• Publications
  • Influence
Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
TLDR
This work empirically demonstrates that its algorithms significantly reduce gender bias in embeddings while preserving the its useful properties such as the ability to cluster related concepts and to solve analogy tasks.
Adaptive Neural Networks for Efficient Inference
TLDR
It is shown that computational time can be dramatically reduced by exploiting the fact that many examples can be correctly classified using relatively efficient networks and that complex, computationally costly networks are only necessary for a small fraction of examples.
The What-If Tool: Interactive Probing of Machine Learning Models
TLDR
The What-If Tool is an open-source application that allows practitioners to probe, visualize, and analyze ML systems, with minimal coding, and lets practitioners measure systems according to multiple ML fairness metrics.
XRAI: Better Attributions Through Regions
TLDR
A novel region-based attribution method that builds upon integrated gradients and contributes an axiom-based sanity check for attribution methods is presented and it is shown that XRAI produces better results than other saliency methods for common models and the ImageNet dataset.
The Language Interpretability Tool: Extensible, Interactive Visualizations and Analysis for NLP Models
TLDR
The Language Interpretability Tool (LIT), an open-source platform for visualization and understanding of NLP models, is presented, which integrates local explanations, aggregate analysis, and counterfactual generation into a streamlined, browser-based interface to enable rapid exploration and error analysis.
Quantifying and Reducing Stereotypes in Word Embeddings
TLDR
A novel gender analogy task is created and combined with crowdsourcing to systematically quantify the gender bias in a given embedding, and an efficient algorithm is developed that reduces gender stereotype using just a handful of training examples while preserving the useful geometric properties of the embedding.
Adaptive Neural Networks for Fast Test-Time Prediction
TLDR
An adaptive network evaluation scheme is posed, where a system to adaptively choose the components of a deep network to be evaluated for each example, to reduce the evaluation time on new examples without loss of classification performance.
BING++: A Fast High Quality Object Proposal Generator at 100fps
TLDR
A novel object proposal algorithm is proposed which inherits the good computational efficiency of BING but significantly improves its proposal localization quality and also recursively improves BING++'s proposals by exploiting the fact that edges in images are typically associated with object boundaries.
Debiasing Embeddings for Reduced Gender Bias in Text Classification
TLDR
It is shown that traditional techniques for debiasing embeddings can actually worsen the bias of the downstream classifier by providing a less noisy channel for communicating gender information, and how these same techniques can be used to simultaneously reduce bias and maintain high classification accuracy.
Model Selection by Linear Programming
TLDR
This work considers a binary tree architecture where each leaf corresponds to a different model and shows that adaptive model selection reduces to a linear program thus realizing substantial computational efficiencies and guaranteed convergence properties.
...
1
2
3
...