# MTGNN

Summary: Key Components Time Convolution (TC) Module Time Convolution The temporal convolution is responsible for capturing temporal patterns in a sequence. Graph Convolution Module Mix-hop Propagation in GNN Mix-hop is a strategy to avoid oversmoothing in GNN Graph Structure Learning Layer Graph Structure Learning in GNN We can learn a graph structure without prior knowledge Architecture Wu et al., 2020

References:
- Wu Z, Pan S, Long G, Jiang J, Chang X, Zhang C. Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks. arXiv [cs.LG]. 2020. Available: http://arxiv.org/abs/2005.11650
- Lässig F. Temporal Convolutional Networks and Forecasting. In: Unit8 [Internet]. 6 Jul 2021 [cited 28 Nov 2022]. Available: https://unit8.com/resources/temporal-convolutional-networks-and-forecasting/

Tags:

Status:

# StemGNN

Summary: What problem is StemGNN solving: intra-series temporal pattern: DFT Each series inter-series correlations At each step, the interactions between nodes reversible operator Example problem Covid cases: DE, AT, NL, … Predicting each country without considering the interactions between them Or introduce the people flow between them GFT: Completes DFT as it takes care of the inter-series correlations one extra slide for this topic Convolutions on Graphs one extra slide for this topic Graph Basics How to build the graph “self-attention”: outer product of key, query, as the adjacency matrix key, query are of length # of ndoes Weights and Biases LC 1DConv GLU FC Experiments Traffic adjacency matrix neighbouring sensors have higher correlations Covid Neighbouring countries have higher correlation Spetral analysis: Some eigenvectors have clear meanings Week spots Paper and code are not consistent https://github.

Tags:

Status:

# Self-supervised Learning: Generative or Constrastive

Summary: Review of self-supervised learning.

Types: { paper }

Tags:

Status:

# Normalizing Flows: An Introduction and Review of Current Methods

Summary: To generate complicated distributions step by step from a simple and interpretable distribution.

Types: { paper }

Tags:

Status:

# The Grammar of Graphics

Summary: Reading notes for the book The Grammar of Graphics

Types: { book }

Tags:

Status:
10%

# Improving Document Ranking with Dual Word Embeddings

Summary: Word2vec produces two embedding spaces, the in-embedding and out-embedding.

References:
- Nalisnick, Eric, Bhaskar Mitra, Nick Craswell, and Rich Caruana. 2016. Improving Document Ranking with Dual Word Embeddings.

Types: { paper }

Tags:

Status:

# The Art of Data Science

Summary: A nice and elegant book on data science

References:
- The Art of Data Science
- Leek, J. T., & Peng, R. D. (2015). What is the question? Science, 347(6228), 1314–1315.

Types: { book }

Tags:

Status:

# Human Graphical Perception of Quantitative Information in Data Visualization

Summary: Data visualization caveats

Types: { paper }

Tags:

Status:

# Information Theory and Statistical Mechanics

Summary: Max entropy principle as a method to infer distributions of statistical systems

References:
- Jaynes, E. T. (1957). Information Theory and Statistical Mechanics. Physical Review, 106(4), 620–630.

Types: { paper }

Tags:

Status:

# Schaum's Outline of Theories and Problems of Elements of Statistics I and II

Summary: The basics and all of modern statistics

References:
- Schaum's Outline of Theories and Problems of Elements of Statistics I and II, by Ruth Bernstein and Stephen Bernstein

Types: { book }

Tags:

Status:
80%

# Data Mining: Concepts and Techniques

Summary: How data mining was done in the past

References:
- Data Mining: Concepts and Techniques

Types: { book }

Tags:

Status:
70%

# Overcoming catastrophic forgetting in neural networks

Summary: Using a newly defined loss function the authors could implement an idea that achieves the multi-task within one network.

References:
- Overcoming catastrophic forgetting in neural networks

Tags:

Status:

# Working Memory and Brain Waves

Summary: Working memory might be related to the background brain waves from theoretical point of view

References:
- Flexible frequency control of cortical oscillations enables computations required for working memory

Tags:

Status:

# Popularity versus similarity in growing networks

Summary: Introduce geometry into the manifold of complex networks

References:
- Popularity versus similarity in growing networks

Tags:

Status: