Through my educational program, I aim to build a diverse professional and research workforce with a deep care for social issues and a passion about using data, models, and algorithms to address them. While varied in nature, my teaching and mentoring have a focus on methods from operations research (OR) and artificial intelligence (AI) and their use to address societal problems such as homelessness, substance use, biodiversity conservation, or suicide prevention. My research expertise in operations research and artificial intelligence for society enable me to help students gain a deep understanding of OR and AI techniques and how to use them in a practical way in their careers, while also encouraging further research in the increasingly active area of AI and OR for society. To achieve this, I continuously incorporate applications, motivating examples, and course projects into my classes, and also invite domain experts and community partners to give guest lectures and seminars, thus situating the methods I teach the students in diverse and inspiring contexts. My end goal is to educate my students and mentees in an integrated way as engineering professionals and researchers, as well as responsible citizens capable of using the skills they learn in our courses to make a difference in society.

In this statement, I describe my objectives as a teacher and mentor, the tenets that I follow to meet these objectives, how I translate these tenets to action, and the feedback mechanisms that I employ to improve upon my teaching. I then briefly discuss my contributions to student mentoring.

My objectives as a teacher are:

1. To instill in my students the idea that decision- and policy-making problems can be solved with mathematical rigor and that data, models, and algorithms can help make better decisions.

2. To equip my students with the quantitative and qualitative skills necessary to understand and conduct data-driven analyses in order to identify real-wold decision-making problems, recommend effective solutions, and justify a course of action.

3. To provide my students with the confidence needed to think critically and creatively to independently generate novel solutions to problems and explore new application areas and to effectively communicate their ideas.

4. To inspire my students by immersing them in the full set of applications and domains where data-driven and algorithmic techniques can have a positive impact.

5. To increase awareness of my field among practitioners and to improve decision-making and policy-making in private and public sector organizations.

In order to meet my objectives, I rely on the following tenets, which have evolved from my experiences as a student, teaching assistant, lecturer, and assistant professor:

1. I believe that education should strive to connect theory to practice. Therefore, I always incorporate diverse, exciting, and socially important real-world examples, applications, and course projects into my classes. These serve to illustrate abstract concepts and highlight the relevance of the theory. I also invite domain experts and community partners to give guest lectures and seminars, thus situating the methods I teach the students in diverse and inspiring contexts.

2. I believe that education should seek and promote diversity. Thus, I leverage my research to infuse my classes with societal applications and ensuing issues surrounding equity, fairness, robustness, and interpretability, which I am convinced can broaden the appeal of our courses and attract a more diverse set of students, including those from underrepresented groups to our programs.

3. I believe that education should leverage and manage diversity. Therefore, I always promote class discussion. This helps me grasp the variety of skills and backgrounds of my students and enables me to tailor my teaching style to each individual and to the particular class composition. It also makes the class more interesting for everyone and facilitates the creation of new knowledge.

4. I believe that good education comes with ownership. Indeed, I am convinced that students only learn concepts in sufficient depth to be able to think critically and creatively if they take ownership of the analysis and solution of a real-world problem. For this reason, I always endeavor to have my students work on a project that they are fully responsible for, see figure below.

5. I believe that education should be rigorous and set demanding yet accomplishable intellectual challenges. Indeed, {by setting} high expectations for my students, I can create an environment that stimulates critical and creative thinking, thus promoting idea generation. In turn, I aim to provide the best possible support by being approachable in and out of class and by creating an atmosphere in which my students feel comfortable to ask questions.

6. I believe that technological education should be an inherent part of the curriculum. Indeed, technology plays a crucial role in today’s economies and most companies increasingly seek even managers that have practical experience in using software and computing to transform data into information and information into optimal decisions. In my classes, I thus emphasize the importance of technology and strive to incorporate lectures on software education. In addition, I provide exercises and projects that enable students to gain hands-on experience.

7. I believe that education does not stop in the classroom. Indeed, as experts in our field, we have the responsibility to educate our communities, organizations, and even government officials about the powers and perils of the technologies that we create.

This spring, I created, developed, and taught a new graduate level course based on my research titled ISE 599 Analytics for Social Impact. This course introduces students to this emerging research area whose goal is to advance descriptive, predictive, and prescriptive analytics to address important societal challenges, e.g., in public health, social services provision, and conservation. Particular methods that the students learn about relate to fairness, interpretability, and robustness, from the lenses of optimization, machine learning, and causal inference.

