High-molecular-weight polysaccharide microgels were synthesized at a range of temperatures above the polymer lower critical solution temperature (LCST) (ΔT = 0.6–18.1 °C), and ΔT was found to strongly influence the structure, dynamics, and volume phase transition temperature of the resulting particles. Static and dynamic light scattering studies and mean field theory analysis of the microgels below and above volume phase transition revealed several distinct regimes. At small ΔT, lower density, larger, and more polydisperse microgels that deswell by a factor 10 in volume were synthesized. At intermediate ΔT = 5–8 °C, the formed microgels were the smallest, densest, and most monodisperse below the transition, but exhibited deswelling only by a factor of 5 above the transition. Synthesis at high ΔT led to the formation of nonuniform microgels with small density, a high degree of polydispersity, and in some cases the apparent presence of the un-cross-linked polymer. Furthermore, the volume phase transition temperature dropped significantly as ΔT increased. This work suggests that synthesis temperature can be used to tune the size, deswelling capacity, and volume phase transition temperature of the polymeric microgels.

No abstract

No abstract

In reinforcement learning (RL), the ability to utilize prior knowledge from previously solved tasks can allow agents to quickly solve new problems. In some cases, these new problems may be approximately solved by composing the solutions of previously solved primitive tasks (task composition). Otherwise, prior knowledge can be used to adjust the reward function for a new problem, in a way that leaves the optimal policy unchanged but enables quicker learning (reward shaping). In this work, we develop a general framework for reward shaping and task composition in entropy-regularized RL. To do so, we derive an exact relation connecting the optimal soft value functions for two entropy-regularized RL problems with different reward functions and dynamics. We show how the derived relation leads to a general result for reward shaping in entropy-regularized RL. We then generalize this approach to derive an exact relation connecting optimal value functions for the composition of multiple tasks in entropy-regularized RL. We validate these theoretical contributions with experiments showing that reward shaping and task composition lead to faster learning in various settings.

This report discusses the application of neural networks (NNs) as small segments of the brain. The networks representing the biological connectome are altered both spatially and temporally. The degradation techniques applied here are "weight degradation", "weight scrambling", and variable activation function. These methods aim to shine light on the study of neurodegenerative diseases such as Alzheimer's, Huntington's and Parkinson's disease as well as strokes and brain tumors disrupting the flow of information in the brain's network. Fundamental insights to memory loss and generalized learning dysfunction are gained by monitoring the network's error function during network degradation. The biological significance of each facet is also discussed.