Course Instructor
Dan Pei
Associate Professor
Department of Computer Science and Technology
Email: peidan(at)tsinghua(dot)edu(dot)cn
Office: East Main Building 9-319
Class Time and Location
Class Time: Wednesday 9:50am – 12:15pm (please see detailed schedule in the course syllabus below)
Location: 三教3507 (Room 3507, Teaching Building 3)
Project: Time Series Forecasting Paper Reading and Reproduction
The entire class collaboratively completes a time series forecasting benchmark. Each student will be assigned papers related to time series forecasting. After carefully reading the papers, students are required to reproduce the code and complete a benchmark report.
Grading Policies
- Attendance: 10%
- Paper Reading & Presentation: 20%
- Algorithm Reproduction & Project Presentation: 40%:
- Report Writing: 30%
Syllabus:
Week | Date | Topic | Techniques | Case Studies | Slides |
1 | Feb 28 (9:50am-12:15pm) | Introduction | 1. Introduction to the course 2. Introduction to AIOps 3. Brief Introduction to assignments and projects 4. ANM website |
Week 1 | |
2 | Mar 6 (9:50am-12:15pm) | Visualization | 1. Student Self-Introduction 2. Storytelling with Visualization 3. Introduction to Visualization |
Week 2 | |
3 | Mar 13 (9:50am-12:15pm) | Correlation | 1. Video Streaming Basics 3. Correlation and Regression 4. Information Gain |
2. SIGCOMM 2011 Case Study | Week 3 |
4 | Mar 20 (9:50am-12:15pm) | Decision Tree | 1. Decision Trees 3. Feature Selection 4. Accuracy_CV_Overfitting |
2. SIGCOMM 2013 Case Study | Week 4 |
5 | Mar 27 (9:50am-12:15pm) | KPI Anomaly Detection with Transfer Learning | 1. Time-Series Anomaly Detection Service at Microsoft 2. Time series algorithms Tutorial 3. Random Forest 4. Deep Learning 5. Deep Generative Model 6. Similarity 7. Clustering 8. Deep Sequence Learning 9. Transfer Learning |
10. Case study: Opprentice (IMC’ 15) 11. Case study: Donut (WWW’ 18) & ROCKA (IWQoS’18) 12. Case study: OmniAnomaly (KDD’ 19) & CTF (INFOCOM’ 21) |
Week 5 |
6 | Apr 3 (9:50am-12:15pm) | Week 6 | |||
7 | Apr 10 (9:50am-12:15pm) | Week 7 | |||
8 | Apr 17 (9:50am-12:15pm) | Foundation Model for KPI Forecasting |
1. Transformer |
2. Case Study: SimMTM (NeurIPS’ 23) |
|
9 | Apr 24 (9:50am-12:15pm) | Week 9 | |||
10 | May 1 (No class due to school holidays) | ||||
11 | May 8 (9:50am-12:15pm) | Multivariate Time Series Clustering and Outlier Detection |
1. STGNN |
3. OmniCluster: WWW 2022 |
|
12 | May 15 (9:50am-12:15pm) | Causal Discovery among Time Series |
1. Hiearchical Clustering |
2. CoFlux (IWQoS 2019) |
|
13 | May 22 (9:50am-12:15pm) | Week 13 | |||
14 | May 29 (9:50am-12:15pm) | Causal Inference on a Graph | 1. Microservices 2. Causal Inference 4. Agent |
3. CIRCA (KDD’22) 4. LatentScope 6. G-RCA |
Week 14 |
15 | Jun 5 (9:50am-12:15pm) | ||||
16 | Jun 12 (9:50am-12:15pm) | Project Presentation |
Course Information
《MIT 6.S191 Introduction to Deep Learning 》 with video and slides.
《Site Reliability Engineering –How Google Runs Production Systems》, by Betsy Beyer, Chris Jones, Jennifer Petoff & Niall Richard Murphy
Course Assistant
Zhe Xie
Email: xiez22(at)mails(dot)tsinghua(dot)edu(dot)cn