REVIEW 31 cited by
DiffTaichi: Differentiable Programming for Physical Simulation
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
DiffTaichi: Differentiable Programming for Physical Simulation
read the original abstract
We present DiffTaichi, a new differentiable programming language tailored for building high-performance differentiable physical simulators. Based on an imperative programming language, DiffTaichi generates gradients of simulation steps using source code transformations that preserve arithmetic intensity and parallelism. A light-weight tape is used to record the whole simulation program structure and replay the gradient kernels in a reversed order, for end-to-end backpropagation. We demonstrate the performance and productivity of our language in gradient-based learning and optimization tasks on 10 different physical simulators. For example, a differentiable elastic object simulator written in our language is 4.2x shorter than the hand-engineered CUDA version yet runs as fast, and is 188x faster than the TensorFlow implementation. Using our differentiable programs, neural network controllers are typically optimized within only tens of iterations.
Forward citations
Cited by 31 Pith papers
-
Differentiable Approximations for Distance Queries
A (1+ε)-approximate Euclidean distance function that is differentiable and returns gradients, using O(n/ε^(d/2)) space and O(log(n/ε)) query time.
-
Microcosmos: Reimagining Artificial Life for the GPU Era
Microcosmos combines Cosserat-style elastic filaments with a lattice-Boltzmann fluid solver in a fully differentiable JAX engine, validated on hand-designed gaits, gradient folding, neuroevolution, quality-diversity s...
-
A matrix-free, differentiable PyTorch solver for phase-field fracture: Formulation, benchmarks, and inverse analysis
A matrix-free, GPU-compatible PyTorch implementation of phase-field fracture with explicit dynamics, custom differentiable implicit damage solve, benchmarks on dynamic and quasi-static cases, and inverse recovery of f...
-
Embedding Hybrid Systems into Continuous Latent Vector Fields
An n-dimensional hybrid system embeds into a continuous vector field in m > 2n dimensions, enabling latent Neural ODEs with consistency losses to recover hybrid flows from time series.
-
Policy Optimization in Hybrid Discrete-Continuous Action Spaces via Mixed Gradients
HPO enables unbiased policy optimization in hybrid action spaces by mixing differentiable simulation gradients with score-function estimates, outperforming PPO as continuous dimensions increase.
-
Differentiable Satellite Constellation Configuration via Relaxed Coverage and Revisit Objectives
Four continuous relaxations turn non-differentiable coverage and revisit calculations into a fully differentiable pipeline that optimizes satellite orbits via gradients and outperforms metaheuristics.
-
VertAX: a differentiable vertex model for learning epithelial tissue mechanics
VertAX supplies a differentiable JAX implementation of vertex models for confluent epithelia that enables forward simulation, mechanical parameter inference, and inverse design of tissue-scale behaviors.
-
Fast and Reliable Gradients for Deformables Across Frictional Contact Regimes
An analytical adjoint with smoothed Fischer–Burmeister contact and a Woodbury-preconditioned Krylov solver yields stable gradients for frictional deformable contact.
-
Early-Terminable Energy-Safe Iterative Coupling for Parallel Simulation of Partitioned Port-Hamiltonian Systems
Douglas–Rachford iteration in scattering coordinates yields an anytime energy-safe coupling for partitioned port-Hamiltonian systems with finite-iteration passivity and hard-coupling convergence.
-
Identifiability-Aware Source Apportionment in City-Scale Advection-Diffusion Systems
IASA makes source apportionment report what a sensor network can actually tell apart: the rank and singular values of the projected response matrix set the finest defensible attribution resolution.
-
jaxdae: A JAX-native Differentiable Solver for Differential-Algebraic Equations in Coupled Multi-physics
jaxdae provides the first JAX-native differentiable DAE solver, using a frozen-grid BDF-2 replay adjoint for reverse-mode gradients and XLA-fused batched sweeps.
-
Differentiable Cardiac Electrophysiology Simulations for Dynamical State and Parameter Estimation
A differentiable solver for cardiac reaction-diffusion waves recovers hidden states and parameters of the Aliev-Panfilov model from sparse, surface-only, or noisy voltage observations, including two experimental monol...
-
Systematic LLM Translation of Legacy Scientific Code to Differentiable Frameworks: Application to a Land Surface Model
An LLM-based pipeline systematically converts the CLM-ml-v2 Fortran land surface model to JAX, delivering full differentiability, 8x fewer optimization steps, and 24x wall-clock speedup at N=2048 while releasing the c...
