Published Apr 1, 2016

Machine Learning Done Wrong

Cheng-tao Chu, a seasoned entrepreneur and machine learning expert, delves into the common pitfalls of machine learning, exploring the advantages of nonlinear models, the essentials of feature engineering, and the crucial role of selecting appropriate loss functions for model accuracy and optimization.

Episode Highlights

Topics covered

Popular Clips

Episode Highlights

Related Episodes

The Computational Complexity of Machine Learning
Answers 383 questions
ML Ops
Answers 383 questions
Trusting Machine Learning Models with LIME
Answers 383 questions
Machine Learning on Images with Noisy Human-centric Labels
Answers 383 questions
Easily Fooling Deep Neural Networks
Answers 383 questions
Visualization and Interpretability
Answers 383 questions
Pandemic Machine Learning Pitfalls
Answers 383 questions
Defending Against Adversarial Attacks
Answers 383 questions
ML Ops Best Practices
Answers 383 questions
4 out of 5 Data Scientists Agree
Answers 383 questions
Big Data Doesn't Exist
Answers 383 questions
The Model Complexity Myth
Answers 383 questions
Drug Discovery with Machine Learning
Answers 383 questions
Blind Spots in Reinforcement Learning
Answers 383 questions
Introduction
Answers 383 questions

Dexa/Kyle Polich

Machine Learning Done Wrong

Topics covered

Popular Clips

Nonlinear Data Modeling

Solving Data Challenges

Addressing Class Imbalance

Multicollinearity Concerns

Uncovering Data Insights

Interpretation Challenges

Machine Learning Insights

Feature Engineering Techniques

Episode Highlights

Interpreting Models

Feature Engineering

Improving Loss Functions

Related Episodes

The Computational Complexity of Machine Learning

ML Ops

Trusting Machine Learning Models with LIME

Machine Learning on Images with Noisy Human-centric Labels

Easily Fooling Deep Neural Networks

Visualization and Interpretability

Pandemic Machine Learning Pitfalls

Defending Against Adversarial Attacks

ML Ops Best Practices

4 out of 5 Data Scientists Agree

Big Data Doesn't Exist

The Model Complexity Myth

Drug Discovery with Machine Learning

Blind Spots in Reinforcement Learning

Introduction

Machine Learning Done Wrong

Topics covered

Popular Clips

Episode Highlights

Interpreting ModelsCheng-tao Chu discusses the pitfalls of logistic regression and the benefits of nonlinear models in machine learning. He emphasizes the importance of choosing the right algorithms to avoid misleading insights and improve model accuracy.

Interpreting Models

Feature EngineeringCheng-tao Chu highlights the importance of proper feature standardization and the challenges of multicollinearity in machine learning. His insights offer valuable guidance for avoiding common pitfalls that can lead to incorrect model predictions and interpretations.

Feature Engineering

Improving Loss FunctionsCheng-tao Chu, an entrepreneur and machine learning expert, discusses the critical role of loss functions in model evaluation and optimization. He highlights the importance of selecting the right loss function based on data characteristics and problem objectives.

Improving Loss Functions

Related Episodes