1. Deep Learning

  1. Dive into Deep Learning
  2. Neural Networks: Zero to Hero by Andrej Karpathy: This is how I got started with AI-ML.
  3. UV-A Deep Learning Course
  4. Deep Learning: I am trying my best to read this book regularly, I post about it here
  5. Pattern Recognition and Machine Learning
  6. On the opportunities and risks of Foundation Models
  7. A cookbook of Self-Supervised Learning

2. Autoencoders

  1. Introduction to Autoencoders
  2. Autoencoders: Chapter from Deep Learning Book
  3. Auto-Encoding Variational Bayes: Foundation paper on VAEs (Variational AutoEncoders)

I am a management/analytics guy, and even today most of the data I work with is structured data (even though it generally needs good bit of data-cleanup, feature engineering to make it useful and so on). I am still working inside the realm of structured data - Tables, Graphs, Time-Series Data as opposed to Text & Images. While the Deep learning research on unstructured data (text & images) today has literally exploded, the same for structured data is slowly picking its pace as I see it. Add to that, the kind of data I am working with is tabular data, but when processed using a normal deteministic algorithm, it results in graph data. So this whole area of research is very relevant for me and I find it quite interesting too. So will be listing stuff about structured data (Tabular-Graph-Time Series data in particular).

1. Graph Data

  1. Introduction to Graph Neural Networks: Foundations, Frontiers & Applications by Lingfei Wu
  2. Introduction to Graph Neural Networks: A Textbook
  3. Introduction to Graph Neural Networks: A Starting Point for Machine Learning Engineers
  4. A Gentle Introduction to Graph Neural Networks
  5. Graph Neural Networks in Big Data Analytics: Introduction
  6. Introduction to Graph Neural Networks: Lecture @ EPFL
  7. Introduction to Graph Neural Networks: Lecture @ Stanford
  8. Graph Foundation Models: Concepts, Opportunities & Challenges: March 2025
  9. Position: Graph Foundation Models are Already Here
  10. A Comprehensive Survey on Graph Neural Networks: August 2019
  11. A Comprehensive Survey on Deep Graph Representation Learning
  12. Deep Learning on Graphs: A Survey: 2020

2. Tabular Data

  1. Deep Neural Networks and Tabular Data: A Survey: A near survey on Deep Learning and Tabular Data. A side note: Dr. Vadim Borisov runs a cool startup called tabularis.ai, check it out if interested.
  2. A Short Chronology of Deep Learning for Tabular Data by Sebastian Raschka (this article is god’s work!): July 2022
  3. Why Tabular Foundation Models Should be a Research Priority: May 2024
  4. Is Deep Learning finally better than Decision Trees on Tabular Data: Feb 2025
  5. Towards Foundation Models for Learning Tabular Data: 2024
  6. TabICL: A Tabular Foundation Model for In-Context Learning on Large Data
  7. Language Models are Realistic Tabular Data Generators: Another paper whose aim I am interested in. Business Processes are my domain of interest, and from what I understand there is not a single large, standard dataset. Old datasets of BPI Challenges (2012-2022) are generally used even today. I think generating realistic synthetic datasets based on features of those smaller, old datasets might help research, just a vague idea.
  8. Deep Learning within Tabular Data: Foundations, Challenges, Advances & Future Directions: 2025
  9. A Survey on Deep Tabular Learning: Oct 2024

3. Time-Series Data

  1. Time Series Data Augmentation for Deep Learning: A Survey: March 2022
  2. Deep Learning for Time Series Forecasting: A Survey

  3. TIme Series Forecasting with Deep Learning: A Survey

  4. From Tables to Time: How TabPFN-v2 Outperforms Specialized Time Series Forecasting Models: This is kind of thing I am looking for. I am working on data that can be interpreted as Tabular Data, Graph Data or Time-Series Data - all 3 are deterministic transformations of one another. So this kind of work where a Tabular Foundation Model works better for a certain Time Series data than a specialized Time Series model itself enthuses me.