-
The Invisible Hand of Physics: When Video Diffusion Models Know More Than They Show
Physical plausibility is linearly decodable from diffusion transformer states in video models at 81.27% accuracy on IntPhys and InfLevel, absent from VAE latents and outperforming V-JEPA and VideoMAE.
-
Critical evaluation of PINN for FWD inverse analysis and differentiable FEM as an alternative
Standard PINNs fail to recover layer moduli in multilayer pavement systems due to domain discontinuities; XPINN improves but remains sensitive to weighting and noise; DiffFEM yields superior accuracy, stability, and e...
-
Batched Differentiable Rigid Body Dynamics in PyTorch for GPU-Accelerated Robot Learning
BARD is a self-contained PyTorch implementation of Featherstone's rigid body dynamics optimized for batched GPU evaluation and differentiation, reporting up to 64x throughput gains over Pinocchio on robot models with ...
-
$\Delta$ynamics: Language-Based Representation for Inferring Rigid-Body Dynamics From Videos
A vision-language framework generates text-based rigid-body scene configurations from videos using motion reasoning and optical flow, reporting 0.30 IoU on CLEVRER (7x over baselines) and transfer to 235 real videos.
-
COSMIC: Concurrent Optimization of Structure, Material, and Integrated Control for robotic systems
COSMIC co-optimizes robot structure, materials, and control simultaneously via differentiable simulation and constrained gradients, yielding locomotion strategies that outperform sequential baselines.
-
RigidFormer: Learning Rigid Dynamics using Transformers
RigidFormer learns mesh-free rigid dynamics from point clouds using object-centric anchors, Anchor-Vertex Pooling, Anchor-based RoPE, and differentiable Kabsch alignment to enforce rigidity.
-
Exploring the Boundaries of Differentiable Radiation Transport and Detector Simulation
Mitigation strategy for exploding gradients at material boundaries in differentiable radiation transport enables stable, optimization-ready derivatives for detector design.
-
Exploring the Boundaries of Differentiable Radiation Transport and Detector Simulation
Targeted halting of gradient flow at unstable material boundaries enables stable derivatives for optimizing detector designs in radiation transport simulations.
-
Neural Control: Adjoint Learning Through Equilibrium Constraints
Neural Control uses adjoint differentiation of equilibrium conditions to compute trajectory-dependent proxy gradients for history-dependent implicit models in deformable object manipulation.
-
SoMA: A Real-to-Sim Neural Simulator for Robotic Soft-body Manipulation
SoMA couples robot joint actions, environmental forces, and learned Gaussian-splat dynamics into a single neural simulator, improving resimulation and generalization on real robot soft-body manipulation by about 20% o...
-
PhysChoreo: Physics-Controllable Video Generation with Part-Aware Semantic Grounding
A two-stage framework that predicts per-part material properties from a single image and uses editable physics simulation to guide video generation.
-
PhysMorph-GS: Render-Guided Volumetric Morphing with Differentiable Physics
PhysMorph-GS injects visual supervision via deformation gradients in differentiable physics simulation and uses phased Chamfer-guided plasticity to reduce silhouette error by up to 49.9% compared to physics-only baselines.
-
Physically Viable World Models: A Case for Query-Conditioned Embodied AI
Embodied AI requires query-conditioned world models that select the simplest physical abstraction sufficient to answer intervention queries.
-
Safety-Critical Control for Smoothed Implicit Contact Dynamics
Introduces boundary-focused rollouts to screen the smoothing parameter κ and augments a discrete-time CBF with a fixed robust margin to eliminate contact force violations in smoothed implicit dynamics simulations.
-
OrbiSim: World Models as Differentiable Physics Engines for Embodied Intelligence
OrbiSim builds a differentiable physics engine from world models to support gradient-based policy optimization and contact modeling in robotics.
-
Neural Control: Adjoint Learning Through Equilibrium Constraints
Neural Control introduces adjoint-based differentiation through implicit equilibrium constraints to enable memory-efficient gradient computation and robust receding-horizon MPC for multi-stable deformable object manip...
-
Fast and Reliable Gradients for Deformables Across Frictional Contact Regimes
The authors develop a differentiable simulator enforcing Markovian dynamics on a position-velocity manifold and using a mass-aligned preconditioner with a soft Fischer-Burmeister operator to produce stable gradients f...
-
Integrating Mechanistic and Data-Driven Models for Neurological Disorders through Differentiable Programming
This perspective paper categorizes hybrid architectures for combining mechanistic and data-driven models using residual learning, Neural ODEs, and solver-in-the-loop to model neurological disorder progression.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.