Zhijie Deng

Affiliations

Shanghai Jiao Tong University

papers

707

total citations

papers (20)

Orthus: Autoregressive Interleaved Image-Text Generation with Modality-Specific Heads

ICML 2025arXiv

citations

Accelerated Linearized Laplace Approximation for Bayesian Deep Learning

NEURIPS 2022arXiv

citations

SpikeZIP-TF: Conversion is All You Need for Transformer-based SNN

ICML 2024arXiv

citations

Meta-Unlearning on Diffusion Models: Preventing Relearning Unlearned Concepts

ICCV 2025arXiv

citations

SCott: Accelerating Diffusion Models with Stochastic Consistency Distillation

AAAI 2025arXiv

citations

Learning Sample Difficulty from Pre-trained Models for Reliable Prediction

NEURIPS 2023arXiv

citations

Confidence-based Reliable Learning under Dual Noises

NEURIPS 2022arXiv

citations

Towards Accelerated Model Training via Bayesian Data Selection

NEURIPS 2023arXiv

citations

Learning Neural Eigenfunctions for Unsupervised Semantic Segmentation

ICCV 2023arXiv

citations

Bayesian Exploration of Pre-trained Models for Low-shot Image Classification

CVPR 2024arXiv

citations

On Calibrating Diffusion Probabilistic Models

NEURIPS 2023arXiv

citations

Understanding and Exploring the Network with Stochastic Architectures

NEURIPS 2020

citations

AutoSync: Learning to Synchronize for Data-Parallel Distributed Deep Learning

NEURIPS 2020

citations

Advancing Text-to-3D Generation with Linearized Lookahead Variational Score Distillation

ICCV 2025arXiv

citations

Zhijie Deng

Affiliations

papers (20)

Adversarial Distributional Training for Robust Deep Learning

Black-Box Detection of Backdoor Attacks With Limited Information and Data

Online Speculative Decoding

LiBRe: A Practical Bayesian Approach to Adversarial Detection

CLLMs: Consistency Large Language Models

Improved Operator Learning by Orthogonal Attention

Orthus: Autoregressive Interleaved Image-Text Generation with Modality-Specific Heads

Accelerated Linearized Laplace Approximation for Bayesian Deep Learning

SpikeZIP-TF: Conversion is All You Need for Transformer-based SNN

Meta-Unlearning on Diffusion Models: Preventing Relearning Unlearned Concepts

SCott: Accelerating Diffusion Models with Stochastic Consistency Distillation

Learning Sample Difficulty from Pre-trained Models for Reliable Prediction

Confidence-based Reliable Learning under Dual Noises

Towards Accelerated Model Training via Bayesian Data Selection

Learning Neural Eigenfunctions for Unsupervised Semantic Segmentation

Bayesian Exploration of Pre-trained Models for Low-shot Image Classification

On Calibrating Diffusion Probabilistic Models

Understanding and Exploring the Network with Stochastic Architectures

AutoSync: Learning to Synchronize for Data-Parallel Distributed Deep Learning

Advancing Text-to-3D Generation with Linearized Lookahead Variational Score Distillation

papers (20)

Adversarial Distributional Training for Robust Deep Learning

Black-Box Detection of Backdoor Attacks With Limited Information and Data

Online Speculative Decoding

LiBRe: A Practical Bayesian Approach to Adversarial Detection

CLLMs: Consistency Large Language Models

Improved Operator Learning by Orthogonal Attention

Orthus: Autoregressive Interleaved Image-Text Generation with Modality-Specific Heads

Accelerated Linearized Laplace Approximation for Bayesian Deep Learning

SpikeZIP-TF: Conversion is All You Need for Transformer-based SNN

Meta-Unlearning on Diffusion Models: Preventing Relearning Unlearned Concepts

SCott: Accelerating Diffusion Models with Stochastic Consistency Distillation

Learning Sample Difficulty from Pre-trained Models for Reliable Prediction

Confidence-based Reliable Learning under Dual Noises

Towards Accelerated Model Training via Bayesian Data Selection

Learning Neural Eigenfunctions for Unsupervised Semantic Segmentation

Bayesian Exploration of Pre-trained Models for Low-shot Image Classification

On Calibrating Diffusion Probabilistic Models

Understanding and Exploring the Network with Stochastic Architectures

AutoSync: Learning to Synchronize for Data-Parallel Distributed Deep Learning

Advancing Text-to-3D Generation with Linearized Lookahead Variational Score Distillation