Methods
During my contact time with students, I use two methods to create a productive learning environment: flipped classroom and active learning.
Flipped classroom:
I provide students with opportunities to learn course content at their own pace before coming to class. Each week, I provide self-learning material including readings, video lectures, and coding exercises students must complete prior to the class time. In addition, I supply a list of guided questions to complement the learning material in order to stimulate students鈥� interests in and attention to the content. Then, during the class time, we devote the majority of our time to interactive discussions and learning activities to further students鈥� understanding of course content. I previously applied flipped classroom in a course on rapid prototyping of user interfaces and in another course on big data. In order to keep each student accountable for their learning, I require each student to keep a learning journal (i.e., in the form of a series of Jupyter notebooks). In the journal, a student can check off each week鈥檚 assigned readings, video lectures, or coding exercises when completed. In addition, the student answer the guided questions.
Active learning:
During the in-class time, I use a variety of active learning techniques to engage students. (a) Retrieval practice typically happens at the start of class. I give students five minutes to write down key points they can remember from the week鈥檚 learning (i.e., readings, videos, and coding exercises they are supposed to complete prior to the class). They write on sticky notes and put them on the whiteboard to share with the entire class. Given the subject matter is about the intersection of AI and neuroscience, I give students specific prompts such as 鈥淲hat are the key points you learned about AI?鈥� and 鈥淲hat are the key points you learned about neuroscience?鈥� (b) Fishbowl usually follows the retrieval practice. I assign a small number of students in an inner circle and the rest of the students in an outer circle. I provide a discussion question based on that week鈥檚 material. Students in the inner circle serve as active participants, while students in the outside are observers. Sticky notes on the whiteboard can aid to the discussion. (c) Role playing is another option. I prepare a list of roles and scenarios designed to simulate real world situations where interests from different stakeholder groups may collide. One scenario may involve a skeptical patient and a researcher who wishes to recruit the patient to participate in a brain imaging study. Two students can take on the role of the patient and the researcher respectively. The student taking the role of patient tries to raise as many questions about the safety of the study as possible, while the student taking on the role of the researcher tries to provide arguments to address the safety concerns. (d) Peer teaching: The rest of the class time is given to students to teach their peers. Students take turns presenting and teaching others. They model the teaching methods I use for this course, rather than just showing slides to the class.
The target audience of this course is graduate students. I intend to use this course to help them develop skills and confidence to conduct independent scholarly works on topics in the intersection of artificial intelligence and neuroscience. The flipped classroom design will shift the burden as well as ownership of learning more to the students. Active learning methods will create an environment for students to actively contribute rather than passively receive. I expect the majority of the students will have some degree of experiences of active learning methods during their undergraduate study.
In terms of content knowledge, the goals are to cover AI, neuroscience, and ethics. The flipped classroom method will expose students to technical content about AI and neuroscience via a series of readings, video lectures, and coding exercises. Understanding of the ethical principles in AI and neuroscience will be achieved through role playing throughout the semester, which is more effective than just receiving a big lecture about ethics.
To measure whether flipped classroom works for an individual student, I look at (a) the learning journal kept by the students, (b) the sticky notes from the retrieval practice, and (c) the quality of in-class participation.
Activities Outside of Class
In addition to weekly learning, students work on a semester long project aimed to answer a research question using a real world brain imaging dataset and artificial intelligence. The components of the project include: proposing a research question, selecting a dataset, exploring the dataset, designing a data analysis plan, executing the plan, identifying key findings, understanding the limitations, reporting the findings to an academic audience, and presenting the findings to the general public. Students can gain technical skills to conduct research using brain imaging datasets and artificial intelligence algorithms and tools. Students can also gain communication skills by writing reports and giving presentations about their findings. Each component of the project is associated with a grading rubric to assess student鈥檚 performance along the way. The final report and presentation provide a summative assessment of students鈥� overall performance.
Course Materials
Given this is a special topic graduate class, there is no established textbook. I curate my own course materials including papers, video lectures, coding exercises, and public datasets. Students use the course materials in a flipped classroom setting.
Broader Curriculum
As expected, the majority of students are already equipped with some technical background in artificial intelligence and data science from previous courses. Some students have limited background in neuroscience. This course is the first time students are learning about the integration of neuroscience and AI; students can apply what they know about AI and data science to a new problem domain---neuroscience.
There is increased emphasis on interdisciplinary research in the broader department curriculum, as computer science in general and artificial intelligence in particular are transforming many other fields. This course contributes to expanding interdisciplinary work between computer science and neuroscience. Moreover, this course assists students to participate in cutting-edge research in artificial intelligence and neuroscience. For master students, one such endeavor could be to apply to a PhD program in computer science or in neuroscience. For PhD students in computer science, they learn how to develop new algorithms or tools to support neuroscience research. For PhD students in neuroscience, on the other hand, they can explore more advanced uses of AI in their dissertation studies.
Back (Background)聽 聽 聽 聽 聽聽 Next (Student Work)