributions in the generative process and efficient computation of the loss as independent terms over the hierarchy.
Image-free Classifier Injection for Zero-Shot Classification
We achieve this with our proposed Image-free Classifier Injection with Semantics (ICIS) that injects classifiers for new, unseen classes into pre-trained classification models in a post-hoc fashion without relying on image data.
Addressing caveats of neural persistence with deep graph persistence
Whilst this captures useful information for linear classifiers, we find that no relevant spatial structure is present in later layers of deep neural networks, making neural persistence roughly equivalent to the variance of weights.
Implicit Transfer Operator Learning: Multiple Time-Resolution Surrogates for Molecular Dynamics
Computing properties of molecular systems rely on estimating expectations of the (unnormalized) Boltzmann distribution.
Graph Neural Network Interatomic Potential Ensembles with Calibrated Aleatoric and Epistemic Uncertainty on Energy and Forces
The proposed method considers both epistemic and aleatoric uncertainty and the total uncertainties are recalibrated post hoc using a nonlinear scaling function to achieve good calibration on previously unseen data, without loss of predictive accuracy.
Dermatological Diagnosis Explainability Benchmark for Convolutional Neural Networks
We pre-trained all architectures on an clinical skin disease dataset, and fine-tuned them on a DermXDB subset.
Unifying Molecular and Textual Representations via Multi-task Language Modelling
Here, we propose the first multi-domain, multi-task language model that can solve a wide range of tasks in both the chemical and natural language domains.
Ranked #3 on
Molecule Captioning
on ChEBI-20
ThoughtSource: A central hub for large language model reasoning data
Large language models (LLMs) such as GPT-4 have recently demonstrated impressive results across a wide range of tasks.
Variational Open-Domain Question Answering
Retrieval-augmented models have proven to be effective in natural language processing tasks, yet there remains a lack of research on their optimization using variational inference.
Ranked #4 on
Multiple Choice Question Answering (MCQA)
on MedMCQA
Explainable Image Quality Assessments in Teledermatological Photography
For poor image quality explanations, our method obtains F1-scores of between 0. 37 +- 0. 01 and 0. 70 +- 0. 01, similar to the inter-rater pairwise F1-score of between 0. 24 +- 0. 15 and 0. 83 +- 0. 06.
Transition1x -- a Dataset for Building Generalizable Reactive Machine Learning Potentials
In this work, we present the dataset Transition1x containing 9. 6 million Density Functional Theory (DFT) calculations of forces and energies of molecular configurations on and around reaction pathways at the wB97x/6-31G(d) level of theory.
NeuralNEB -- Neural Networks can find Reaction Paths Fast
We also compare with and outperform Density Functional based Tight Binding (DFTB) on both accuracy and computational resource.
Can large language models reason about medical questions?
Although large language models (LLMs) often produce impressive outputs, it remains unclear how they perform in real-world scenarios requiring strong reasoning skills and expert domain knowledge.
Ranked #5 on
Question Answering
on PubMedQA
Multiple-choice
Multiple Choice Question Answering (MCQA)
+3
Few-Shot Diffusion Models
At test time, the model is able to generate samples from previously unseen classes conditioned on as few as 5 samples from that class.
Inductive Biases for Object-Centric Representations in the Presence of Complex Textures
Understanding which inductive biases could be helpful for the unsupervised learning of object-centric representations of natural scenes is challenging.
Image Super-Resolution With Deep Variational Autoencoders
Image super-resolution (SR) techniques are used to generate a high-resolution image from a low-resolution image.
DermX: an end-to-end framework for explainable automated dermatological diagnosis
We assess the explanation performance in terms of identification and localization by comparing model-selected with dermatologist-selected explanations, and gradient-weighted class-activation maps with dermatologist explanation maps, respectively.
SCHA-VAE: Hierarchical Context Aggregation for Few-Shot Generation
In few-shot learning the model is trained on data from many sets from distributions sharing some underlying properties such as sets of characters from different alphabets or objects from different categories.
Calibrated Uncertainty for Molecular Property Prediction using Ensembles of Message Passing Neural Networks
In this work we extend a message passing neural network designed specifically for predicting properties of molecules and materials with a calibrated probabilistic predictive distribution.
The Role of Pretrained Representations for the OOD Generalization of Reinforcement Learning Agents
By training 240 representations and over 10, 000 reinforcement learning (RL) policies on a simulated robotic setup, we evaluate to what extent different properties of pretrained VAE-based representations affect the OOD generalization of downstream agents.
Generalization and Robustness Implications in Object-Centric Learning
The idea behind object-centric representation learning is that natural scenes can better be modeled as compositions of objects and their relations as opposed to distributed representations.
Representation Learning for Out-of-distribution Generalization in Reinforcement Learning
Learning data representations that are useful for various downstream tasks is a cornerstone of artificial intelligence.
Out-of-Distribution Generalization
reinforcement-learning
+2
Optimal Variance Control of the Score-Function Gradient Estimator for Importance-Weighted Bounds
Empirically, for the training of both continuous and discrete generative models, the proposed method yields superior variance reduction, resulting in an SNR for IWAE that increases with $K$ without relying on the reparameterization trick.
On the Transfer of Disentangled Representations in Realistic Settings
Learning meaningful representations that disentangle the underlying structure of the data generating process is considered to be of key importance in machine learning.
Optimal Variance Control of the Score Function Gradient Estimator for Importance Weighted Bounds
This paper introduces novel results for the score function gradient estimator of the importance weighted variational bound (IWAE).
SurVAE Flows: Surjections to Bridge the Gap between VAEs and Flows
Normalizing flows and variational autoencoders are powerful generative models that can represent complicated density functions.
