Michael I. Jordan
YOU?
Author Swipe
View article: Stopping Rules for Stochastic Gradient Descent via Anytime-Valid Confidence Sequences
Stopping Rules for Stochastic Gradient Descent via Anytime-Valid Confidence Sequences Open
We study stopping rules for stochastic gradient descent (SGD) for convex optimization from the perspective of anytime-valid confidence sequences. Classical analyses of SGD provide convergence guarantees in expectation or at a fixed horizon…
View article: Structured Matrix Scaling for Multi-Class Calibration
Structured Matrix Scaling for Multi-Class Calibration Open
Post-hoc recalibration methods are widely used to ensure that classifiers provide faithful probability estimates. We argue that parametric recalibration functions based on logistic regression can be motivated from a simple theoretical sett…
View article: Cross-Validated Causal Inference: a Modern Method to Combine Experimental and Observational Data
Cross-Validated Causal Inference: a Modern Method to Combine Experimental and Observational Data Open
We develop new methods to integrate experimental and observational data in causal inference. While randomized controlled trials offer strong internal validity, they are often costly and therefore limited in sample size. Observational data,…
View article: Safety versus performance: How multi-objective learning reduces barriers to market entry
Safety versus performance: How multi-objective learning reduces barriers to market entry Open
Emerging marketplaces for large language models and other large-scale machine learning models appear to exhibit market concentration, which has raised concerns about whether there are insurmountable barriers to entry in such markets. In th…
View article: Deep generative modeling of sample-level heterogeneity in single-cell genomics
Deep generative modeling of sample-level heterogeneity in single-cell genomics Open
Single-cell genomic studies were recently conducted on hundred of samples exhibiting complex designs. These data have tremendous potential for discovering how sample- or tissue-level phenotypes relate to cellular and molecular composition.…
View article: Adaptive Coverage Policies in Conformal Prediction
Adaptive Coverage Policies in Conformal Prediction Open
Traditional conformal prediction methods construct prediction sets such that the true label falls within the set with a user-specified coverage level. However, poorly chosen coverage levels can result in uninformative predictions, either p…
View article: Decoding of image properties from single-trial visual evoked potentials recorded by ultra-high-density EEG
Decoding of image properties from single-trial visual evoked potentials recorded by ultra-high-density EEG Open
Visual evoked potentials (VEPs) recorded by encephalography (EEG) allow us to study the neuronal activity non-invasively and in high temporal resolution. Traditionally, EEG analyses have relied on univariate group-level statistics and tria…
View article: A Collectivist, Economic Perspective on AI
A Collectivist, Economic Perspective on AI Open
Information technology is in the midst of a revolution in which omnipresent data collection and machine learning are impacting the human world as never before. The word ``intelligence'' is being used as a North Star for the development of …
View article: Valid Selection among Conformal Sets
Valid Selection among Conformal Sets Open
Conformal prediction offers a distribution-free framework for constructing prediction sets with coverage guarantees. In practice, multiple valid conformal prediction sets may be available, arising from different models or methodologies. Ho…
View article: Sample Complexity and Representation Ability of Test-time Scaling Paradigms
Sample Complexity and Representation Ability of Test-time Scaling Paradigms Open
Test-time scaling paradigms have significantly advanced the capabilities of large language models (LLMs) on complex tasks. Despite their empirical success, theoretical understanding of the sample efficiency of various test-time strategies …
View article: Online Decision-Focused Learning
Online Decision-Focused Learning Open
Decision-focused learning (DFL) is an increasingly popular paradigm for training predictive models whose outputs are used in decision-making tasks. Instead of merely optimizing for predictive accuracy, DFL trains models to directly minimiz…
View article: Backward Conformal Prediction
Backward Conformal Prediction Open
We introduce $\textit{Backward Conformal Prediction}$, a method that guarantees conformal coverage while providing flexible control over the size of prediction sets. Unlike standard conformal prediction, which fixes the coverage level and …
View article: On Finding Local Nash Equilibria (and only Local Nash Equilibria) in Zero-Sum Games
On Finding Local Nash Equilibria (and only Local Nash Equilibria) in Zero-Sum Games Open
We propose local symplectic surgery , a two-timescale procedure for finding local Nash equilibria in two-player zero-sum games. We first show that previous gradient-based algorithms cannot guarantee convergence to local Nash equilibria due…
View article: Understanding In-context Learning of Addition via Activation Subspaces
Understanding In-context Learning of Addition via Activation Subspaces Open
To perform few-shot learning, language models extract signals from a few input-label pairs, aggregate these into a learned prediction rule, and apply this rule to new inputs. How is this implemented in the forward pass of modern transforme…
View article: A Quadratic Speedup in Finding Nash Equilibria of Quantum Zero-Sum Games
