Three off-hour lectures by Bijit Sarma (Faculty, CMI):

Title: Artificial Intelligence for Physicists: Fundamentals and Frontiers

Abstract: This series of talks will introduce machine learning (ML) fundamentals and applications in physics, starting with the basic structure of artificial neural networks, detailing the operation of artificial neurons, non-linear activation functions (such as sigmoids and ReLUs), and network layouts, while highlighting the universality

theorem that allows hidden layers to approximate arbitrary smooth functions. Building on this foundation, we will discuss the process of training a network. This section will cover the definition of cost functions, the mechanics of stochastic gradient descent, and the backpropagation algorithm - a highly efficient method that utilizes the chain rule to update millions of weights in a single backward pass.
We will then transition to advanced network architectures designed to exploit specific data structures. This part will cover Convolutional Neural Networks (CNNs), which utilize identical shifting weights (kernels) to achieve translational invariance in image processing, and autoencoders that perform unsupervised learning and feature extraction

by forcing data through a compressed bottleneck layer. Further, to handle temporal, sequential, or text-based data, we will introduce Recurrent Neural Networks (RNNs) and their robust variant, Long Short-Term Memory (LSTM) networks.
Moving beyond data classification, the talk will explore Reinforcement Learning (RL), a framework where an artificial agent learns to discover optimal strategies through trial-and-error interactions with an environment. Key RL concepts such as policy gradients, optimal baselines, Q-learning, and the critical balance between exploration and exploitation will be discussed.
In the final section, we will discuss several applications of ML in physics, specifically to quantum physics - showing where ML is actually making a difference right now.

We had sessions on careers beyond academia on Wednesday, 24th June:

Himalaya Senapati: An in-person discussion on career pathways from academia to industry: Himalaya Senapati (Equity Derivative Quant, HSBC).
Career Session: An online career insights session featuring professionals who transitioned from academia to industry across quant, analytics, and technology sectors.

Resources shared by Ritabrata Bhattacharya: Slides for the career session; CV used for quant-finance jobs.