Building Out a Sociotechnical Systems and Control Paradigm for Designing and Governing Algorithmic Systems

Dr.ir. Roel Dobbe

Assistant Professor, Engineering Systems and Services Department, Delft University of Technology

Seminar Information

Seminar Series
Dynamic Systems & Controls

Seminar Date - Time
April 19, 2023, 3:00 pm
-
4 PM

Seminar Location
SME 248, ASML Conference Center Room


Abstract

Data-driven, algorithmic and intelligent systems are informing, mediating and automating increasingly more parts of our daily lives as well as of our public infrastructures, services and democratic processes. Opportunities abound, ubiquitous experimentation has led to many emerging forms of undesirable and sometimes harmful system outcomes. In an effort to address algorithmic harms and injustices, a plethora of technical, ethical and policy efforts has been proposed. However, while most efforts have their merit, the actors developing or applying these lack an integral understanding of how all these instruments and solutions collectively contribute to outcomes that are safe, responsible, just, sustainable, etc.

In the STEAD Systems Lab at Delft University of Technology, we ask what it means to properly design, operate and govern algorithmic systems, and how we can ensure that the needs and interests of those subject to the system, be they workers or citizens, are taken into account in legitimate and effective ways. We hence study systems, but also design, policy and management processes, and the institutional and sociopolitical mechanisms involved in those.

Systems and control theory provides us with valuable lessons, principles and ways of understanding and designing for the complexity associated with the above challenges. In this talk, I aim to give you an overview of these lessons and how these and my training as an interdisciplinary systems and control scholar are shaping a new research, engineering and governance paradigm for algorithmic systems, with applications in (mostly) public administration, and (emerging) energy systems and healthcare.

Speaker Bio

Roel Dobbe received a BSc in Mechanical Engineering (2007) and a MSc in Systems & Control (2010) from TU Delft. During his masters, he conducted thesis research at UC Berkeley in the Hybrid Systems Lab, under the guidance of Professor Alessandro Abate and Professor Claire Tomlin.

From 2013 to 2018, he completed a PhD degree in Electrical Engineering & Computer Sciences at UC Berkeley, under the guidance of Professor Claire Tomlin in the Hybrid Systems Group and co-advised by Duncan Callaway in the Energy & Resources Group. From 2018 to 2020, he was an inaugural postdoc with the AI Now Institute at New York University.

His main interests are in modernizing critical infrastructure and sensitive decision-making through developing data-driven tools for analysis and control that promote safe, sustainable and democratically just societal systems.

In addition, Roel enjoys improving organizational culture and standards around employee wellness. While at Berkeley, he started a student-run organization to stimulate cross-disciplinary and engaged scholarship around technology and its societal implications, called Graduates for Engaged and Extended Scholarship in and around Engineering (GEESE).

After graduating from Delft and before starting my PhD, Roel gained experience in industry and the public sector. First, he participated in the Nationale DenkTank where he worked on trust and citizen participation in public institutions. Consecutively, he was a management consultant with A.T. Kearney at their Amsterdam office - working on a variety of projects for utility, healthcare and financial organizations. In 2016, Roel worked as a Data Scientist at C3 IoT in Silicon Valley, helping them to deliver better machine learning products to their energy customers (utilities and providers), by developing tools to increase interpretability and diagnosis and aiding the launch of platform tools with which end users can do data science and machine learning without explicit programming. In 2017, Roel was a Research Affiliate at Lawrence Berkeley National Labs in the Grid Integration Group with Daniel Arnold, working on state estimation and decentralized learning for distribution grids.