Laurence Illing Midgley

Laurence Illing Midgley, Vincent Stimper, Gregor N. C. Simm et al. · cambridge

Normalizing flows are tractable density models that can approximate complicated target distributions, e.g. Boltzmann distributions of physical systems. However, current methods for training flows either suffer from mode-seeking behavior, use samples from the target generated beforehand by expensive MCMC methods, or use stochastic losses that have high variance. To avoid these problems, we augment flows with annealed importance sampling (AIS) and minimize the mass-covering $α$-divergence with $α=2$, which minimizes importance weight variance. Our method, Flow AIS Bootstrap (FAB), uses AIS to generate samples in regions where the flow is a poor approximation of the target, facilitating the discovery of new modes. We apply FAB to multimodal targets and show that we can approximate them very accurately where previous methods fail. To the best of our knowledge, we are the first to learn the Boltzmann distribution of the alanine dipeptide molecule using only the unnormalized target density, without access to samples generated via Molecular Dynamics (MD) simulations: FAB produces better results than training via maximum likelihood on MD samples while using 100 times fewer target evaluations. After reweighting the samples, we obtain unbiased histograms of dihedral angles that are almost identical to the ground truth.

Clément Bonnet, Laurence Midgley, Alexandre Laterre

Meta-gradient Reinforcement Learning (RL) allows agents to self-tune their hyper-parameters in an online fashion during training. In this paper, we identify a bias in the meta-gradient of current meta-gradient RL approaches. This bias comes from using the critic that is trained using the meta-learned discount factor for the advantage estimation in the outer objective which requires a different discount factor. Because the meta-learned discount factor is typically lower than the one used in the outer objective, the resulting bias can cause the meta-gradient to favor myopic policies. We propose a simple solution to this issue: we eliminate this bias by using an alternative, \emph{outer} value function in the estimation of the outer loss. To obtain this outer value function we add a second head to the critic network and train it alongside the classic critic, using the outer loss discount factor. On an illustrative toy problem, we show that the bias can cause catastrophic failure of current meta-gradient RL approaches, and show that our proposed solution fixes it. We then apply our method to a more complex environment and demonstrate that fixing the meta-gradient bias can significantly improve performance.

Laurence Illing Midgley, Vincent Stimper, Gregor N. C. Simm et al.

Normalizing flows are flexible, parameterized distributions that can be used to approximate expectations from intractable distributions via importance sampling. However, current flow-based approaches are limited on challenging targets where they either suffer from mode seeking behaviour or high variance in the training loss, or rely on samples from the target distribution, which may not be available. To address these challenges, we combine flows with annealed importance sampling (AIS), while using the $α$-divergence as our objective, in a novel training procedure, FAB (Flow AIS Bootstrap). Thereby, the flow and AIS improve each other in a bootstrapping manner. We demonstrate that FAB can be used to produce accurate approximations to complex target distributions, including Boltzmann distributions, in problems where previous flow-based methods fail.

Laurence Illing Midgley

This paper demonstrates the application of reinforcement learning (RL) to process synthesis by presenting Distillation Gym, a set of RL environments in which an RL agent is tasked with designing a distillation train, given a user defined multi-component feed stream. Distillation Gym interfaces with a process simulator (COCO and ChemSep) to simulate the environment. A demonstration of two distillation problem examples are discussed in this paper (a Benzene, Toluene, P-xylene separation problem and a hydrocarbon separation problem), in which a deep RL agent is successfully able to learn within Distillation Gym to produce reasonable designs. Finally, this paper proposes the creation of Chemical Engineering Gym, an all-purpose reinforcement learning software toolkit for chemical engineering process synthesis.