Browsing by Subject "cs.AI"

Now showing 1 - 4 of 4
  • Thumbnail Image
    ItemOpen Access
    Hölder Bounds for Sensitivity Analysis in Causal Reasoning.
    (CoRR, 2021) Assaad, Serge; Zeng, Shuxi; Pfister, Henry; Li, Fan; Carin, Lawrence
    We examine interval estimation of the effect of a treatment T on an outcome Y given the existence of an unobserved confounder U. Using H\"older's inequality, we derive a set of bounds on the confounding bias |E[Y|T=t]-E[Y|do(T=t)]| based on the degree of unmeasured confounding (i.e., the strength of the connection U->T, and the strength of U->Y). These bounds are tight either when U is independent of T or when U is independent of Y given T (when there is no unobserved confounding). We focus on a special case of this bound depending on the total variation distance between the distributions p(U) and p(U|T=t), as well as the maximum (over all possible values of U) deviation of the conditional expected outcome E[Y|U=u,T=t] from the average expected outcome E[Y|T=t]. We discuss possible calibration strategies for this bound to get interval estimates for treatment effects, and experimentally validate the bound using synthetic and semi-synthetic datasets.
  • Thumbnail Image
    ItemOpen Access
    On the Ontological Modeling of Trees
    (2017-10-24) Carral, David; Hitzler, Pascal; Lapp, Hilmar; Rudolph, Sebastian
    Trees -- i.e., the type of data structure known under this name -- are central to many aspects of knowledge organization. We investigate some central design choices concerning the ontological modeling of such trees. In particular, we consider the limits of what is expressible in the Web Ontology Language, and provide a reusable ontology design pattern for trees.
  • Thumbnail Image
    ItemOpen Access
    Thyroid Cancer Malignancy Prediction From Whole Slide Cytopathology Images
    Dov, David; Kovalsky, Shahar; Cohen, Jonathan; Range, Danielle; Henao, Ricardo; Carin, Lawrence
    We consider preoperative prediction of thyroid cancer based on ultra-high-resolution whole-slide cytopathology images. Inspired by how human experts perform diagnosis, our approach first identifies and classifies diagnostic image regions containing informative thyroid cells, which only comprise a tiny fraction of the entire image. These local estimates are then aggregated into a single prediction of thyroid malignancy. Several unique characteristics of thyroid cytopathology guide our deep-learning-based approach. While our method is closely related to multiple-instance learning, it deviates from these methods by using a supervised procedure to extract diagnostically relevant regions. Moreover, we propose to simultaneously predict thyroid malignancy, as well as a diagnostic score assigned by a human expert, which further allows us to devise an improved training strategy. Experimental results show that the proposed algorithm achieves performance comparable to human experts, and demonstrate the potential of using the algorithm for screening and as an assistive tool for the improved diagnosis of indeterminate cases.
  • Thumbnail Image
    ItemOpen Access
    Unachievable Region in Precision-Recall Space and Its Effect on Empirical Evaluation.
    (Proceedings of the ... International Conference on Machine Learning. International Conference on Machine Learning, 2012-12) Boyd, Kendrick; Santos Costa, Vítor; Davis, Jesse; Page, C David
    Precision-recall (PR) curves and the areas under them are widely used to summarize machine learning results, especially for data sets exhibiting class skew. They are often used analogously to ROC curves and the area under ROC curves. It is known that PR curves vary as class skew changes. What was not recognized before this paper is that there is a region of PR space that is completely unachievable, and the size of this region depends only on the skew. This paper precisely characterizes the size of that region and discusses its implications for empirical evaluation methodology in machine learning.