github: robinsonkwame email: kwamepr@umich.edu

linkedin: kporterrobinson twitter: @kporterrobinson

About

Kwame is a socio-technical action researcher. Multimodal machine learning coupled with mixed-method system designs is how he often takes action with others. Sustainable social, economic, and environmental relationships with human AI interaction are his areas of inquiry. His three current interrelated research projects are Computing for Community Economies, Sustainability Assessment, and Social Justice in Human AI Interaction, particularly in robotics.

In his Computing in Community Economics area, Kwame is working under PI Ron Eglash on a NSF funded grant “Race, Gender and Class Equity in the Future of Work: Automation for the Artisanal Economy.” In collaboration with a diverse ecosystem of artisanal enterprises, this project combines artificial intelligence with a Science and Technology Studies framework, Generative Justice, to enhance wealth equity through experiments with bottom up social technologies to bolster collective production, supply chain making, and other dimensions of work.

Kwame’s second project is Sustainability Assessment. The manufacturing of products affects the environment, changes how people socialize, and determines who profits. Wary of negative impacts, individuals, corporations and governments often refer to top-down assessments developed by knowledgeable institutions, such as green certifications. In other cases, an individual must make a bottom up assessment alone. This project investigates the intersection of top-down and bottom-up strategies: bridged product assessment. Through the use of disparate data, interpersonal frameworks, and artificial intelligence technologies, this work seeks to combine different ways of understanding and assessing impact. Current work includes layering personal sustainability frameworks over product listings generated by search engines and learning how to decompose products into their constituent materials and manufacturing processes.

The final interrelated area, Social Justice for Human AI interaction (HAI), determines ways that interaction can be proactive instead of reactive. In this area, Kwame focuses on the problem of human consent for machine learning inference and trust repair strategies after negative HAI interactions. Examples include recent work on justice as trust repair and trustless models of machine learning consent.

In his research, he develops interdisciplinary theory within case-based work across various problem domains. His primary investigative techniques are systems building, field studies, and experimental designs. For evaluation he largely uses quantitative and mixed-methods methods. Prior to academia he worked for several years in industry as CEO, research engineer, data scientist and machine learning engineer for both government and private corporations. Kwame is a Ph.D. candidate at the University of Michigan in the School of Information. He is advised by Dr. Robert and Dr. Eglash.

Publications

Conference

C.5 Ron Eglash, Lionel Robert, Audrey Bennett, Kwame Robinson, Matthew Garvin, Deborah Hammond-Sowah (2022) “Navigating the open/closed spectrum: the need for layered access in platforms forgenerative justice”, ICA72

C.4 Robinson, K.P., Robert, L.P. Eglash, R. (2021). “Extrapolating significance of text-based autonomous vehicle scenarios to multimedia scenarios and implications for user-centered design”, RO-MAN 2021, 10.7302/1691

C.3 Robinson, K.P., Eglash, R., Bennett, A., Nandakumar, S. and Robert, L.P. (2020). “Authente-Kente: Enabling Authentication for Artisanal Economies with Deep Learning”, AI & Society, 10.13140/RG.2.2.27020.95362/2

C.2 Ron Eglash, Lionel P. Robert, Audrey Bennett, Kwame Porter Robinson, Michael L Lachney, William Babbitt. “Automation for the artisanal economy: enhancing the economic and environmental sustainability of crafting professions with human–machine collaboration”. September 2019. AI & Society. DOI: 10.1007/s00146-019-00915-w.

C.1 Ron Eglash, Lionel P. Robert, Audrey Bennett, Kwame Porter Robinson, Michael L Lachney, William Babbitt. “AI for a Generative Economy: The Role of Intelligent Systems in Sustaining Unalienated Labor, Environment, and Society”. August 2019. Conference: AAAI Fall 2019 Symposium on AI and Work At: Arlington, Virginia USA

Journal

J.1 Ron Eglash, Kwame Porter Robinson, Audrey Bennett, Lionel Robert, Matthew Garvin (2023; Under Review) “Computational Reparations as Generative Justice: Decolonial Transitions to Unalienated Circular Value Flow”. Big Data & Society

J.2 Kwame Porter Robinson, Ron Eglash, Lionel P. Robert, Audrey Bennett, Mark Guzdial (2023; Under Review) “Computing for Community-Based Economies: A Sociotechnical Ecosystem for Democratic, Egalitarian and Sustainable Futures”, The Information Society

Book Chapter

B.1 Ron Eglash, Audrey Bennett, Kwame Porter Robinson, Matthew Garvin, Lionel P. Robert, Mark Guzdial (2022). “Decolonization, Computation, Propagation: Phyto-human alliances in the pathways towards generative justice”. In Plants by Numbers, Bloomsbury Visual Arts.

B.2 Matthew Garvin, Ron Eglash, Kwame Porter Robinson, Lionel P. Robert, Audrey Bennett (2023). “Counter-hegemonic AI: The Role of Artisanal Identity in the Design of Automation for a Liberated Economy”. In Algorithms and Society, Taylor & Francis.

