IEEE C/DC 2941 Working Group
IEEE C/DC AI Model Representation, Compression, Distribution and Management Working Group

Meeting Minutes of the 3rd IEEE DC P2941 Working Group Plenary Meeting

The 3rd AIM Working Group plenary meeting was held via conference call, and quorum reached.

IEEE Patent Policy: Call for Patents

The call for patents was issued; none raises or if any concerns were raised for consideration.

IEEE Copyright Policy

The copyright policy was presented. There were no questions or concerns.

Approval of the Agenda

Motion to approve the meeting agenda. (Mover: Ms. Xiaoxu Luan; Second: Dr. Chao Yang). The agenda was unanimously approved as presented without objection.

14 technical contributions (proposals) received and accepted on this meeting. Below is the list:

AIM-3-P0001: Coding method for highly unified neural network (Wei Wang, Wei Jiang, Shan Liu)

AIM-3-P0002: A Data-Free Neural Network Compression Method (Chen Hanting, Xu Yixing, Liu Chuanjian, Wang Yunhe, Xu Chunjing, Qian Li)

AIM-3-P0003: Neural Network Representation (Yonghong Tian, Guangyao Chen, Yemin Shi)

AIM-3-P0004: Adding operations in P2941 (xiaoguang hu, xiaoru li, xiaolin jiang)

AIM-3-P0005: Interval shrinkage quantization (Yonghong Tian, Quan Zhang, Lechun Zhang, Zhaoyi Yan, Yemin Shi, Zheyang li)

AIM-3-P0006: Residual quantization compression (Wenhong Duan, Zhenghui Zhao, Kai Lin, Chuanmin Jia, Shanshe Wang, Siwei Ma, Li wang, Shiliang Pu)

AIM-3-P0007: Neural network model encryption (Bo Peng, Zheyang Li, Wenming Tan, Xueming Chen, Wenqi Ren)

AIM-3-P0008: Integer-Arithmetic-Only Convolutional Neural Networks with Bounded ReLU (Hengrui Zhao, Dong Liu, Houqiang Li, Feng wu)

AIM-3-P0009: Comments to sections: Structured Matrix, Data Generation and Multiple Models (Huawei Technologies Co.,Ltd.)

AIM-3-P0010: A Quantization Method for Image Super Resolution Reconstruction with Trainable Upper-Bound Cutoff. (Rongrong Ji, Huixia Li, Chenqian Yan, Shaohui Lin, Xiawu Zheng, Yuchao Li, Fan Yang)

AIM-3-P0011: Multi-bit neural network quantization algorithm (Rongrong Ji, Shaohui Lin, Yuchao Li, Yichen Zhang, Xiawu Zheng, Chenqian Yan, Huixia Li)

AIM-3-P0012: Deep neural network compression based on structured sparse filter pruning (Rongrong Ji, Shaohui Lin, Yuchao Li, Fei Chao)

AIM-3-P0013: Acceleration and compression of deep neural network via tensor low-rank group-sparse decomposition (Peisong Wang, Zhenmeng Zuo, Yifan Zhang, Jian Cheng)AIM-3-P0014: Structured Pruning for Efficient ConvNets via Incremental Regularization (Haoji Hu, Xiang Li, Huan Wang, Yang Luo, Xinyi Hu, Hualiang Wang)

Approval of the Draft Standard Writing Plan

Motion to approve the Draft Standard Writing Plan. (Mover: Mr. Guangyao Chen; Second: Dr. Xiaoxu Luan).

The writing plan was unanimously approved without objection, the draft standard version 1 shall be collected by May 10th, 2021, and the completed draft standard will be available by May 25th 2021. The standard will be written based on the 14 contribution received and accepted on the meeting.

The introduction of Obtaining Copyright Permission for Contribution was given by Ms. Xiaoxu Luan, and no question raised.

Approval of the Liaison with Audio Video Coding Standard Workgroup of China

Motion to approve the liaison with Audio Video Coding Standard Workgroup of China. (Mover: Mr. Xiaoxu Luan; Second: Dr. Chao Yang). The liaison was unanimously approved as presented without objection. The coordinator is Dr. Yonghong Tian.

The next meeting is tentatively scheduled at 09:30am (Beijing Time Zone, GMT 8:00am) on June 19th at Hainan, China. Remote attendees can request to set up teleconference service at least one week ahead of the meeting.

The meeting closed at 10:30am, April 19th, 2021.


Annex 1: Meeting Agenda

Annex 2: Attendee List