Mitigating Bias in Calibration Error Estimation

google-research/google-research • • 15 Dec 2020

We find that binning-based estimators with bins of equal mass (number of instances) have lower bias than estimators with bins of equal width.

Similarity of Neural Network Representations Revisited

google-research/google-research • • ICML 2019 2019

We introduce a similarity index that measures the relationship between representational similarity matrices and does not suffer from this limitation.

Practical and Consistent Estimation of f-Divergences

google-research/google-research • • NeurIPS 2019

The estimation of an f-divergence between two probability distributions based on samples is a fundamental problem in statistics and machine learning.

BIG-bench Machine Learning Mutual Information Estimation +1

Differentiable Ranking and Sorting using Optimal Transport

google-research/google-research • • NeurIPS 2019

From this observation, we propose extended rank and sort operators by considering optimal transport (OT) problems (the natural relaxation for assignments) where the auxiliary measure can be any weighted measure supported on $m$ increasing values, where $m \ne n$.

A Benchmark for Interpretability Methods in Deep Neural Networks

google-research/google-research • • NeurIPS 2019

We propose an empirical measure of the approximate accuracy of feature importance estimates in deep neural networks.

Feature Importance Image Classification

Optimizing Generalized Rate Metrics with Three Players

google-research/google-research • • NeurIPS 2019

We present a general framework for solving a large class of learning problems with non-linear functions of classification rates.

Fairness

Memory Efficient Adaptive Optimization

google-research/google-research • • NeurIPS 2019

Adaptive gradient-based optimizers such as Adagrad and Adam are crucial for achieving state-of-the-art performance in machine translation and language modeling.

Language Modelling Machine Translation +1

Search on the Replay Buffer: Bridging Planning and Reinforcement Learning

google-research/google-research • • NeurIPS 2019

We introduce a general control algorithm that combines the strengths of planning and reinforcement learning to effectively solve these tasks.

reinforcement-learning Reinforcement Learning (RL)

Charformer: Fast Character Transformers via Gradient-based Subword Tokenization

google-research/google-research • • ICLR 2022

In this paper, we propose a new model inductive bias that learns a subword tokenization end-to-end as part of the model.

Inductive Bias Linguistic Acceptability +3

Meta-Learning Requires Meta-Augmentation

google-research/google-research • • NeurIPS 2020

Meta-learning algorithms aim to learn two components: a model that predicts targets for a task, and a base learner that quickly updates that model when given examples from a new task.

Meta-Learning