No abstract

No abstract

Understand and optimize industrial processes via machine learning and chemical engineering principles.

Traumatic brain injury is a leading cause of mortality worldwide, often affecting individuals at their most economically active yet no primary disease-modifying interventions exist for their treatment. Real-time direct spectroscopic examination of the brain tissue within the context of traumatic brain injury has the potential to improve the understanding of injury heterogeneity and subtypes, better target management strategies and organ penetrance of pharmacological agents, identify novel targets for intervention, and allow a clearer understanding of fundamental biochemistry evolution. Here, a novel device is designed and engineered, delivering Raman spectroscopy-based measurements from the brain through clinically established cranial access techniques. Device prototyping is undertaken within the constraints imposed by the acquisition and site dimensions (standard intracranial access holes, probe’s dimensions), and an artificial skull anatomical model with cortical impact is developed. The device shows a good agreement with the data acquired via a standard commercial Raman, and the spectra measured are comparable in terms of quality and detectable bands to the established traumatic brain injury model. The developed proof-of-concept device demonstrates the feasibility for real-time optical brain spectroscopic interface while removing the noise of extracranial tissue and with further optimization and in vivo validation, such technology will be directly translatable for integration into currently available standards of neurological care.

Reinforcement learning (RL) is a data-driven approach to synthesizing an optimal control policy. A barrier to wide implementation of RL-based controllers is its datahungry nature during online training and its inability to extract useful information from human operator and historical process operation data. Here, we present a twostep framework to resolve this challenge. First, we employ apprenticeship learning via inverse RL to analyze historical process data for synchronous identification of a reward function and parameterization of the control policy. This is conducted offline.Second, the parameterization is improved online efficiently under the ongoing process via RL within only a few iterations. Significant advantages of this framework include to allow for the hot-start of RL algorithms for process optimal control, and robust abstraction of existing controllers and control knowledge from data. The framework is demonstrated on three case studies, showing its potential for chemical process control. K E Y W O R D Sapprenticeship learning, inverse reinforcement learning, machine learning, optimal control, reinforcement learning | INTRODUCTIONRecent initiatives for efficiency improvements in industrial process operation has driven interest in the development of high performance, advanced process control (APC) schemes. Reinforcement learning (RL) has achieved impressive results on benchmark game-based control tasks, 1,2 providing an avenue for research in translation to APC. In