Picture for Andrew Slavin Ross

Andrew Slavin Ross

Learning Predictive and Interpretable Timeseries Summaries from ICU Data

Sep 22, 2021
Nari Johnson, Sonali Parbhoo, Andrew Slavin Ross, Finale Doshi-Velez

Figure 1 for Learning Predictive and Interpretable Timeseries Summaries from ICU Data

Figure 2 for Learning Predictive and Interpretable Timeseries Summaries from ICU Data

Figure 3 for Learning Predictive and Interpretable Timeseries Summaries from ICU Data

Figure 4 for Learning Predictive and Interpretable Timeseries Summaries from ICU Data

* 10 pages, 3 figures, AMIA 2021 Annual Symposium

Access Paper or Ask Questions

Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglement

Feb 09, 2021
Andrew Slavin Ross, Finale Doshi-Velez

Figure 1 for Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglement

Figure 2 for Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglement

Figure 3 for Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglement

Figure 4 for Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglement

Access Paper or Ask Questions

Evaluating the Interpretability of Generative Models by Interactive Reconstruction

Feb 02, 2021
Andrew Slavin Ross, Nina Chen, Elisa Zhao Hang, Elena L. Glassman, Finale Doshi-Velez

Figure 1 for Evaluating the Interpretability of Generative Models by Interactive Reconstruction

Figure 2 for Evaluating the Interpretability of Generative Models by Interactive Reconstruction

Figure 3 for Evaluating the Interpretability of Generative Models by Interactive Reconstruction

Figure 4 for Evaluating the Interpretability of Generative Models by Interactive Reconstruction

* CHI 2021 accepted paper

Access Paper or Ask Questions

Ensembles of Locally Independent Prediction Models

Nov 27, 2019
Andrew Slavin Ross, Weiwei Pan, Leo Anthony Celi, Finale Doshi-Velez

Figure 1 for Ensembles of Locally Independent Prediction Models

Figure 2 for Ensembles of Locally Independent Prediction Models

Figure 3 for Ensembles of Locally Independent Prediction Models

Figure 4 for Ensembles of Locally Independent Prediction Models

* This is an expansion of arXiv:1806.08716 with different applications and focus, accepted to AAAI 2020

Access Paper or Ask Questions

Human-in-the-Loop Interpretability Prior

Oct 30, 2018
Isaac Lage, Andrew Slavin Ross, Been Kim, Samuel J. Gershman, Finale Doshi-Velez

Figure 1 for Human-in-the-Loop Interpretability Prior

Figure 2 for Human-in-the-Loop Interpretability Prior

Figure 3 for Human-in-the-Loop Interpretability Prior

Figure 4 for Human-in-the-Loop Interpretability Prior

* To appear at NIPS 2018, selected for a spotlight. 13 pages (incl references and appendix)

Access Paper or Ask Questions

Training Machine Learning Models by Regularizing their Explanations

Sep 29, 2018
Andrew Slavin Ross

Figure 1 for Training Machine Learning Models by Regularizing their Explanations

Figure 2 for Training Machine Learning Models by Regularizing their Explanations

Figure 3 for Training Machine Learning Models by Regularizing their Explanations

Figure 4 for Training Machine Learning Models by Regularizing their Explanations

* Harvard CSE master's thesis; includes portions of arxiv:1703.03717 and arxiv:1711.09404

Access Paper or Ask Questions

Learning Qualitatively Diverse and Interpretable Rules for Classification

Jul 19, 2018
Andrew Slavin Ross, Weiwei Pan, Finale Doshi-Velez

Figure 1 for Learning Qualitatively Diverse and Interpretable Rules for Classification

Figure 2 for Learning Qualitatively Diverse and Interpretable Rules for Classification

Figure 3 for Learning Qualitatively Diverse and Interpretable Rules for Classification

Figure 4 for Learning Qualitatively Diverse and Interpretable Rules for Classification

* Presented at 2018 ICML Workshop on Human Interpretability in Machine Learning (WHI 2018), Stockholm, Sweden (revision fixes minor issues)

Access Paper or Ask Questions

Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients

Nov 26, 2017
Andrew Slavin Ross, Finale Doshi-Velez

Figure 1 for Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients

Figure 2 for Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients

Figure 3 for Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients

Figure 4 for Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients

* To appear in AAAI 2018

Access Paper or Ask Questions