Search for: All records

Abstract Digital health data are multimodal and high-dimensional. A patient’s health state can be characterized by a multitude of signals including medical imaging, clinical variables, genome sequencing, conversations between clinicians and patients, and continuous signals from wearables, among others. This high volume, personalized data stream aggregated over patients’ lives has spurred interest in developing new artificial intelligence (AI) models for higher-precision diagnosis, prognosis, and tracking. While the promise of these algorithms is undeniable, their dissemination and adoption have been slow, owing partially to unpredictable AI model performance once deployed in the real world. We posit that one of the rate-limiting factors in developing algorithms that generalize to real-world scenarios is the very attribute that makes the data exciting—their high-dimensional nature. This paper considers how the large number of features in vast digital health data can challenge the development of robust AI models—a phenomenon known as “the curse of dimensionality” in statistical learning theory. We provide an overview of the curse of dimensionality in the context of digital health, demonstrate how it can negatively impact out-of-sample performance, and highlight important considerations for researchers and algorithm designers. 

Though it is often taken as a truism that communication contributes to organizational productivity, there are surprisingly few empirical studies documenting a relationship between observable interaction and productivity. This is because comprehensive, direct observation of communication in organizational settings is notoriously difficult. In this paper, we report a method for extracting network and speech characteristics data from audio recordings of participants talking with each other in real time. We use this method to analyze communication and productivity data from seventy-nine employees working within a software engineering organization who had their speech recorded during working hours for a period of approximately 3 years. From the speech data, we infer when any two individuals are talking to each other and use this information to construct a communication graph for the organization for each week. We use the spectral and temporal characteristics of the produced speech and the structure of the resultant communication graphs to predict the productivity of the group, as measured by the number of lines of code produced. The results indicate that the most important speech and network features for predicting productivity include those that measure the number of unique people interacting within the organization, the frequency of interactions, and the topology of the communication network.