arXiv subjects

Shenbo Xu

Publications and source records attributed to Shenbo Xu.

3 recordsLinked to original sources

arXiv2025

Finding the Minimal Parameter Budget for Implicit Reasoning: A Data Complexity Driven Scaling Law for Language Models

Reasoning is a core capability of language models (LMs), yet it remains unclear how much model capacity is necessary to support reasoning during pretraining. In this work, we study the minimal parameter budget required for implicit reasoning, defined as the ability to infer new facts from learned knowledge without explicit chain-of-thought supervision. To isolate this phenomenon, we pretrain LMs from scratch in a controlled synthetic environment that mimics the structure and distribution of real-world knowledge graphs, and evaluate their ability to complete missing edges via multi-hop inference. From both a theoretical and an empirical perspective, we identify a scaling law linking this optimal parameter budget to a graph search entropy measure. Across a wide range of model sizes, training steps, and graph complexities, we show that an optimally sized language model can reliably reason over approximately 0.008 bits of information per parameter at most. Our results characterize the minimal sufficient capacity for implicit reasoning during pretraining. Our findings provide principled guidance for matching model size to data complexity and offer new insights into the scaling behavior of reasoning in large language models.

cs.AI↗

arXiv2024

Estimating Heterogeneous Treatment Effects on Survival Outcomes Using Counterfactual Censoring Unbiased Transformations

Methods for estimating heterogeneous treatment effects (HTE) from observational data have largely focused on continuous or binary outcomes, with less attention paid to survival outcomes and almost none to settings with competing risks. In this work, we develop censoring unbiased transformations (CUTs) for survival outcomes both with and without competing risks. After converting time-to-event outcomes using these CUTs, direct application of HTE learners for continuous outcomes yields consistent estimates of heterogeneous cumulative incidence effects, total effects, and separable direct effects. Our CUTs enable application of a much larger set of state of the art HTE learners for censored outcomes than had previously been available, especially in competing risks settings. We provide generic model-free learner-specific oracle inequalities bounding the finite-sample excess risk. The oracle efficiency results depend on the oracle selector and estimated nuisance functions from all steps involved in the transformation. We demonstrate the empirical performance of the proposed methods in simulation studies.

stat.ME↗

arXiv2023

Efficient estimation of weighted cumulative treatment effects by double/debiased machine learning

In empirical studies with time-to-event outcomes, investigators often leverage observational data to conduct causal inference on the effect of exposure when randomized controlled trial data is unavailable. Model misspecification and lack of overlap are common issues in observational studies, and they often lead to inconsistent and inefficient estimators of the average treatment effect. Estimators targeting overlap weighted effects have been proposed to address the challenge of poor overlap, and methods enabling flexible machine learning for nuisance models address model misspecification. However, the approaches that allow machine learning for nuisance models have not been extended to the setting of weighted average treatment effects for time-to-event outcomes when there is poor overlap. In this work, we propose a class of one-step cross-fitted double/debiased machine learning estimators for the weighted cumulative causal effect as a function of restriction time. We prove that the proposed estimators are consistent, asymptotically linear, and reach semiparametric efficiency bounds under regularity conditions. Our simulations show that the proposed estimators using nonparametric machine learning nuisance models perform as well as established methods that require correctly-specified parametric nuisance models, illustrating that our estimators mitigate the need for oracle parametric nuisance models. We apply the proposed methods to real-world observational data from a UK primary care database to compare the effects of anti-diabetic drugs on cancer clinical outcomes.

stat.ME↗