Search Results for author: Biswadip Dey

Found 18 papers, 4 papers with code

Physics-Informed Machine Learning for Modeling and Control of Dynamical Systems

no code implementations24 Jun 2023 Truong X. Nghiem, Ján Drgoňa, Colin Jones, Zoltan Nagy, Roland Schwan, Biswadip Dey, Ankush Chakrabarty, Stefano Di Cairano, Joel A. Paulson, Andrea Carron, Melanie N. Zeilinger, Wenceslao Shaw Cortez, Draguna L. Vrabie

Specifically, the paper covers an overview of the theory, fundamental concepts and methods, tools, and applications on topics of: 1) physics-informed learning for system identification; 2) physics-informed learning for control; 3) analysis and verification of PIML models; and 4) physics-informed digital twins.

Physics-informed machine learning

RANS-PINN based Simulation Surrogates for Predicting Turbulent Flows

no code implementations9 Jun 2023 Shinjan Ghosh, Amit Chakraborty, Georgia Olympia Brikis, Biswadip Dey

Physics-informed neural networks (PINNs) provide a framework to build surrogate models for dynamical systems governed by differential equations.

Improving Gradient Computation for Differentiable Physics Simulation with Contacts

1 code implementation28 Apr 2023 Yaofeng Desmond Zhong, Jiequn Han, Biswadip Dey, Georgia Olympia Brikis

We find that existing differentiable simulation methods provide inaccurate gradients when the contact normal direction is not fixed - a general situation when the contacts are between two moving objects.

A Neural ODE Interpretation of Transformer Layers

no code implementations12 Dec 2022 Yaofeng Desmond Zhong, Tongtao Zhang, Amit Chakraborty, Biswadip Dey

Our experiments show that this simple modification improves the performance of transformer networks in multiple tasks.

Image Classification Numerical Integration

Flocks, Games, and Cognition: A Geometric Approach

no code implementations7 Aug 2022 Udit Halder, Vidya Raju, Matteo Mischiati, Biswadip Dey, P. S. Krishnaprasad

These behaviors may be viewed as flock-scale strategies, emerging from interactions between individuals, accomplishing some collective adaptive purpose such as finding a roost, or mitigating the danger from predator attacks.

Demystifying the Data Need of ML-surrogates for CFD Simulations

no code implementations5 May 2022 Tongtao Zhang, Biswadip Dey, Krishna Veeraraghavan, Harshad Kulkarni, Amit Chakraborty

Computational fluid dynamics (CFD) simulations, a critical tool in various engineering applications, often require significant time and compute power to predict flow properties.

Physics-informed neural networks for modeling rate- and temperature-dependent plasticity

no code implementations20 Jan 2022 Rajat Arora, Pratik Kakkar, Biswadip Dey, Amit Chakraborty

This work presents a physics-informed neural network (PINN) based framework to model the strain-rate and temperature dependence of the deformation fields in elastic-viscoplastic solids.

A Decentralized Reinforcement Learning Framework for Efficient Passage of Emergency Vehicles

no code implementations30 Oct 2021 Haoran Su, Yaofeng Desmond Zhong, Biswadip Dey, Amit Chakraborty

Consequently, the decentralized RL agents learn network-level cooperative traffic signal phase strategies that reduce EMV travel time and the average travel time of non-EMVs in the network.

reinforcement-learning Reinforcement Learning (RL)

EMVLight: A Decentralized Reinforcement Learning Framework for Efficient Passage of Emergency Vehicles

no code implementations12 Sep 2021 Haoran Su, Yaofeng Desmond Zhong, Biswadip Dey, Amit Chakraborty

EMVLight extends Dijkstra's algorithm to efficiently update the optimal route for the EMVs in real time as it travels through the traffic network.

reinforcement-learning Reinforcement Learning (RL)

Extending Lagrangian and Hamiltonian Neural Networks with Differentiable Contact Models

1 code implementation NeurIPS 2021 Yaofeng Desmond Zhong, Biswadip Dey, Amit Chakraborty

In this paper, we introduce a differentiable contact model, which can capture contact mechanics: frictionless/frictional, as well as elastic/inelastic.

Contact mechanics Friction +1

Benchmarking Energy-Conserving Neural Networks for Learning Dynamics from Data

no code implementations3 Dec 2020 Yaofeng Desmond Zhong, Biswadip Dey, Amit Chakraborty

The last few years have witnessed an increased interest in incorporating physics-informed inductive bias in deep learning frameworks.

Benchmarking Inductive Bias +2

On Using Hamiltonian Monte Carlo Sampling for Reinforcement Learning Problems in High-dimension

no code implementations11 Nov 2020 Udari Madhushani, Biswadip Dey, Naomi Ehrich Leonard, Amit Chakraborty

Value function based reinforcement learning (RL) algorithms, for example, $Q$-learning, learn optimal policies from datasets of actions, rewards, and state transitions.

Matrix Completion Q-Learning +2

Frequency-compensated PINNs for Fluid-dynamic Design Problems

no code implementations3 Nov 2020 Tongtao Zhang, Biswadip Dey, Pratik Kakkar, Arindam Dasgupta, Amit Chakraborty

We demonstrate this approach by predicting simulation results over out of range time interval and for novel design conditions.

Dissipative SymODEN: Encoding Hamiltonian Dynamics with Dissipation and Control into Deep Learning

1 code implementation ICLR Workshop DeepDiffEq 2019 Yaofeng Desmond Zhong, Biswadip Dey, Amit Chakraborty

In this work, we introduce Dissipative SymODEN, a deep learning architecture which can infer the dynamics of a physical system with dissipation from observed state trajectories.

A Conditional Generative Model for Predicting Material Microstructures from Processing Methods

no code implementations4 Oct 2019 Akshay Iyer, Biswadip Dey, Arindam Dasgupta, Wei Chen, Amit Chakraborty

Microstructures of a material form the bridge linking processing conditions - which can be controlled, to the material property - which is the primary interest in engineering applications.

Feature Engineering Image Generation

Symplectic ODE-Net: Learning Hamiltonian Dynamics with Control

1 code implementation ICLR 2020 Yaofeng Desmond Zhong, Biswadip Dey, Amit Chakraborty

In this paper, we introduce Symplectic ODE-Net (SymODEN), a deep learning framework which can infer the dynamics of a physical system, given by an ordinary differential equation (ODE), from observed state trajectories.

Inductive Bias

A graph-theoretic approach to multitasking

no code implementations NeurIPS 2017 Noga Alon, Daniel Reichman, Igor Shinkar, Tal Wagner, Sebastian Musslick, Jonathan D. Cohen, Tom Griffiths, Biswadip Dey, Kayhan Ozcimder

A key feature of neural network architectures is their ability to support the simultaneous interaction among large numbers of units in the learning and processing of representations.

Cannot find the paper you are looking for? You can Submit a new open access paper.