There exist so many machine learning paradigms, like isolated villages hidden in a jungle. But how do they differ from each other? How can we relate them? What are their fundamental assumptions, formulations, and motivations? Where to use them? How to formulate your problem into one of them, or, when and how to create your own learning paradigms? In this course, we will take an in-depth tour in the jungle of machine learning paradigms.

Cross-listed with ENEE759N. Credit only granted for CMSC838C, CMSC498F, or ENEE759N.

AR, VR, and MR, collectively referred to as XR, are becoming ubiquitous for human-computer interaction with limitless applications and potentialuse. This course examines advances on real-time multi-modal XR systems in which the user is 'immersed' in and interacts with a simulated 3D environment. The topics will include display, modeling, 3D graphics, haptics, audio, locomotion, animation, applications, immersionand presence.

Prerequisites: Successful completion of a compiler course (CMSC430 or equivalent) and a computer organization/architecture (CMSC 411 or equivalent) is strongly recommended. Exploration of the interplay between computer architecture and programming languages, with a focus on applying PL formalisms and techniques to emerging computer architecture research. The course is structured into three parts: 1) Topics in various non-traditional computer architectures and computing paradigms (including dataflow processing, intermittent computing, persistent memory, reconfigurable architectures,etc.); 2) Programming languages *for* computer architecture (including design of hardware description languages and high-level synthesis languages, etc.); 3) Problems of end-to-end correctness guarantees (including verified and secure compilation, full-stack correctness proofs, etc.)

Exploration of the potential of human augmentation technologies, such as wearable computing, haptics, virtual reality, and more, to enhance human physical, psychological, and cognitive capabilities. Students willread relevant literature from the fields of Psychology and Human-Computer Interaction. Additionally, students will create low-fidelity paper mockups and a prototype using contemporary tools, such as an open-sourcehardware platform and a programming environment.

Prerequisites: Successful completion (minimum grade of C-) in CMSC351, CMSC420, and CMSC330, as well as one course from (MATH240, MATH341, MATH461); and permission of the CMNS-Computer Science department. This course is open to Master's or Doctoral students in Computer Science, Electrical and Computer Engineering, or Mechanical Engineering programs.

Cognitive Robotics explores the application of human cognitive intelligence to the design and development of intelligent robots. The course delves into the fundamental principles of human cognitive intelligence and its integration with robotics and machine learning. Students will learn to develop cognitive robot learning architectures and implement them using simulators like Pybullet, NVIDA Issac-Gym, and Meta Habitat 2.0. Through engaging class projects, students will apply their newly acquired knowledge to solve novel, challenging and practically useful problems, enabling them to make meaningful contributions to the field. This unique opportunity to bridge the gap between cognitive science and robot learning that empowers students to develop smarter andand more capable robotic systems.