Uniting Statistical Testing and Machine Learning for Safe Predictions

Introduction

Recent breakthroughs in machine learning (ML) have brought about a transformative impact on decision-making, autonomous systems, medical diagnosis, and the creation of new scientific knowledge. However, this progress has a major drawback: modern predictive systems are extremely complex and hard to interpret, a problem known as the ‘black-box effect’.

Challenges of Black-Box Systems

The opaque nature of modern ML models, trained on increasingly diverse, incomplete, and noisy data, and later deployed in varying environments, hinders our ability to comprehend what drives inaccurate predictions, biased outcomes, and test-time failures.

Perhaps the most pressing question of our times is this: can we trust the predictions for future unseen instances obtained by black-box systems? The lack of practical guarantees on the limits of predictive performance poses a significant obstacle to deploying ML in applications that affect people's lives, opportunities, and science.

Goals and Objectives

My overarching goal is to put precise, interpretable, and robust error bounds on ML predictions, communicating rigorously what can be honestly inferred from data. I call for the development of protective ecosystems that can be seamlessly plugged into any ML model to monitor and guarantee its safety.

Methodology

This proposal introduces a unique interplay between statistics—the grammar of science—and ML—the art of learning from experience. Leveraging my expertise in both domains, I will show how statistical methodologies such as:

1. Conformal prediction
2. Test-martingales

can empower ML, and how recent breakthroughs in ML such as:

1. Semi-supervised learning
2. Domain adaptation technologies

can empower statistics.

Addressing Real-World Problems

I will tackle challenges rooted in real-world problems concerning:

1. Availability of training data
2. Quality of training data
3. Test-time drifting data

Expected Outcomes

A successful outcome would not only lead to a timely and rigorous way toward safe ML but may also significantly reform the way we develop, deploy, and interact with learning systems.