Machine Learning and Optimization Seminar

The Machine Learning and Optimization is a student-led seminar in the Department of Mathematical Sciences at NJIT. Its goal is to expose the participants to topics in machine learning and optimization by:

1. Facilitating hands-on workshops and group discussions to explore and gain experience
2. Inviting speakers to introduce machine learning and optimization concepts (both established and under development)

Spring 2023

Workshops

• 2/16/23: Transformer neural networks (DALLE, ChatGPT, etc.) - David Mazowiecki
• 3/16/23: Inverse problems for acoustics - Diego Rios

Presentations

• 2/9/23: Generative adversarial networks - Soheil Saghafi
• 2/23/23: A deep learning approach to infer connectivity and neuronal dynamics from spike trains - Rodrigo Pena (NJIT)

---

Past Presentations

Fall 2022

Workshops

• 9/15/22: Python set up, machine learning basics, gradient descent, and automatic differentiation - Connor Robertson
  • Document (with code)
• 10/6/22: Neural networks - structure, building, and training - Jake Brusca
  • Presentation
  • Document (with code)
  • Google Colab notebook
• 10/13/22: Data-driven model discovery - Connor Robertson
  • Document (with code)
  • Video data
• 10/20/22: Convolutional neural networks - Soheil Saghafi
  • Presentation
  • Document (with code)
  • Image data
• 11/3/22: Recurrent neural networks - Austin Juhl
  • Presentation
  • Code repository
• 11/10/22: Recurrent neural networks for beat prediction - Prianka Bose
• 12/1/22: Reinforcement learning - Sepideh Nikookar
  • Presentation (pdf)
  • Presentation (pptx)
  • Document (with code)
  • Data (gifs)

Spring 2022

• Wasserstein GANs Work Because They Fail - Axel Turnquist
• Mean-field Theory: Drift and the Mean Drift - Binan Gu

Fall 2021

• Computing the Distance Between Probability Measures - Axel Turnquist
• Full Waveform Inversion Using the Wasserstein Metric - Brittany Hamfeldt
• Image Sharpening - Axel Turnquist
• Neural Networks for function approximation and data-driven modeling - Connor Robertson
• Optimal control of systems governed by PDEs with uncertainty - Georg Stadler
• Stochastic Temporal Networks - Binan Gu
• Topological Data Analysis Applied to Interaction Networks in Particulate Systems - Lou Kondic

Spring 2021

• Spin Glass and High Dimensional Energy Landscape - Binan Gu
• Diffusion Approximations and Applications in Non-convex Optimization - Binan Gu
• Stochastic Gradient Descent and How I Learned to Love the Randomness - Joe McCann

Fall 2020

• Why does stochastic gradient descent work so well? - Axel Turnquist
• Information Geometry & Learning - Axel Turnquist
• Graph-based Learning Beyond the Paradigm of Neural Networks - Binan Gu
• Effective Dimension in High-Dimensional Problems - Axel Turnquist
• Bayesian Statistics and Machine Learning - Gan Luan
• Matrix Completion and Sparse Recovery - Axel Turnquist
• Wasserstein GAN - Yixuan Sun
• Graphical Model Selection - Binan Gu
• Learning Frameworks - Axel Turnquist