Deep Reinforcement Learning (DRL) stock trading system with LSTM-PPO, integrating technical indicators and FinBERT-based news sentiment analysis.

Learnable MAPF. “Distributed Heuristic Multi-Agent Path Finding with Communication” (DHC) algorithm from ICRA 2021 is implemented and benchmarked in out-of-distribution (OOD) scenarios. A new robust training loop to handle communication failures is introduced.

Multi-Agent Deep Reinforcement Learning for Collaborative Computation Offloading in Mobile Edge-Computing

Code for the paper "Age-of-Information-based Scheduling in Multiuser Uplinks with Stochastic Arrivals: A POMDP Approach"

This repository contains Dongming Shen's code and documentation for the research projects conducted at the AIDyS Lab, USC. The project focuses on integrating Reinforcement Learning (RL) to solve partially observable Markov decision processes (POMDP) under finite linear temporal logic (LTL) constraints.

Application of Reinforcement Learning algorithms (DQN,DRQN,PPO,A2C) to gym's MountainCar-v0

Climate change-related risk mitigation for infrastructure systems often requires adaptation. A computational framework for optimal decision-making under uncertainty based on dynamically changing conditions observed in time is developed in response.

Project exploring Policy Space Response Oracles (PSRO) in a Normative POMDPs

Course Project - Advanced Topics in Machine Learning - Autumn Semester 2023 - Indian Institute of Technology Bombay