A Quadratic Speedup in Finding Nash Equilibria of Quantum Zero-Sum Games Open
Recent developments in domains such as non-local games, quantum interactive proofs, and quantum generative adversarial networks have renewed interest in quantum game theory and, specifically, quantum zero-sum games. Central to classical ga…
View article: Relying on the Metrics of Evaluated Agents
Relying on the Metrics of Evaluated Agents Open
View article: Accelerated first-order optimization under nonlinear constraints
Accelerated first-order optimization under nonlinear constraints Open
We exploit analogies between first-order algorithms for constrained optimization and non-smooth dynamical systems to design a new class of accelerated first-order algorithms for constrained optimization. Unlike Frank–Wolfe or projected gra…
View article: Stochastic Optimization with Optimal Importance Sampling
Stochastic Optimization with Optimal Importance Sampling Open
Importance Sampling (IS) is a widely used variance reduction technique for enhancing the efficiency of Monte Carlo methods, particularly in rare-event simulation and related applications. Despite its power, the performance of IS is often h…
View article: Universal Log-Optimality for General Classes of e-processes and Sequential Hypothesis Tests
Universal Log-Optimality for General Classes of e-processes and Sequential Hypothesis Tests Open
We consider the problem of sequential hypothesis testing by betting. For a general class of composite testing problems -- which include bounded mean testing, equal mean testing for bounded random tuples, and some key ingredients of two-sam…
View article: E-Values Expand the Scope of Conformal Prediction
E-Values Expand the Scope of Conformal Prediction Open
Conformal prediction is a powerful framework for distribution-free uncertainty quantification. The standard approach to conformal prediction relies on comparing the ranks of prediction scores: under exchangeability, the rank of a future te…
View article: Processing and optimization of laser-activated Cu and Ni inks for ceramic tile inkjet printing
Processing and optimization of laser-activated Cu and Ni inks for ceramic tile inkjet printing Open
View article: Marketplace Operators Can Induce Competitive Pricing
Marketplace Operators Can Induce Competitive Pricing Open
As e-commerce marketplaces continue to grow in popularity, it has become increasingly important to understand the role and impact of marketplace operators on competition and social welfare. We model a marketplace operator as an entity that…
View article: An Overview of Large Language Models for Statisticians
An Overview of Large Language Models for Statisticians Open
Large Language Models (LLMs) have emerged as transformative tools in artificial intelligence (AI), exhibiting remarkable capabilities across diverse tasks such as text generation, reasoning, and decision-making. While their success has pri…
View article: How Do LLMs Perform Two-Hop Reasoning in Context?
How Do LLMs Perform Two-Hop Reasoning in Context? Open
``Socrates is human. All humans are mortal. Therefore, Socrates is mortal.'' This form of argument illustrates a typical pattern of two-hop reasoning. Formally, two-hop reasoning refers to the process of inferring a conclusion by making tw…
View article: Conformal Prediction under Levy-Prokhorov Distribution Shifts: Robustness to Local and Global Perturbations
Conformal Prediction under Levy-Prokhorov Distribution Shifts: Robustness to Local and Global Perturbations Open
Conformal prediction provides a powerful framework for constructing prediction intervals with finite-sample guarantees, yet its robustness under distribution shifts remains a significant challenge. This paper addresses this limitation by m…
View article: Statistical Collusion by Collectives on Learning Platforms
Statistical Collusion by Collectives on Learning Platforms Open
As platforms increasingly rely on learning algorithms, collectives may form and seek ways to influence these platforms to align with their own interests. This can be achieved by coordinated submission of altered data. To evaluate the poten…
View article: Online Decision-Making in Tree-Like Multi-Agent Games with Transfers
Online Decision-Making in Tree-Like Multi-Agent Games with Transfers Open
The widespread deployment of Machine Learning systems everywhere raises challenges, such as dealing with interactions or competition between multiple learners. In that goal, we study multi-agent sequential decision-making by considering pr…
View article: Rethinking Early Stopping: Refine, Then Calibrate
Rethinking Early Stopping: Refine, Then Calibrate Open
Machine learning classifiers often produce probabilistic predictions that are critical for accurate and interpretable decision-making in various domains. The quality of these predictions is generally evaluated with proper losses, such as c…
View article: Prediction-Aware Learning in Multi-Agent Systems
Prediction-Aware Learning in Multi-Agent Systems Open
The framework of uncoupled online learning in multiplayer games has made significant progress in recent years. In particular, the development of time-varying games has considerably expanded its modeling capabilities. However, current regre…
View article: The Sample Complexity of Online Reinforcement Learning: A Multi-model Perspective
The Sample Complexity of Online Reinforcement Learning: A Multi-model Perspective Open
We study the sample complexity of online reinforcement learning in the general setting of nonlinear dynamical systems with continuous state and action spaces. Our analysis accommodates a large class of dynamical systems ranging from a fini…