Search Results for author: Zhezhi He

Found 22 papers, 11 papers with code

CLLMs: Consistency Large Language Models

1 code implementation28 Feb 2024 Siqi Kou, Lanxiang Hu, Zhezhi He, Zhijie Deng, Hao Zhang

Parallel decoding methods such as Jacobi decoding show promise for more efficient LLM inference as it breaks the sequential nature of the LLM decoding process and transforms it into parallelizable computation.

ResSFL: A Resistance Transfer Framework for Defending Model Inversion Attack in Split Federated Learning

1 code implementation CVPR 2022 Jingtao Li, Adnan Siraj Rakin, Xing Chen, Zhezhi He, Deliang Fan, Chaitali Chakrabarti

While such a scheme helps reduce the computational load at the client end, it opens itself to reconstruction of raw data from intermediate activation by the server.

Federated Learning

CP-ViT: Cascade Vision Transformer Pruning via Progressive Sparsity Prediction

1 code implementation9 Mar 2022 Zhuoran Song, Yihong Xu, Zhezhi He, Li Jiang, Naifeng Jing, Xiaoyao Liang

We explore the sparsity in ViT and observe that informative patches and heads are sufficient for accurate image recognition.

N3H-Core: Neuron-designed Neural Network Accelerator via FPGA-based Heterogeneous Computing Cores

1 code implementation15 Dec 2021 Yu Gong, Zhihan Xu, Zhezhi He, Weifeng Zhang, Xiaobing Tu, Xiaoyao Liang, Li Jiang

From the software perspective, we mathematically and systematically model the latency and resource utilization of the proposed heterogeneous accelerator, regarding varying system design configurations.


RADAR: Run-time Adversarial Weight Attack Detection and Accuracy Recovery

1 code implementation20 Jan 2021 Jingtao Li, Adnan Siraj Rakin, Zhezhi He, Deliang Fan, Chaitali Chakrabarti

In this work, we propose RADAR, a Run-time adversarial weight Attack Detection and Accuracy Recovery scheme to protect DNN weights against PBFA.

MetaGater: Fast Learning of Conditional Channel Gated Networks via Federated Meta-Learning

no code implementations25 Nov 2020 Sen Lin, Li Yang, Zhezhi He, Deliang Fan, Junshan Zhang

In this work, we advocate a holistic approach to jointly train the backbone network and the channel gating which enables dynamical selection of a subset of filters for more efficient local computation given the data input.

Meta-Learning Quantization

KSM: Fast Multiple Task Adaption via Kernel-wise Soft Mask Learning

no code implementations CVPR 2021 Li Yang, Zhezhi He, Junshan Zhang, Deliang Fan

Thus motivated, we propose a new training method called \textit{kernel-wise Soft Mask} (KSM), which learns a kernel-wise hybrid binary and real-value soft mask for each task, while using the same backbone model.

Continual Learning

A Progressive Sub-Network Searching Framework for Dynamic Inference

no code implementations11 Sep 2020 Li Yang, Zhezhi He, Yu Cao, Deliang Fan

Many techniques have been developed, such as model compression, to make Deep Neural Networks (DNNs) inference more efficiently.

Model Compression

T-BFA: Targeted Bit-Flip Adversarial Weight Attack

2 code implementations24 Jul 2020 Adnan Siraj Rakin, Zhezhi He, Jingtao Li, Fan Yao, Chaitali Chakrabarti, Deliang Fan

Prior works of BFA focus on un-targeted attack that can hack all inputs into a random output class by flipping a very small number of weight bits stored in computer memory.

Adversarial Attack Image Classification

Defending and Harnessing the Bit-Flip Based Adversarial Weight Attack

1 code implementation CVPR 2020 Zhezhi He, Adnan Siraj Rakin, Jingtao Li, Chaitali Chakrabarti, Deliang Fan

Recently, a new paradigm of the adversarial attack on the quantized neural network weights has attracted great attention, namely, the Bit-Flip based adversarial weight attack, aka.

Adversarial Attack Binarization +1

TBT: Targeted Neural Network Attack with Bit Trojan

3 code implementations CVPR 2020 Adnan Siraj Rakin, Zhezhi He, Deliang Fan

However, when the attacker activates the trigger by embedding it with any input, the network is forced to classify all inputs to a certain target class.

Non-Structured DNN Weight Pruning -- Is It Beneficial in Any Platform?

no code implementations3 Jul 2019 Xiaolong Ma, Sheng Lin, Shaokai Ye, Zhezhi He, Linfeng Zhang, Geng Yuan, Sia Huat Tan, Zhengang Li, Deliang Fan, Xuehai Qian, Xue Lin, Kaisheng Ma, Yanzhi Wang

Based on the proposed comparison framework, with the same accuracy and quantization, the results show that non-structrued pruning is not competitive in terms of both storage and computation efficiency.

Model Compression Quantization

Robust Sparse Regularization: Simultaneously Optimizing Neural Network Robustness and Compactness

no code implementations30 May 2019 Adnan Siraj Rakin, Zhezhi He, Li Yang, Yanzhi Wang, Liqiang Wang, Deliang Fan

In this work, we show that shrinking the model size through proper weight pruning can even be helpful to improve the DNN robustness under adversarial attack.

Adversarial Attack

Bit-Flip Attack: Crushing Neural Network with Progressive Bit Search

1 code implementation ICCV 2019 Adnan Siraj Rakin, Zhezhi He, Deliang Fan

Several important security issues of Deep Neural Network (DNN) have been raised recently associated with different applications and components.

Parametric Noise Injection: Trainable Randomness to Improve Deep Neural Network Robustness against Adversarial Attack

1 code implementation CVPR 2019 Adnan Siraj Rakin, Zhezhi He, Deliang Fan

Training the network with Gaussian noise is an effective technique to perform model regularization, thus improving model robustness against input variation.

Adversarial Attack Adversarial Defense +1

Simultaneously Optimizing Weight and Quantizer of Ternary Neural Network using Truncated Gaussian Approximation

no code implementations CVPR 2019 Zhezhi He, Deliang Fan

In the past years, Deep convolution neural network has achieved great success in many artificial intelligence applications.

Blind Pre-Processing: A Robust Defense Method Against Adversarial Examples

no code implementations5 Feb 2018 Adnan Siraj Rakin, Zhezhi He, Boqing Gong, Deliang Fan

Blind pre-processing improves the white box attack accuracy of MNIST from 94. 3\% to 98. 7\%.

Adversarial Attack

Developing All-Skyrmion Spiking Neural Network

no code implementations8 May 2017 Zhezhi He, Deliang Fan

In this work, we have proposed a revolutionary neuromorphic computing methodology to implement All-Skyrmion Spiking Neural Network (AS-SNN).

Handwritten Digit Recognition

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