-
Leveraging Contextual Relatedness to Identify Suicide Documentation in Clinical Notes through Zero Shot Learning
Authors:
Terri Elizabeth Workman,
Joseph L. Goulet,
Cynthia A. Brandt,
Allison R. Warren,
Jacob Eleazer,
Melissa Skanderson,
Luke Lindemann,
John R. Blosnich,
John O Leary,
Qing Zeng Treitler
Abstract:
Identifying suicidality including suicidal ideation, attempts, and risk factors in electronic health record data in clinical notes is difficult. A major difficulty is the lack of training samples given the small number of true positive instances among the increasingly large number of patients being screened. This paper describes a novel methodology that identifies suicidality in clinical notes by…
▽ More
Identifying suicidality including suicidal ideation, attempts, and risk factors in electronic health record data in clinical notes is difficult. A major difficulty is the lack of training samples given the small number of true positive instances among the increasingly large number of patients being screened. This paper describes a novel methodology that identifies suicidality in clinical notes by addressing this data sparsity issue through zero-shot learning. U.S. Veterans Affairs clinical notes served as data. The training dataset label was determined using diagnostic codes of suicide attempt and self-harm. A base string associated with the target label of suicidality was used to provide auxiliary information by narrowing the positive training cases to those containing the base string. A deep neural network was trained by map** the training documents contents to a semantic space. For comparison, we trained another deep neural network using the identical training dataset labels and bag-of-words features. The zero shot learning model outperformed the baseline model in terms of AUC, sensitivity, specificity, and positive predictive value at multiple probability thresholds. In applying a 0.90 probability threshold, the methodology identified notes not associated with a relevant ICD 10 CM code that documented suicidality, with 94 percent accuracy. This new method can effectively identify suicidality without requiring manual annotation.
△ Less
Submitted 9 January, 2023;
originally announced January 2023.
-
The K-Theoretic Formulation of D-Brane Aharonov-Bohm Phases
Authors:
Aaron R. Warren
Abstract:
The topological calculation of Aharonov-Bohm phases associated with D-branes in the absence of a Neveu-Schwarz B-field is explored. The K-theoretic classification of Ramond-Ramond fields in Type II and Type I theories is used to produce formulae for the Aharonov-Bohm phase associated with a torsion flux. A topological construction shows that K-theoretic pairings to calculate such phases exist and…
▽ More
The topological calculation of Aharonov-Bohm phases associated with D-branes in the absence of a Neveu-Schwarz B-field is explored. The K-theoretic classification of Ramond-Ramond fields in Type II and Type I theories is used to produce formulae for the Aharonov-Bohm phase associated with a torsion flux. A topological construction shows that K-theoretic pairings to calculate such phases exist and are well-defined. An analytic perspective is then taken, obtaining a means for determining Aharonov-Bohm phases by way of the reduced eta-invariant. This perspective is used to calculate the phase for an experiment involving the $(-1)-8$ system in Type I theory, and compared with previous calculations performed using different methods.
△ Less
Submitted 25 November, 2012;
originally announced November 2012.
-
Stellar Collisions and the Interior Structure of Blue Stragglers
Authors:
James C. Lombardi,
Jessica Sawyer Warren,
Frederic A. Rasio,
Alison Sills,
Aaron R. Warren
Abstract:
Collisions of main sequence stars occur frequently in dense star clusters. In open and globular clusters, these collisions produce merger remnants that may be observed as blue stragglers. Detailed theoretical models of this process require lengthy hydrodynamic computations in three dimensions. However, a less computationally expensive approach, which we present here, is to approximate the merger…
▽ More
Collisions of main sequence stars occur frequently in dense star clusters. In open and globular clusters, these collisions produce merger remnants that may be observed as blue stragglers. Detailed theoretical models of this process require lengthy hydrodynamic computations in three dimensions. However, a less computationally expensive approach, which we present here, is to approximate the merger process (including shock heating, hydrodynamic mixing, mass ejection, and angular momentum transfer) with simple algorithms based on conservation laws and a basic qualitative understanding of the hydrodynamics. These algorithms have been fine tuned through comparisons with the results of our previous hydrodynamic simulations. We find that the thermodynamic and chemical composition profiles of our simple models agree very well with those from recent SPH (smoothed particle hydrodynamics) calculations of stellar collisions, and the subsequent stellar evolution of our simple models also matches closely that of the more accurate hydrodynamic models. Our algorithms have been implemented in an easy to use software package, which we are making publicly available (see http://vassun.vassar.edu/~lombardi/mmas/). This software could be used in combination with realistic dynamical simulations of star clusters that must take into account stellar collisions.
△ Less
Submitted 2 November, 2001; v1 submitted 19 July, 2001;
originally announced July 2001.