Closing the Dequantization Gap: PixelCNN as a Single-Layer Flow
Flow models have recently made great progress at modeling ordinal discrete data such as images and audio.
Towards Hierarchical Discrete Variational Autoencoders
We introduce the Hierarchical Discrete Variational Autoencoder (HD-VAE): a hi- erarchy of variational memory layers.
LAVAE: Disentangling Location and Appearance
We propose a probabilistic generative model for unsupervised learning of structured, interpretable, object-based representations of visual scenes.
BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling
In this paper we close the performance gap by constructing VAE models that can effectively utilize a deep hierarchy of stochastic variables and model complex covariance structures.
Ranked #18 on
Image Generation
on ImageNet 32x32
(bpd metric)
Attend, Copy, Parse -- End-to-end information extraction from documents
We believe our proposed architecture can be used on many real life information extraction tasks where word classification cannot be used due to a lack of the required word-level labels.
A deep learning approach to identify local structures in atomic-resolution transmission electron microscopy images
Recording atomic-resolution transmission electron microscopy (TEM) images is becoming increasingly routine.
Materials Science
Recurrent Relational Networks for complex relational reasoning
Humans possess an ability to abstractly reason about objects and their interactions, an ability not shared with state-of-the-art deep learning models.
Feature Map Variational Auto-Encoders
There have been multiple attempts with variational auto-encoders (VAE) to learn powerful global representations of complex data using a combination of latent stochastic variables and an autoregressive model over the dimensions of the data.
Zero-shot Cross Language Text Classification
Labeled text classification datasets are typically only available in a few select languages.
Recurrent Relational Networks
We achieve state of the art results on the bAbI textual question-answering dataset with the recurrent relational network, consistently solving 20/20 tasks.
Ranked #3 on
Question Answering
on bAbi
(Mean Error Rate metric)
A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learning
This paper takes a step towards temporal reasoning in a dynamically changing video, not in the pixel space that constitutes its frames, but in a latent space that describes the non-linear dynamics of the objects in its world.
Hash Embeddings for Efficient Word Representations
In hash embeddings each token is represented by $k$ $d$-dimensional embeddings vectors and one $k$ dimensional weight vector.
CloudScan - A configuration-free invoice analysis system using recurrent neural networks
We describe a recurrent neural network model that can capture long range context and compare it to a baseline logistic regression model corresponding to the current CloudScan production system.
End-to-End Information Extraction without Token-Level Supervision
End-to-end (E2E) models, which take raw text as input and produce the desired output directly, need not depend on token-level labels.
Semi-Supervised Generation with Cluster-aware Generative Models
Deep generative models trained with large amounts of unlabelled data have proven to be powerful within the domain of unsupervised learning.
Neural Machine Translation with Characters and Hierarchical Encoding
Most existing Neural Machine Translation models use groups of characters or whole words as their unit of input and output.
Self-Averaging Expectation Propagation
Our approach extends the framework of (generalized) approximate message passing -- assumes zero-mean iid entries of the measurement matrix -- to a general class of random matrix ensembles.
Sequential Neural Models with Stochastic Layers
How can we efficiently propagate uncertainty in a latent state representation with recurrent neural networks?
An Adaptive Resample-Move Algorithm for Estimating Normalizing Constants
The estimation of normalizing constants is a fundamental step in probabilistic model comparison.
Auxiliary Deep Generative Models
The auxiliary variables leave the generative model unchanged but make the variational distribution more expressive.
Ranked #49 on
Image Classification
on SVHN
Ladder Variational Autoencoders
Variational Autoencoders are powerful models for unsupervised learning.
Autoencoding beyond pixels using a learned similarity metric
We present an autoencoder that leverages learned representations to better measure similarities in data space.
Recurrent Spatial Transformer Networks
We investigate different down-sampling factors (ratio of pixel in input and output) for the SPN and show that the RNN-SPN model is able to down-sample the input images without deteriorating performance.
Bayesian inference for spatio-temporal spike-and-slab priors
In this work, we address the problem of solving a series of underdetermined linear inverse problems subject to a sparsity constraint.
Spatio-temporal Spike and Slab Priors for Multiple Measurement Vector Problems
We are interested in solving the multiple measurement vector (MMV) problem for instances, where the underlying sparsity pattern exhibit spatio-temporal structure motivated by the electroencephalogram (EEG) source localization problem.
Convolutional LSTM Networks for Subcellular Localization of Proteins
Machine learning is widely used to analyze biological sequence data.
Deep Belief Nets for Topic Modeling
Applying traditional collaborative filtering to digital publishing is challenging because user data is very sparse due to the high volume of documents relative to the number of users.
Protein Secondary Structure Prediction with Long Short Term Memory Networks
Recurrent neural networks are an generalization of the feed forward neural network that naturally handle sequential data.
Bayesian leave-one-out cross-validation approximations for Gaussian latent variable models
The future predictive performance of a Bayesian model can be estimated using Bayesian cross-validation.
Bayesian Inference for Structured Spike and Slab Priors
Sparse signal recovery addresses the problem of solving underdetermined linear inverse problems subject to a sparsity constraint.
Scalable Bayesian Modelling of Paired Symbols
We present a novel, scalable and Bayesian approach to modelling the occurrence of pairs of symbols (i, j) drawn from a large vocabulary.
Perturbative Corrections for Approximate Inference in Gaussian Latent Variable Models
A perturbative expansion is made of the exact but intractable correction, and can be applied to the model's partition function and other moments of interest.
Bayesian Sparse Factor Models and DAGs Inference and Comparison
In this paper we present a novel approach to learn directed acyclic graphs (DAG) and factor models within the same framework while also allowing for model comparison between them.
Improving on Expectation Propagation
We develop as series of corrections to Expectation Propagation (EP), which is one of the most popular methods for approximate probabilistic inference.
