English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Paper

Pointwise Representational Similarity

MPS-Authors
/persons/resource/persons290873

Kolling,  Camila
Group K. Gummadi, Max Planck Institute for Software Systems, Max Planck Society;

/persons/resource/persons231773

Speicher,  Till
Group K. Gummadi, Max Planck Institute for Software Systems, Max Planck Society;

/persons/resource/persons246687

Nanda,  Vedant
Group K. Gummadi, Max Planck Institute for Software Systems, Max Planck Society;

/persons/resource/persons281834

Toneva,  Mariya
Group M. Toneva, Max Planck Institute for Software Systems, Max Planck Society;

/persons/resource/persons144524

Gummadi,  Krishna
Group K. Gummadi, Max Planck Institute for Software Systems, Max Planck Society;

External Resource
No external resources are shared
Fulltext (restricted access)
There are currently no full texts shared for your IP range.
Fulltext (public)

arXiv:2305.19294.pdf
(Preprint), 5MB

Supplementary Material (public)
There is no public supplementary material available
Citation

Kolling, C., Speicher, T., Nanda, V., Toneva, M., & Gummadi, K. (2023). Pointwise Representational Similarity. Retrieved from https://arxiv.org/abs/2305.19294.


Cite as: https://hdl.handle.net/21.11116/0000-000D-6B93-5
Abstract
With the increasing reliance on deep neural networks, it is important to
develop ways to better understand their learned representations. Representation
similarity measures have emerged as a popular tool for examining learned
representations However, existing measures only provide aggregate estimates of
similarity at a global level, i.e. over a set of representations for N input
examples. As such, these measures are not well-suited for investigating
representations at a local level, i.e. representations of a single input
example. Local similarity measures are needed, for instance, to understand
which individual input representations are affected by training interventions
to models (e.g. to be more fair and unbiased) or are at greater risk of being
misclassified. In this work, we fill in this gap and propose Pointwise
Normalized Kernel Alignment (PNKA), a measure that quantifies how similarly an
individual input is represented in two representation spaces. Intuitively, PNKA
compares the similarity of an input's neighborhoods across both spaces. Using
our measure, we are able to analyze properties of learned representations at a
finer granularity than what was previously possible. Concretely, we show how
PNKA can be leveraged to develop a deeper understanding of (a) the input
examples that are likely to be misclassified, (b) the concepts encoded by
(individual) neurons in a layer, and (c) the effects of fairness interventions
on learned representations.