article thumbnail

Explaining black-box models using attribute importance, PDPs, and LIME

Domino Data Lab

Model distillation – this approach builds a separate explainable model that mimics the input-output behaviour of the deep network. Because this separate model is essentially a white-box, it can be used for extraction of rules that explain the decisions behind the ANN. 2016) for an example of this technique (LIME).

Modeling 139
article thumbnail

Tackling Bias in Machine Learning

Insight

Bias in Machine Learning Algorithms (Bottom Photos Source: ProPublica ; Top Photos Source: Pexels.com) Biases in predictive modeling are a widespread issue Machine learning and AI applications are used across industries, from recommendation engines to self-driving cars and more. 5 is labeled as low.

Insiders

Sign Up for our Newsletter

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Trending Sources

article thumbnail

2019 US Open Predictions: Doubling Down on the Data

DataRobot Blog

Using this data, we built a historical dataset containing past results, current Elo scores (both overall and surface-specific) and tournament information, then used DataRobot to determine the best model and predict the probability that a player would win a set. The US Open has begun, and the world is watching.

article thumbnail

Topics to watch at the Strata Data Conference in New York 2019

O'Reilly on Data

Machine learning, artificial intelligence, data engineering, and architecture are driving the data space. The Strata Data Conferences helped chronicle the birth of big data, as well as the emergence of data science, streaming, and machine learning (ML) as disruptive phenomena. 1 again in proposals this year.

IoT 20
article thumbnail

Techniques for Collecting, Prepping, and Plotting Data: Predicting Social Media-Influence in the NBA

Domino Data Lab

Eighty percent of this problem is collecting the data and then transforming the data. The other 20 percent is ML- and data science–related tasks like finding the right model, doing EDA, and feature engineering. Gathering the Data. there is a list of data sources to extract and transform. In Figure 6.1,

article thumbnail

Using random effects models in prediction problems

The Unofficial Google Data Science Blog

We have many routine analyses for which the sparsity pattern is closer to the nested case and lme4 scales very well; however, our prediction models tend to have input data that looks like the simulation on the right. A Scalable Blocked Gibbs Sampling Algorithm For Gaussian And Poisson Regression Models." 7] Nicholas A.

article thumbnail

Deep Learning Illustrated: Building Natural Language Processing Models

Domino Data Lab

Data scientists and researchers require an extensive array of techniques, packages, and tools to accelerate core work flow tasks including prepping, processing, and analyzing data. Utilizing NLP helps researchers and data scientists complete core tasks faster. Note: Google Translate has incorporated NMT since 2016.