Unpublished and In-progress

Kwame Porter Robinson, Matthew Garvin, Ron Eglash, Lionel Robert, Mark Guzdial, Audrey Bennett “Making Exploratory Search Engines using Qualitative Case Studies: a content-aware mixed- methods implementation using interviews with Detroit Artisans” (In Progress)

UNP. Kwame Tacumah Porter-Robinson (Kwame Porter Robinson). “An Energy Efficient Rate Adaptive Distributed Source Coding Algorithm: RSWITCH”. December 2012. unpublished.

Conference Presentations

• 2023 Ethical AI Forum University of Michigan - From Extraction to Empowerment: Recent developments in Community-Based Computing (5/16/2023) 🏆 Press: UMSI News article

• RO-MAN 2021 - Extrapolating significance of text-based autonomous vehicle scenarios to multimedia scenarios and implications for user-centered design (8/10/2021)

🏆 - Recieved award for best student talk

Funding

• Rackham Merit Fellowship, stipend with full tuition
• Rackham Professional Development Grant, 2022 ($400)
• Rackham Graduate Student Research Grant, 2020 ($925)
• Rackham Conference Travel Grant, 2019 ($800)
• AWS Cloud Credits for Research, 2019 (~$1,000)

Industry Biography

Prior to becoming a Ph.D. candidate at the University of Michigan, Kwame was lead data scientist at (Brighthive) where he designed scalable natural language processing systems and algorithms for workforce artificial intelligence applications, including unstructured taxonomy matching and multi-level semantic similarity. In 2015 Kwame created and led a data science consultancy that served a variety of private and public organizations, including the WKKF Foundation and the World Bank. Additionally, Kwame has worked on classified projects spanning data science, blockchain, cyber security and telecommunications research for the Department of Defense.

Academic Biography

Kwame holds a master’s degree in Computer Science (University of Maryland, Baltimore County), with a thesis on Slepian-Wolf probabilistic source code correlation, a Bachelor’s degree in Electrical Engineering (New Mexico State University), with a specialization in control systems and a Bachelor’s of Fine Art (Boston University). Kwame is a Ph.D. candidate at the University of Michigan.

Ad-Hoc Review

• ACM/IEEE International Conference on Human-Robot Interaction 2024
• IEEE Transactions on Technology and Society
• AIS Transactions on Human-Computer Interaction
• New Media and Society Journal
• CHI’20, April 25–30, 2020
• ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO’22)

Graduate Student Instructor Courses

• SI 671/721: Data Mining: Methods and Applications - (98 students) Automatic, robust, and intelligent data mining techniques have become essential tools to handle heterogeneous, noisy, nontraditional, and large-scale data sets. This is a doctoral seminar course of advanced topics in data mining. The course provides an overview of recent research topics in the field of data mining, the state-of-the-art methods to analyze different genres of information, and the applications to many real world problems.
• SI 699: Big Data Analytics - (34 students) The big data analytics mastery course will require students to demonstrate mastery of data collection,processing, analysis, visualization, and prediction. To develop these skills students will work onsemester-long projects that deal with large or industry-scale data sets, and solve real-world problems.Aligned with best industry practices, students will be expected to work in a fast-paced, collaborativeenvironment, while demonstrating independence and leadership. Students must be able to create and usetools to handle very large transactional, text, network, behavioral, and/or multimedia data sets.

Service

• University of Michigan Doctoral Executive Committee (DEC) officer (2023 - 2024 Acacmdic year) - Canvas support for and help enact PhD programmatic changes benefiting UM School of Information Students through collaboration with adminstiration, faculty, and students. DEC is part of an organization, DSO, that includes all School of Information (SI) doctoral students. The mission of the DSO is to provide both academic and social support to all SI doctoral students.
• Advisory Search Committee, School of Information University of Michigan, Member - Appointed to the advisory search committee to support school search for next dean. Collaborated over multiple months, with search firm Heidrick & Struggles, School of Information faculty and staff, and contributed to committee recommendation to University Michigan Office of the Provost. Dr Andrea Forte recommended as School of Information dean.
• Kalamazoo Area Mathematics and Science Center (KAMSC) - Keynote for Big Data Day; Presented principles and questions for effective analysis of big data to Kalamazoo High School students. Introduced Information as a discipline, the importance of theory, and contextualized learnings with a brief introduction of Authene-Kente (Robinson et al 2021). Slides available here.
• Black at SI (B@SI) - One of several founding executive memebers. The Black@SI is an academic and social support network that serves Black students, alumni, and accomplices of the University of Michigan School of Information (UMSI). Through our programming, we help ensure that underrepresented students can successfully contribute to both scientific research and the college’s learning environment, and positively impact our communities.

Contact

Feel free to reach out to Kwame at kwamepr@umich.edu and he welcomes focused collaboration across a variety of disciplines.