Discovery Teaching is a lightweight clicker-based audience response system with additional features for supporting back channel communication and feedback from the students. It also provides statistics on student performance in formative assessments and participation. Discovery Teaching allows students and instructors to engage in online dialogs and discussions about class material. These discussions can be driven by the instructor or left to the back channel among students alone. We provide functionality for students to give instantaneous personal cognitive and affective feedback to the instructor on how the students are feeling about the class. Students can give feedback at various points during the lecture of their own choice. Alternatively, an instructor can ask all students for feedback at any particular point. Discovery Teaching allows for a highly interactive pedagogy that involves all of the students, TAs and the instructor.

Visit Discovery Teaching

The Affective Tutor, envisioned to help instructors keep all students in a large lecture class engaged and actively learning material related to the course. We have found, by asking for feedback after every class, that in a large class of over 50 students, a significant number of students will feel the class went too slow and another group will feel the class went too fast. One of the most challenging aspects of teaching a large class is to keep the entire class engaged in the course material. The Affective Tutor helps instructors to determine in real-time how students feel about the pace of the class.
The tool also helps instructors to pin-point confusing topics, and to tackle the problem of keeping students engaged by providing a monitored back-channel that allows the bored students to help the confused students get back on track and into an engaged mode.

Visit The Affective Tutor

This work was based on an internship with Prof. Gary Parker at Connecticut College. The internship was based on a research project to develop a machine learning system that utilizes machine learning using a computer simulation of a robot, wireless communication and an over-head camera vision to learn control programs for on-ground mobile robots. The internship involved intense computer programming, robotics design and construction, systems design, assembly and construction, and utilization of theories of artificial intelligence and evolutionary computation. I spent about 10 fun and busy weeks in a Computer Science lab where I mostly worked alone under Prof Parker’s supervision and help whenever necessary.

This research uses periodic tests on the actual robot to test the control programs learned on the robot model by evolutionary computation and improve the learning process by altering the learning algorithm (Fitness Biasing) or changing the model (Co-Evolving Model Parameters).

In project we use a Cyclic Genetic Algorithm (CGA) to learn optimal gaits for a quadruped servo-robot with three degrees of movement per leg. An actual robot was used to generate a simulation model of the movement and states of the robot. The CGA used the robot’s unique features and capabilities to develop gaits specific for that particular robot. Tests done in simulation show the success of the CGA in evolving a reasonable control program and preliminary tests on the robot show that the resultant control program produces a suitable gait.

This project investigates a new selection scheme that is tested using a Cyclic Genetic Algorithm (CGA) for learning gaits for a hexapod servo-robot. The effectiveness of CGA in learning optimal gaits with roulette wheel selection and selection with greedy factors is compared. The results were analyzed based on fitness of the individual generated gaits, convergence time of the evolution process, and the fitness of the entire populations evolved. In this work, we use a previously developed CGA to learn gaits using a simulation model created from an actual hexapod robot. Results demonstrate that greedy factors tested tend to prematurely converge with sub-optimal fitness, whereas training with roulette wheel selection evolves more diverse populations with individuals that result in the desired optimal gaits.