Understanding Is a Process.

How do we gauge understanding? Tests of understanding, such as Turing's imitation game, are numerous; yet, attempts to achieve a state of understanding are not satisfactory assessments. Intelligent agents designed to pass one test of understanding often fall short of others. Rather than approaching understanding as a system state, in this paper, we argue that understanding is a process that changes over time and experience. The only window into the process is through the lens of natural language. Usefully, failures of understanding reveal breakdowns in the process. We propose a set of natural language-based probes that can be used to map the degree of understanding a human or intelligent system has achieved through combinations of successes and failures.

Visual perception and mixed-initiative interaction for assisted visualization design.

This paper describes the integration of perceptual guidelines from human vision with an AI-based mixed-initiative search strategy. The result is a visualization assistant called ViA, a system that collaborates with its users to identify perceptually salient visualizations for large, multidimensional datasets. ViA applies knowledge of low-level human vision to: (1) evaluate the effectiveness of a particular visualization for a given dataset and analysis tasks; and (2) rapidly direct its search towards new visualizations that are most likely to offer improvements over those seen to date. Context, domain expertise, and a high-level understanding of a dataset are critical to identifying effective visualizations. We apply a mixed-initiative strategy that allows ViA and its users to share their different strengths and continually improve ViA's understanding of a user's preferences. We visualize historical weather conditions to compare ViA's search strategy to exhaustive analysis, simulated annealing, and reactive tabu search, and to measure the improvement provided by mixed-initiative interaction. We also visualize intelligent agents competing in a simulated online auction to evaluate ViA's perceptual guidelines. Results from each study are positive, suggesting that ViA can construct high-quality visualizations for a range of real-world datasets.

Utility of a genomic-based, personalized medicine test in patients presenting with symptoms suggesting coronary artery disease.

PURPOSE: Better methods are needed to assess patients presenting with symptoms suggestive of obstructive coronary artery disease (CAD). We hypothesized that the use of a gene expression score (GES) would lead to a change in the diagnostic evaluation. METHODS: The Primary Care Providers Use of a Gene Expression Test in Coronary Artery Disease Diagnosis (IMPACT-PCP) trial (clinical trial identifier NCT01594411, clinicaltrials.gov) was a prospective study of stable, nonacute, nondiabetic patients presenting with chest pain and related symptoms at 4 primary care practices. All patients underwent GES testing, with clinicians documenting their planned diagnostic strategy both before and after GES. The GES was derived from a peripheral blood draw measuring expression of 23 genes and has been shown to have a 96% negative predictive value for excluding the diagnosis of obstructive CAD. RESULTS: Of the 251 study patients, 140 were women (56%); the participants had a mean age of 56 years (standard deviation, 13.0) and a mean body mass index of 30 mg/kg(2) (standard deviation, 6.7). The mean GES was 16 (range, 1-38), and 127 patients (51%) had a low GES ([ltqeu]15). A change in the diagnostic testing pattern before and after GES testing was noted in 145 of 251 patients (58% observed vs. 10% predefined expected change; P < .001). CONCLUSIONS: Incorporation of the GES into the diagnostic workup showed clinical utility above and beyond conventional clinical factors by optimizing the patient's diagnostic evaluation.