Neural networks present a useful framework for learning complex dynamics, and are increasingly being considered as components to closed loop predictive control algorithms. However, if they are to be utilized in such safety-critical advisory settings, they must be provably "conformant" to the governing scientific (biological, chemical, physical) laws which underlie the modeled process. Unfortunately, this is not easily guaranteed as neural network models are prone to learn patterns which are artifacts of the conditions under which the training data is collected, which may not necessarily conform to underlying physiological laws.

Recent years have seen a rise in social media platforms that provide peer-to-peer support to individuals suffering from mental distress. Studies on the impact of these platforms have focused on either short-term scales of single-post threads, or long-term changes over arbitrary period of time (months or years). While important, such periods of time do not necessarily follow users’ progressions through acute periods of distress. Using data from Talklife, a mental health platform, we find that user activity follows a distinct pattern of high activity periods with interleaving periods of no activity, and propose a method for identifying such bursts & breaks in activity. We then show how studying activity during bursts can provide a personalized, medium-term analysis for a key question in online mental health communities: What characteristics of user activity lead some users to find support and help, while others fall short? Using two independent outcome metrics, moments of cognitive change and self-reported changes in mood during a burst of activity, we identify two actionable features that can improve outcomes for users: persistence within bursts, and giving complex emotional support to others. Our results demonstrate the value of considering bursts as a natural unit of analysis for psychosocial change in online mental health communities.

Neural networks are increasingly used as data-driven models for a wide variety of physical systems such as ground vehicles, airplanes, human physiology and automobile engines. These models are in-turn used for designing and verifying autonomous systems. The advantages of using neural networks include the ability to capture characteristics of particular systems using the available data. This is particularly advantageous for medical systems, wherein the data collected from individuals can be used to design devices that are well-adapted to a particular individual's unique physiological characteristics. At the same time, neural network models remain opaque: their structure makes them hard to understand and interpret by human developers. One key challenge lies in checking that neural network models of processes are “conformant” to the well established scientific (physical, chemical and biological) laws that underlie these models. In this paper, we will show how conformance often fails in models that are otherwise accurate and trained using the best practices in machine learning, with potentially serious consequences. We motivate the need for learning and verifying key conformance properties in data-driven models of the human insulin-glucose system and data-driven automobile models. We survey verification approaches for neural networks that can hold the key to learning and verifying conformance.