NeurIPS 22 accepted paper list

NeurIPS 22 accepted papers list

EBM

• Diffusion Models as Plug-and-Play Priors
• video diffusion models
• Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding
• On analyzing generative and denoising capabilities of diffusion-based generative models
• BinauralGrad: A two stage conditional diffusion probabilistic model for binaural audio synthesis
• Thompson sampling efficiently learns to control diffusion process
• Generative Time Series Forecasting with Diffusion, Denoise and Disentanglement
• CARD: Classification and regression diffusion models
• Improve diffusion models for inverse problems using manifold constraints
• Elucidating the design space of diffusion-based generative models
• Score-based diffusion meets annealed importance sampling
• Deep equilibrium approaches to diffusion models
• Flexible diffusion modeling of long videos
• Conditional diffusion process for inverse halftoning
• Diffusion-LM improves controllable text generation
• Riemannian diffusion models
• Denoising diffusion restoration models
• DPM-Solver: A fast ODE solver for diffusion probabilistic model sampling in around 10 steps.
• First hitting diffusion models
• GENIE: High -order denoising diffusion solvers.
• Antigen-Specific antibody design and optimization with diffusion-based generative models
• Unsupervised representation learning from pre-trained diffusion probabilistic models.
• Exponential family model-based reinforcement learning via score matching
• Convergence for score-based generative modeling with polynomial complexity.
• Score-based models detect manifolds
• Concrete score matching: generalized score matching for discrete data.
• Score-based generative modeling secretly minimizes the Wasserstein distance.
• Wavelet score-based generative modeling
• Riemannian score-based generative modeling.
• End-to-end stochastic programming with energy-based model
• Adaptive multi-stage density ratio estimation for learning latent space energy-based model
• EGSDE: Unpaired image-to-image translation via energy-guided stochastic differential equations.
• A continuous time framework for discrete denoising models.

GAN

• Masked GANs are robust generation learners.
• Improving GANs via adversarial learning in latent spaces.
• Amortized projection optimization for sliced wasserstein generative models.

Dynamic systems

ODE

• Neural differential equations for learning to program neural nets though continuous learning rules.
• Do residual neural networks discretize neural ordinary differential equations?
• Constraining Gaussian processes to systems of linear ordinary differential equations.
• Imrpoving neural ordinary equations with Nesterov’s accelerated gradient method

SDE

• Learning white noises in neural stochastic differential equations
• Riemannian neural SDE: learning stochastic representation on manifolds.

PDE

• Neural Stochastic PDEs: Resolution invariant learning of continuous spatiotemporal dynamics

Control

• Neural stochastic control
• Markov Chain Score Ascent: A unifying framework of variational inference with Markovian gradients.

Time Series

• Self-supervised contrastive pre-training for time seires via time-frequency consistency
• BILCO: An efficient algorithm for joint alignment of time series.
• GT-GAN: General purpose time series synthesis with generative adversarial networks.
• Generative Time Series Forecasting with Diffusion, Denoise and Disentanglement
• Non-stationary transformers: rethinking the staionarity in time series forecasting
• Time dimension dances with simplicial complexes: Zigzag filtration curve based supra-hodge convolution networks for time-series forecasting
• FILM: Frequency improved legendre memory model for long-term time series forecasting.
• Learning latent seasonal-trend representations for time series forecasting.
• SCINet: Time series modeling and forecasting with sample convolution and interaction.
• WaveBound: Dynamically bounding error for stable time series forecasting.
• Causal disentanglement for time series.
• Dynamic sparse network for time series classification: Learning what to “See”
• Efficient learning of nonlinear prediction models with time-series privileged information.
• Multivariate time-series forecasting with temporal polynomial graph neural networks.
• Dynamic graph neural networks under spatio-temporal distribution shift.
• AutoST: Towards the universal modeling of spatio-temporal sequences.
• Neural Stochastic PDEs: Resolution invariant learning of continuous spatio-temporal dynamics
• Variational context adjustment for temporal event prediction under distribution shifts.
• Practical adversarial attacks on spatio-temporal traffic forecasting models.
• Quo Vadis: Is trajectory forecasting the key towards long-term multi-objective tracking
• Contact-aware human motion forecasting.
• Forecasting Human Trajectory from scene history
• Motion forecasting transformer with global intention localization and local movement refinement.
• Representing spatial trajectories as distributions.