The following student quotes from my teaching evaluations at USC showcase how I take action on my tenets:

1. Connecting theory to practice:

   • "I loved this class! I finally saw how the math I've been learning for years can be applied to the real world.''

   • "Professor Vayanos knows the course material extremely well and is really knowledgeable and it was really interesting to hear about how she uses the course material in her research."

   • "She also provided many real world applications on how linear optimization can be used which was very cool. She was genuinely engaging, and passionate about the subject. She really cared for her students and wants all her students to do well."

2. Leverage and manage diversity:

   • "She could sense when students couldn't understand material [...]. In response, she would note other ways of understanding the material better and develop examples if needed."

   • "This dedication to students' learning provided a very open and conducive learning environment, where it was very much encouraged to ask questions and get full understanding."

   • "She really cares about the outcome of the course for students and uses very individual approach, which I can tell is really hard in class with a larger number of students."

3. Ownership:

   • "She did a great job keep students on track and involved with the learning process.''

   • "Prof. Vayanos is the most eager professor I've had at USC so far. She is the ONLY professor that knows my name. She is the ONLY professor that motivates me to get engaged in the class. I can definitely say that she is one of the best professors I've had at USC."

   • "The final project was something actually applicable to a life concept which I thought was very cool and useful in my future!"

4. Rigorous and demanding, yet accomplishable:

   • "A great, difficult class that challenged me and pushed me to learn about a topic that will help me in my career down the line."

   • "The class was always engaging and challenged me to think on my feet as well as really think through the problems presented."

   • "I really liked how motivating questions and purpose of questions were emphasized throughout the course, so that \ul{we could begin developing intuition and thinking regarding proofs and the material."

   • "Professor Vayanos was always excited about teaching OR and taught something new and challenging everyday. Her homework was organized and taught the material very well."

   • "She is very interested in her course and I think this is the main strength, that coupled with great explanation abilities, makes this course one of the best I ever attended, especially considering technical courses."

5. Technological education:

   • "I liked programming problems and Julia, although I am not very experienced with coding."

Across all the courses I taught at USC, my average evaluation score as instructor is 4.62/5 (4.618/5 and 4.628/5 for graduate and undergraduate teaching, respectively).

I am a strong supporter of interdisciplinary work and have a track record of presenting in workshops that involve students and practitioners across fields, such as the Microsoft Research India Workshops on AI for Social Good, the Partnership on AI/IBM Explainability Workshop, and the philosophy-computer science Algorithmic Ethics Workshop. Finally, in 2021 I co-organized the CAIS Symposium on Equity and AI which brought together practitioners and researchers in engineering and social work to share research and open challenges, and to educate practitioners on the potential of OR and AI to address issues of equity and bias.

In the new course I developed as part of my NSF CAREER award, I hosted several guest lecturers from the social sciences who presented a problem and data from their domain (e.g., homelessness, suicide, substance use) and the students worked in groups throughout the semester, leveraging the techniques they learned in class, to address the problem posed in collaboration with the domain experts. The problems they tackled centered around algorithmic fairness, interpretability, and robustness to tackle these challenges. This gave the students direct practical experience into how the methods and tools learned in class (many of which were devised in my research group) could be used to tackle a real problem faced in the social sciences. The students have complete ownership of their work and several of these projects are likely to turn into publications over the summer.

Over the past few years, I have given invited tutorials on research in AI and OR for social good at top conferences in our field. Notably, I gave tutorials at the Canadian Operations Research Society (CORS) 2021 and at the joint INFORMS International/CORS 2022 Conference. I also gave a tutorial in the Masterclass of the International Conference on the Integration of Constraint Programming, Artificial Intelligence, and Operations Research (CPAIOR) in 2019 and at the Doctoral Consortium on Computational Sustaintability in 2017. Finally, I co-organized the CPAIOR Masterclass on "CP, AI, and OR for Social Good". Through these activities, I am helping raise awareness of the growing area of research on OR and AI for society, aiming to attract more students (in particular those from underrepresented groups) to this field and growing the workforce trained to address societal problems through our tools.

As teachers we must continuously adapt to remain aligned with the needs of the real world. For this reason, I use every opportunity to measure my performance and improve my teaching:

1. I believe that student reviews and feedback provide very constructive ways to better my teaching. For this reason, I have setup a biweekly anonymous feedback system for students to share their opinions on the course throughout the semester. I then adjust my teaching methods and style dynamically, in response to the students inputs, as highlighted by the following quotes drawn my teacher evaluation forms:

   • "Phebe also cares about student’s feedback, by sending out intermediate anonymous feedback forms biweekly throughout the semester. Results were discussed and used to actively make amends to both her own and the TA’s teaching. She was one of the most engaging professors I’ve had at USC."

