Sebastien Martin

Xabi Azagirre, Akshay Balwally, Guillaume Candeli et al.

To better match drivers to riders in our ridesharing application, we revised Lyft's core matching algorithm. We use a novel online reinforcement learning approach that estimates the future earnings of drivers in real time and use this information to find more efficient matches. This change was the first documented implementation of a ridesharing matching algorithm that can learn and improve in real time. We evaluated the new approach during weeks of switchback experimentation in most Lyft markets, and estimated how it benefited drivers, riders, and the platform. In particular, it enabled our drivers to serve millions of additional riders each year, leading to more than $30 million per year in incremental revenue. Lyft rolled out the algorithm globally in 2021.

Sebastien Martin

Electrical impedance tomography (EIT) is a promising technique for biomedical imaging. The strength of EIT is its ability to reconstruct images of the body's internal structures through radiation-safe techniques. EIT is regarded as safe for patients' health, and it is currently being actively researched. This paper investigates the application of EIT during deep brain stimulation (DBS) surgery as a means to identify targets during operations. DBS involves a surgical procedure in which a lead or electrode array is implanted in a specific target area in the brain. Electrical stimulations are then used to modulate neural circuits within the target area to reduce disabling neurological symptoms. The main difficulty in performing DBS surgery is to accurately position the lead in the target area before commencing the treatment. Brain tissue shifts during DBS surgery can be as large as the target size when compared with the pre-operative magnetic resonance imaging (MRI) or computed tomography (CT) images. To address this problem, a solution based on open-domain EIT to reconstruct images surrounding the probe during DBS surgery is proposed. Data acquisition and image reconstruction were performed, and artificial intelligence was applied to enhance the resulting images. The results showed that the proposed method is rapid, produces valuable high-quality images, and constitutes a first step towards in-vivo study.

Dimitris Bertsimas, Arthur Delarue, Patrick Jaillet et al.

When predictive models are used to support complex and important decisions, the ability to explain a model's reasoning can increase trust, expose hidden biases, and reduce vulnerability to adversarial attacks. However, attempts at interpreting models are often ad hoc and application-specific, and the concept of interpretability itself is not well-defined. We propose a general optimization framework to create explanations for linear models. Our methodology decomposes a linear model into a sequence of models of increasing complexity using coordinate updates on the coefficients. Computing this decomposition optimally is a difficult optimization problem for which we propose exact algorithms and scalable heuristics. By solving this problem, we can derive a parametrized family of interpretability metrics for linear models that generalizes typical proxies, and study the tradeoff between interpretability and predictive accuracy.

Dimitris Bertsimas, Arthur Delarue, Patrick Jaillet et al.

When quantitative models are used to support decision-making on complex and important topics, understanding a model's ``reasoning'' can increase trust in its predictions, expose hidden biases, or reduce vulnerability to adversarial attacks. However, the concept of interpretability remains loosely defined and application-specific. In this paper, we introduce a mathematical framework in which machine learning models are constructed in a sequence of interpretable steps. We show that for a variety of models, a natural choice of interpretable steps recovers standard interpretability proxies (e.g., sparsity in linear models). We then generalize these proxies to yield a parametrized family of consistent measures of model interpretability. This formal definition allows us to quantify the ``price'' of interpretability, i.e., the tradeoff with predictive accuracy. We demonstrate practical algorithms to apply our framework on real and synthetic datasets.