Phy-Taylor: Partially Physics-Knowledge-Enhanced Deep Neural Networks via NN Editing.

Purely data-driven deep neural networks (DNNs) applied to physical engineering systems can infer relations that violate physics laws, thus leading to unexpected consequences. To address this challenge, we propose a physics-knowledge-enhanced DNN framework called Phy-Taylor, accelerating learning-compliant representations with physics knowledge. The Phy-Taylor framework makes two key contributions; it introduces a new architectural physics-compatible neural network (PhN) and features a novel compliance mechanism, which we call physics-guided neural network (NN) editing. The PhN aims to directly capture nonlinear physical quantities, such as kinetic energy, electrical power, and aerodynamic drag force. To do so, the PhN augments NN layers with two key components: 1) monomials of the Taylor series for capturing physical quantities and 2) a suppressor for mitigating the influence of noise. The NN editing mechanism further modifies network links and activation functions consistently with physics knowledge. As an extension, we also propose a self-correcting Phy-Taylor framework for safety-critical control of autonomous systems, which introduces two additional capabilities: 1) safety relationship learning and 2) automatic output correction when safety violations occur. Through experiments, we show that Phy-Taylor features considerably fewer parameters and a remarkably accelerated training process while offering enhanced model robustness and accuracy.

ControlVAE: Tuning, Analytical Properties, and Performance Analysis.

This paper reviews the novel concept of a controllable variational autoencoder (ControlVAE), discusses its parameter tuning to meet application needs, derives its key analytic properties, and offers useful extensions and applications. ControlVAE is a new variational autoencoder (VAE) framework that combines automatic control theory with the basic VAE to stabilize the KL-divergence of VAE models to a specified value. It leverages a non-linear PI controller, a variant of the proportional-integral-derivative (PID) controller, to dynamically tune the weight of the KL-divergence term in the evidence lower bound (ELBO) using the output KL-divergence as feedback. This allows us to precisely control the KL-divergence to a desired value (set point) that is effective in avoiding posterior collapse and learning disentangled representations. While prior work developed alternative techniques for controlling the KL divergence, we show that our PI controller has better stability properties and thus better convergence, thereby producing better disentangled representations from finite training data. In order to improve the ELBO of ControlVAE over that of the regular VAE, we provide a simplified theoretical analysis to inform the choice of set point for the KL-divergence of ControlVAE. We evaluate the proposed method on three tasks: image generation, language modeling, and disentangled representation learning. The results show that ControlVAE can achieve much better reconstruction quality than the other methods for comparable disentanglement. On the language modeling task, our method can avoid posterior collapse (KL vanishing) and improve the diversity of generated text. Moreover, it can change the optimization trajectory, improving the ELBO and the reconstruction quality for image generation.

Computational Modeling of Hierarchically Polarized Groups by Structured Matrix Factorization.

The paper extends earlier work on modeling hierarchically polarized groups on social media. An algorithm is described that 1) detects points of agreement and disagreement between groups, and 2) divides them hierarchically to represent nested patterns of agreement and disagreement given a structural guide. For example, two opposing parties might disagree on core issues. Moreover, within a party, despite agreement on fundamentals, disagreement might occur on further details. We call such scenarios hierarchically polarized groups. An (enhanced) unsupervised Non-negative Matrix Factorization (NMF) algorithm is described for computational modeling of hierarchically polarized groups. It is enhanced with a language model, and with a proof of orthogonality of factorized components. We evaluate it on both synthetic and real-world datasets, demonstrating ability to hierarchically decompose overlapping beliefs. In the case where polarization is flat, we compare it to prior art and show that it outperforms state of the art approaches for polarization detection and stance separation. An ablation study further illustrates the value of individual components, including new enhancements.

Human-centric sensing.

The first decade of the century witnessed a proliferation of devices with sensing and communication capabilities in the possession of the average individual. Examples range from camera phones and wireless global positioning system units to sensor-equipped, networked fitness devices and entertainment platforms (such as Wii). Social networking platforms emerged, such as Twitter, that allow sharing information in real time. The unprecedented deployment scale of such sensors and connectivity options ushers in an era of novel data-driven applications that rely on inputs collected by networks of humans or measured by sensors acting on their behalf. These applications will impact domains as diverse as health, transportation, energy, disaster recovery, intelligence and warfare. This paper surveys the important opportunities in human-centric sensing, identifies challenges brought about by such opportunities and describes emerging solutions to these challenges.