• Corpus ID: 235446566

Amortized Synthesis of Constrained Configurations Using a Differentiable Surrogate

  title={Amortized Synthesis of Constrained Configurations Using a Differentiable Surrogate},
  author={Xingyuan Sun and Tianju Xue and Szymon M. Rusinkiewicz and Ryan P. Adams},
In design, fabrication, and control problems, we are often faced with the task of synthesis, in which we must generate an object or configuration that satisfies a set of constraints while maximizing one or more objective functions. The synthesis problem is typically characterized by a physical process in which many different realizations may achieve the goal. This many-to-one map presents challenges to the supervised learning of feed-forward synthesis, as the set of viable designs may have a… 
More Stiffness with Less Fiber: End-to-End Fiber Path Optimization for 3D-Printed Composites
In 3D printing, stiff fibers ( e.g ., carbon fiber) can reinforce thermoplastic polymers with limited stiffness. However, existing commercial digital manufacturing software only provides a few simple
On Diverse System-Level Design Using Manifold Learning and Partial Simulated Annealing
This work uses deep generative models to learn a manifold of the valid design space, followed by Monte Carlo sampling to explore and optimize design over the learned manifold, producing a diverse set of optimal designs.
Physics-Integrated Variational Autoencoders for Robust and Interpretable Generative Modeling
This work introduces an architecture of variational autoencoders (VAEs) in which a part of the latent space is grounded by physics, and proposes a regularized learning method that controls the effect of the trainable components and preserves the semantics of the physics-based latent variables as intended.


PyTorch: An Imperative Style, High-Performance Deep Learning Library
This paper details the principles that drove the implementation of PyTorch and how they are reflected in its architecture, and explains how the careful and pragmatic implementation of the key components of its runtime enables them to work together to achieve compelling performance.
Amortized Finite Element Analysis for Fast PDE-Constrained Optimization
This paper proposes amortized finite element analysis (AmorFEA), in which a neural network learns to produce accurate PDE solutions, while preserving many of the advantages of traditional finite element methods.
Deep learning identifies synergistic drug combinations for treating COVID-19
A neural network architecture that jointly learns drug−target interaction and drug−drug synergy is proposed that is effective in silico search for synergistic combinations against SARS-CoV-2 and can be readily extended to other diseases for which a dearth of chemical−chemical combination data exists.
Conservative Objective Models for Effective Offline Model-Based Optimization
Conservative objective models (COMs) are proposed, a method that learns a model of the objective function which lower bounds the actual value of the ground-truth objective on outof-distribution inputs and uses it for optimization.
Synthesizing Modular Manipulators For Tasks With Time, Obstacle, And Torque Constraints
This work describes a framework that automatically synthesizes the design and controls for a serial chain modular manipulator given a task description and demonstrates the approach on a complex specification in which the robot navigates a constrained environment while holding an object.
Self-powered soft robot in the Mariana Trench.
This work develops an untethered soft robot for deep-sea exploration, with onboard power, control and actuation protected from pressure by integrating electronics in a silicone matrix, and validates the pressure resilience of the electronic components and soft actuators.
Offline Model-Based Optimization via Normalized Maximum Likelihood Estimation
This work proposes to tackle data-driven optimization problems where one must maximize a function given only queries at a fixed set of points by leveraging the normalized maximum-likelihood (NML) estimator, which provides a principled approach to handling uncertainty and out-of-distribution inputs.
Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classification
This work proposes a new margin-based surrogate loss function for the A UC score (named as the AUC margin loss) that is more robust than the commonly used AUC square loss, while enjoying the same advantage in terms of large-scale stochastic optimization.
An Optimal Architectural Design for Unconventional Modular Reconfigurable Manipulation System
An optimal architectural design for unconventional parameters for customized reconfigurability for non-repetitive applications is presented and a comparative study is presented based upon the payload to the manipulator weight ratio, involving other reported architectures.