1 citations · 1 across the 7 of their papers we have counts for
8 papers
How to post-train on a surrogate: Envelope sampling mitigates reward hacking
Sanjit Dandapanthula, Shuvom Sadhuka, Samir Khan +3
Large language models (LLMs) are commonly post-trained against LLM judges and other cheap surrogates because the true reward, such as human preference, is too expensive to query at…
Are we really tilting? The mechanics of reward guidance in flow and diffusion models
Sanjit Dandapanthula, Nicholas M. Boffi
Reward guidance algorithms steer a learned generative process toward the reward-tilted measure at inference time. While empirically powerful, these methods are prone to reward hack…
Downscaling land surface temperature data using edge detection and block-diagonal Gaussian process regression
Sanjit Dandapanthula, Margaret Johnson, Madeleine Pascolini-Campbell +2
Accurate and high-resolution estimation of land surface temperature (LST) is crucial in estimating evapotranspiration, a measure of plant water use and a central quantity in agricu…
Optimal Transportation and Alignment Between Gaussian Measures
Sanjit Dandapanthula, Aleksandr Podkopaev, Shiva Prasad Kasiviswanathan +2
Optimal transport (OT) and Gromov-Wasserstein (GW) alignment provide interpretable geometric frameworks for comparing, transforming, and aggregating heterogeneous datasets---tasks…
Gradient descent dynamics for deep equilibrium models
Sanjit Dandapanthula, Aaditya Ramdas
Deep equilibrium models (DEQs) have recently emerged as a powerful paradigm for training infinitely deep weight-tied neural networks that achieve state of the art performance acros…
Offline changepoint localization using a matrix of conformal p-values
Sanjit Dandapanthula, Aaditya Ramdas
Changepoint localization is the problem of estimating the index at which a change occurred in the data generating distribution of an ordered list of data, or declaring that no chan…
Multiple testing in multi-stream sequential change detection
Sanjit Dandapanthula, Aaditya Ramdas
Multi-stream sequential change detection involves simultaneously monitoring many data streams and raising an alarm when their distributions change. We study the multiple testing is…
Anytime-valid FDR control with the stopped e-BH procedure
Hongjian Wang, Sanjit Dandapanthula, Aaditya Ramdas
The recent e-Benjamini-Hochberg (e-BH) procedure for multiple hypothesis testing is known to control the false discovery rate (FDR) under arbitrary dependence between the input e-v…