Abstract
In the paper «Machine Learning (ML) – A New Kind of Cultural Tool? A 'Recontextualisation' Perspective on Machine Learning + Interprofessional Learning», I argued that
- (a) Machine Learning (ML) is a cultural tool containing a model of learning which enables it perceive patterns in data,
- (b) this learning ML is a more circumscribed kind of learning compared with how that concept has been interpreted in sociocultural theory
- (c) but nevertheless ML is further extending and distributing the complex relationship between human and machine cognition and learning.
That paper was however unable to address an extremely important issue about the ML model of learning namely the assumptions on which it was predicated, and the extent to which they were similar or different from the assumptions underpinning sociocultural theories of learning.
To address this issue, this paper firstly, explicates the assumptions that underpin Connectionism, the theoretical tradition informing ML algorithmic model of learning, and argues that they share with Vygotsky a conception of cognition as an internalisation-externalisation process.
Secondly, depeens this contention by arguing that neural networks do not replicate the operation of the brain, but instead perform a slight of hand and recontextualise cultural practices represented by visual textual or numerical data as neurologically generated patterns.
Thirdly, explains the reason that ML is involved with an inferential recontextualises process, rather than a neurological classification process, is because inferential activity is always more affected by experience and external factors than by internal physiological circuits (hence the ‘famous’ ML bias issue).
The paper concludes by reaffirming why human and machine learning is best viewed therefore as a process of continuous recontextualization.
About Dr. David Guile
Dr. David Guile is Professor in Education and Work and Co-Director of the Centre of Engineering Education and the Post 14 Centre for Education and Work, UCL.
David is an interdisciplinary Practice theorist who draws on Cultural Anthropology, Cultural Psychology and Cultural Sociology, to explore conceptually and empirically the changing relationship between work, technology (in particular Machine Learning) and education in the fields of Professional, Vocational and Workplace Learning.
With Professor (Emeritus) David Livingstone, University of Toronto, he co-edits the Brill Series The Knowledge Economy and Education.
Organisers and contact information
The seminar is organised by the HEDWORK and LiDA research groups
The seminar lecture is open to anyone interested. You do not have to register upfront.
If you have any questions, please contact Professor Monika Nerland.