当前位置:网站首页>CIKM 2022 AnalytiCup Competition: Federal Heterogeneous Task Learning
CIKM 2022 AnalytiCup Competition: Federal Heterogeneous Task Learning
2022-08-11 07:54:00 【Alibaba Cloud Developer】
Introduction:In order to explore the heterogeneity in federated learning and promote the development of federated learning, Alibaba DAMO Academy Intelligent Computing Laboratory and Tianchi jointly held "CIKM 2022 AnalytiCup Competition: Federation Heterogeneous Task Learning"We look forward to using this competition to help break the "data island" in real applications and promote the sharing of data value.
CIKM 2022 AnalytiCup Competition: Federal Heterogeneous Task Learning
Federated Learning is a new machine learning paradigm that allows multiple participants to jointly train machine learning models without directly sharing their respective data.Its core challenge is how to deal with the heterogeneity between participants (Heterogeneity), among which the heterogeneity of data distribution (non-IID) has attracted extensive attention in the research community and has quickly become one of the research hotspots in the field of federated learning.one.However, in many practical applications of federated learning, the heterogeneity among the participants in the federated task is often more complex and challenging: not only the data distribution, but even the tasks of the participants will exhibit large differences.For example, in a molecular graph federated learning task, the goal of some participants is to judge the type of molecules, that is, a classification task, and the goal of other participants is to predict the strength of molecular chemical properties, that is, a regression task.In this task scenario, although the participants all require the trained model to have the ability to understand the representation of molecular graphs, the specific learning goals are completely different and more challenging than the heterogeneity of the data distribution.
In order to meet the above-mentioned challenges of federated learning in real-world applications, the Intelligent Computing Laboratory of Alibaba DAMO Academy proposed a new federated learning setting: Federated hetero-task learning.Compared with traditional federated learning, this setting encourages researchers to combine federated learning with multi-task learning (Multi-task learning), model pre-training (Model pre-training), automatic machine learning (Auto-ML) and other fieldsIt integrates the research concepts of the real-world applications, so as to open up the "data silos" in practical applications, and finally realize the sharing of data value.At the same time, the Intelligent Computing Laboratory of DAMO Academy designed and implemented an open-source federated learning platform-FederatedScope[1,2] to help researchers more easily explore, design, and implement federated heterogeneous task learning algorithms, and fully implement them.ground verification.
At the same time, the Intelligent Computing Laboratory of Alibaba DAMO Academy and Tianchi jointly held the "CIKM 2022 AnalytiCup Competition: Federal Heterogeneous Task Learning" competition and prepared generous rewards. We look forward to more students participating in the Commonwealth Heterogeneous Task.In the exploration of task learning.
In order to help contestants get started as soon as possible, we have prepared a detailed tutorial
a>, and provides a playground that can be used directly by the contestants.- Official website of the competition: https://tianchi.aliyun.com/competition/entrance/532008/introduction
- FederatedScope Open Source Federated Learning Platform: https://github.com/alibaba/FederatedScope
Contest prize money:
- First Place: 5000 USD
- Second Place: 3000 USD
- Third Place: 1500 USD
- Fourth-Tenth: $500 each
Competition Certificate:
- Number 1 to 20: Certificate of Ranking
- Other: Certificate of Entry
Competition schedule (all times are calculated at 11:59PM UTC on the deadline)
- July 15, 2022: The competition starts, and the dataset and code are available for download.Forum and tournament leaderboards are starting to update.
- September 01, 2022: Contest registration closes.
- September 11, 2022: Deadline for submission of forecast results.
- September 12, 2022: The organizer conducts a code review, and the top 30 teams on the leaderboard will automatically enter the code review stage.
- September 18, 2022: Announcement of code review results.
- September 21, 2022: "CIKM 2022 AnalytiCup Competition" results announced.
- October 17, 2022: The CIKM 2022 conference begins.
Citation
- FederatedScope: A Flexible Federated Learning Platform for Heterogeneity. arXiv preprint 2022. pdf
- FederatedScope-GNN: Towards a Unified, Comprehensive and Efficient Package for Federated Graph Learning. KDD 2022. pdf
Original link:https://developer.aliyun.com/article/991177?a>
Copyright Statement: The content of this article is contributed by Alibaba Cloud real-name registered users, the copyright belongs to the original author, and the Alibaba Cloud developer community does not own the copyright and does not assume the corresponding legal responsibility.For specific rules, please refer to the "Alibaba Cloud Developer Community User Service Agreement" and "Alibaba Cloud Developer Community Intellectual Property Protection Guidelines".If you find any content suspected of plagiarism in this community, fill out the infringement complaint form to report it. Once verified, this community will delete the allegedly infringing content immediately.
边栏推荐
- 1046 划拳 (15 分)
- 【LaTex-错误和异常】\verb ended by end of line.原因是因为闭合边界符没有在\verb命令所属行中出现;\verb命令的正确和错误用法、verbatim环境的用法
- 3.2-分类-Logistic回归
- 零基础SQL教程: 主键、外键和索引 04
- XXL-JOB 分布式任务调度中心搭建
- 3.1-Classification-probabilistic generative model
- 1101 B是A的多少倍 (15 分)
- leetcode: 69. Square root of x
- 【LeetCode每日一题】——682.棒球比赛
- 线程交替输出(你能想出几种方法)
猜你喜欢
通过记账,了解当月收支情况
The most complete documentation on Excel's implementation of grouped summation
TF中的One-hot
机器学习总结(二)
查找最新人员工资和上上次人员工资的变动情况
My creative anniversary丨Thank you for being with you for these 365 days, not forgetting the original intention, and each is wonderful
tf.reduce_mean() and tf.reduce_sum()
1081 Check Password (15 points)
我的创作纪念日丨感恩这365天来有你相伴,不忘初心,各自精彩
关于#sql#的问题:怎么将下面的数据按逗号分隔成多行,以列的形式展示出来
随机推荐
经典论文-MobileNet V1论文及实践
Pico neo3 Unity Packaging Settings
软件测试常用工具的用途及优缺点比较(详细)
1076 Wifi密码 (15 分)
matrix multiplication in tf
【sdx62】XBL设置共享内存变量,然后内核层获取变量实现
Four states of Activity
CSDN21天学习挑战赛——封装(06)
查找最新人员工资和上上次人员工资的变动情况
Decrement operation in tf; tf.assign_sub()
Activity的四种启动模式
1061 判断题 (15 分)
流式结构化数据计算语言的进化与新选择
零基础SQL教程: 主键、外键和索引 04
matplotlib
Square, multi-power, square root calculation in Tf
项目2-年收入判断
Tensorflow中使用tf.argmax返回张量沿指定维度最大值的索引
The softmax function is used in TF;
线程交替输出(你能想出几种方法)