   • "Phebe is a super approachable instructor. She is always there to help students either through email or office hour. She tries her best to adjust her teaching with students' feedback."

2. I believe that tracking student progress throughout the course of the semester (in class, through homework, during office hours, recitations, etc.) is a very good way of measuring my teaching effectiveness and identifying improvement strategies.

3. I believe that teaching assistants are also a great source of constructive feedback. For example, when I started teaching the Linear Programming class, I held regular meetings with the courses TA and used their observations and feedback to improve the design of the course and to develop my class management skills.

University of Southern California, Los Angeles, CA, 2015-present

ISE 599 Analytics for Social Impact, 2022-present

• Created, developed, and taught the course

• Development of this course was supported by my NSF CAREER award

• This course, which is tightly coupled with my research, introduces students to the emerging area of research on Analytics for Social Impact whose goal is to advance descriptive, predictive, and prescriptive analytics to address important societal challenges. It exposes the students to applications in e.g., public health, social services provision, and conservation. Particular methods that the students learn about relate to fairness, interpretability, and robustness, from the lenses of optimization, machine learning, and causal inference.

ISE631 Linear Programming, 2015-present

• Developed and taught the course

• This is a graduate level class corresponding to the first doctoral course in the field of optimization that serves as the foundation for all subsequent courses in the area

• After six iterations, the class now regularly involves students from Computer Science, Industrial \& Systems Engineering, Electrical Engineering, and from the Business School.

Note: At USC, 500 level courses and above are graduate level courses. All courses for which I was the main instructor at USC involved 150 minutes per week of in classroom contact hours and one and a half hours per week of office hours on average over 15 weeks, for a total of 60 contact hours per course on average.

University of Southern California, Los Angeles, CA, 2015-present

Instructor, Viterbi School of Engineering

• ISE599 Analytics for Social Impact, Spring 2022 (18 students)

• ISE330 Introduction to Operations Research: Deterministic Models, Spring 2022 (25 students)

• ISE330 Introduction to Operations Research: Deterministic Models, Spring 2021 (37 students)

• ISE631 Linear Programming, Fall 2020 (9 students)

• ISE330 Introduction to Operations Research: Deterministic Models, Spring 2020 (30 students)

• ISE631 Linear Programming, Fall 2019 (11 students)

• ISE330 Introduction to Operations Research: Deterministic Models, Spring 2019 (36 students)

• ISE631 Linear Programming, Fall 2018 (10 students)

• ISE330 Introduction to Operations Research: Deterministic Models, Spring 2018 (20 students)

• ISE631 Linear Programming, Fall 2017 (16 students)

• ISE330 Introduction to Operations Research: Deterministic Models, Spring 2017 (18 students)

• ISE631 Linear Programming, Fall 2016 (8 students)

• ISE631 Linear Programming, Fall 2015 (9 students)

Guest Lecturer, Viterbi School of Engineering

• ISE105 Introduction to Industrial \& Systems Engineering, Fall 2021

• CSCI697 Seminar in Computer Science Research, Spring 2021

• CSCI697 Seminar in Computer Science Research, Fall 2019

• CSCI697 Seminar in Computer Science Research, Fall 2018

• CSCI599 Artificial Intelligence for Social Good, Spring 2017

• ISE105 Introduction to Industrial \& Systems Engineering, Fall 2016

• ISE105 Introduction to Industrial \& Systems Engineering, Fall 2015

• ENGR102 Engineering Freshman Academy, Fall 2015

MIT Sloan School of Management, Cambridge, MA, 2014

Instructor, Operations Research & Statistics Group

• 15.093 Optimization Methods, Fall 2014 (68 students)

Guest Lecturer and Teaching Assistant, {\em Operations Research Center}

• 15.094 Robust Modeling, Optimization & Computation, Spring 2014 (20 students)

Massachusetts Institute of Technology, Cambridge, MA, 2013

Guest Lecturer, Department of Chemical Engineering

• 10.557 Mixed-Integer and Non-Convex Optimization (~30 students)

Imperial College London, London, UK, 2009-2011

Guest Lecturer and Teaching Assistant, Department of Computing

• CO343 Operations Research (~80 students)a

Teaching Assistant, Department of Computing

• CO422 Computational Finance (~130 students)