10 papers
A Profile-Separation Framework for Quantitative Convergence of No-U-Turn Samplers
Krishnakumar Balasubramanian
We study multinomial and biased-progressive No-U-Turn Samplers for strongly log-concave targets satisfying and $\|\nabla^2U(x)-\nabla^2U(y)\…
Uniform-in-time Propagation-of-Chaos for Stein Variational Gradient Descent
Krishnakumar Balasubramanian, Sayan Banerjee, Anna Korba
We study uniform-in-time propagation-of-chaos for continuous-time Stein Variational Gradient Descent (SVGD). Classical finite-time propagation-of-chaos estimates for mean-field sys…
Total Variation Rates for Riemannian Flow Matching
Yunrui Guan, Krishnakumar Balasubramanian, Shiqian Ma
Riemannian flow matching (RFM) extends flow-based generative modeling to data supported on manifolds by learning a time-dependent tangent vector field whose flow-ODE transports a s…
Finite-Particle Rates for Regularized Stein Variational Gradient Descent
Ye He, Krishnakumar Balasubramanian, Sayan Banerjee +1
We derive finite-particle rates for the regularized Stein variational gradient descent (R-SVGD) algorithm introduced by He et al. (2024) that corrects the constant-order bias of th…
Dependence-Aware Label Aggregation for LLM-as-a-Judge via Ising Models
Krishnakumar Balasubramanian, Aleksandr Podkopaev, Shiva Prasad Kasiviswanathan
Large-scale AI evaluation increasingly relies on aggregating binary judgments from annotators, including LLMs used as judges. Most classical methods, e.g., Dawid-Skene or (weig…
On the Structure of Stationary Solutions to McKean-Vlasov Equations with Applications to Noisy Transformers
Krishnakumar Balasubramanian, Sayan Banerjee, Philippe Rigollet
We study stationary solutions of McKean-Vlasov equations on the circle. Our main contributions stem from observing an exact equivalence between solutions of the stationary McKean-V…
Statistical Inference for Linear Functionals of Online Least-squares SGD when
Bhavya Agrawalla, Krishnakumar Balasubramanian, Promit Ghosal
Stochastic Gradient Descent (SGD) has become a cornerstone method in modern data science. However, deploying SGD in high-stakes applications necessitates rigorous quantification of…
Restricted Spectral Gap Decomposition for Simulated Tempering Targeting Mixture Distributions
Jhanvi Garg, Krishna Balasubramanian, Quan Zhou
Simulated tempering is a widely used strategy for sampling from multimodal distributions. In this paper, we consider simulated tempering combined with an arbitrary local Markov cha…
Improved Finite-Particle Convergence Rates for Stein Variational Gradient Descent
Sayan Banerjee, Krishnakumar Balasubramanian, Promit Ghosal
We provide finite-particle convergence rates for the Stein Variational Gradient Descent (SVGD) algorithm in the Kernelized Stein Discrepancy () and Wasserstein-2 metr…
Minimax Optimal Goodness-of-Fit Testing with Kernel Stein Discrepancy
Omar Hagrass, Bharath Sriperumbudur, Krishnakumar Balasubramanian
We explore the minimax optimality of goodness-of-fit tests on general domains using the kernelized Stein discrepancy (KSD). The KSD framework offers a flexible approach for goodnes…