Overview

We are so excited you have decided to begin this journey with us! Research is fascinating: it fashions what the future will be; you can tackle some of the hardest problems in the world; you explore and learn deeply as you try to find the best solutions.

Our research is focused in helping robots learn efficiently and robustly directly in the real world. Imagine a robot performing in the real world and learning new tasks as it encounters things or situations it has never seen before. Imagine a human assigning a robot a new task it wants it to do and having that robot learning it quickly and robustly.

We are currently exploring deep reinforcement learning algorithms. Deep Reinforcement Learning algorithms have had a tremendous impact in the field of AI and robotics.

Here are a few examples:

  1. AlphaFold transformed structural biology by showing that deep neural networks could predict protein structures with near-experimental accuracy, dramatically accelerating biological and drug-discovery research—although importantly, AlphaFold is not a reinforcement-learning system (Jumper et al., Nature, 2021).
  2. AlphaZero demonstrated the extraordinary potential of deep reinforcement learning by learning chess, shogi, and Go from self-play with essentially no game-specific human knowledge, reaching superhuman performance using a common learning algorithm (Silver et al., Science, 2018).
  3. In robotics, systems such as QT-Opt brought deep reinforcement learning into the physical world, learning closed-loop visual manipulation from more than 580,000 real robot grasp attempts and achieving a reported 96% grasp success rate on previously unseen objects (Kalashnikov et al., CoRL, 2018).

How can you get started?

  1. Research Focus Get to know some of the work I have been involved with via my personal research home page in the area of intelligent robotic manipulation.

It spans control, failure recovery and robot introspection, human-robot collaboration, and self-learning: https://www.JuanRojas.net.

  1. Student Advice Look at the advice I give to students who want to work with me: https://www.juanrojas.net/students/

  2. Intelligent Robotics Track at Lipscomb Consider participating in our Intelligent Robotics track in the EECE department (Will formally start in Fall 27, but you can take all the classes right now). Students from other majors can still take many of the classes including:

    Intelligent Robotic Courses

    1. EECE/ME 4523 Mechatronics This course will cover fundamental robot physics and ROS2
    2. EECE 395v: Special Topics in Advanced Robotics This is a research based course. We look at a state-of-the-art publication, dissect it, implement it, and discuss its strengths and weaknesses. This course prepares students to perform research over the summer.

    Math Courses

    1. MA 3213 Linear Algebra
    2. MA3113 Theory of Statistics
    3. MA4113 Abstract Algebra

    Programming Courses

    1. CS2623 Design and Analysis of Algorithms In this class you will learn sophisticated algorithms and their time and space complexity.

    AI Courses

    1. EECE 4483 Fundamentals of Machine Learning In this course you will learn supervised and unsupervised machine learning (prediction, classification, Kmeans, and Neural Networks); Markov Decision Processes (MDPs); Reinforcement Learning (RL), and Deep Reinforcement Learning. For the final project you will deploy a DRL algorithm in simulation.
    2. One of our other AI offerings in CS
      1. CS 3453 Principles of AI (S)
      2. CS 4523 Modern AI Abstractions
      3. CS 4333 - Modern Implementations of AI (3)

  3. Workshops Check out one of the many workshops we have prepared to get quick overviews of important topics here.

  4. Our Tutorials You do not need to take all those classes to get started. As part of our undergraduate research lab vision, we are creating tutorials to help you get started as quickly as possible and develop proficiency in both theory and systems connected to intelligent robotics.

Theory Fundamentals

Our Theory Fundamentals page presents AI, ROS, and Robot Manipulation basics. They are arranged to be covered in 14 sittings. This will help you to get started in this area.

Theoretical Foundations