When is Machine Learning Data Good?: Valuing in Public Health Datafication

Thakkar, Divy; Ismail, Azra; Kumar, Pratyush; Hanna, Alex; Sambasivan, Nithya; Kumar, Neha

doi:10.1145/3491102.3501868

Cited by 19 publications

(4 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These studies point to the discretionary, collaborative, and technologically-mediated human labour required to do data science [50,68,75,101]. Data science teams are made up of variously esteemed experts [68,81,95] with heterogeneous skill sets [48,58,82]. These practitioners collaborate with both one another as well as with counterparts beyond their organisational boundaries [51,77].…”

Section: Related Workmentioning

confidence: 99%

Designing Data Science Software for Social Care Organisations

Reinmund

2023

Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

Within the last few decades, data science has risen towards the top of agendas in public services such as adult social care. Concomitant with this ever-increasing appetite for data science is an expanding catalogue of challenges associated with developing, deploying, and maintaining data science software: choosing the appropriate, if any, analytical technique to use; balancing competing conceptions of success; and sustaining user adoption. In my thesis, I argue that these are sociotechnical challenges: issues that arise in the tension between what people do, and how data science software supports, limits, and, crucially, changes what they do. Through design science research, this thesis will develop, demonstrate, and evaluate a design approach for data science software that attempts to address such sociotechnical challenges. The empirical sites in which these research activities will take place are live data science projects with local government organisations responsible for adult social care services in England. The resultant approach will include a conceptual model for sociotechnical data science, process guidance and methods for applying the design approach, and results from applying the approach in a real-world data science project in collaboration with an industry partner. CCS CONCEPTS• Human-centered computing → Collaborative and social computing design and evaluation methods; • Software and its engineering → Requirements analysis.

show abstract

Section: Related Workmentioning

confidence: 99%

Designing Data Science Software for Social Care Organisations

Reinmund

2023

Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

show abstract

“…As noted by philosophers long ago, the concept of good is very subjective. In relation to ML, it was recently demonstrated that the understanding of "good data" varies considerably for different stakeholders [10]. The concept of quality, though more objective and operational, is domain specific [11] and multi-dimensional [12].…”

Section: Data Quality Control In MLmentioning

confidence: 99%

Quality of Labeled Data in Machine Learning: Common Sense and the Controversial Effect for User Behavior Models

Bakaev

Khvorostov

2023

Intels’22

View full text Add to dashboard Cite

show abstract

“…Users, developers, and outsourced workers carry out these tasks at any point in the development and deployment of AI systems. For example, medical professionals in the case of AI for healthcare [6,55,79], education professionals [50], or internet users when answering ReCAPTCHA tests [40]. This paper will employ the term "data work" to refer exclusively to the labor outsourced through crowdsourcing platforms and specialized business process outsourcing (BPO) companies, instead of the broader data work carried out by other professionals and users, while acknowledging the role of the former within the dispositif as requesters.…”

Section: Data Work For Machine Learningmentioning

confidence: 99%

The Data-Production Dispositif

Miceli¹,

Posada²

2022

Preprint

View full text Add to dashboard Cite

Machine learning (ML) depends on data to train and verify models. Very often, organizations outsource processes related to data work (i.e., generating and annotating data and evaluating outputs) through business process outsourcing (BPO) companies and crowdsourcing platforms. This paper investigates outsourced ML data work in Latin America by studying three platforms in Venezuela and a BPO in Argentina. We lean on the Foucauldian notion of dispositif to define the data-production dispositif as an ensemble of discourses, actions, and objects strategically disposed to (re)produce power/knowledge relations in data and labor. Our dispositif analysis comprises the examination of 210 data work instruction documents, 55 interviews with data workers, managers, and requesters, and participant observation. Our findings show that discourses encoded in instructions reproduce and normalize the worldviews of requesters. Precarious working conditions and economic dependency alienate workers, making them obedient to instructions. Furthermore, discourses and social contexts materialize in artifacts, such as interfaces and performance metrics, limiting workers' agency and normalizing specific ways of interpreting data. We conclude by stressing the importance of counteracting the data-production dispositif by fighting alienation and precarization, and empowering data workers to become assets in the quest for high-quality data.

show abstract

When is Machine Learning Data Good?: Valuing in Public Health Datafication

Cited by 19 publications

References 63 publications

Designing Data Science Software for Social Care Organisations

Designing Data Science Software for Social Care Organisations

Quality of Labeled Data in Machine Learning: Common Sense and the Controversial Effect for User Behavior Models

The Data-Production Dispositif

Contact Info

Product

Resources

About