REVIEW 6 cited by
On the "steerability" of generative adversarial networks
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
An open secret in contemporary machine learning is that many models work beautifully on standard benchmarks but fail to generalize outside the lab. This has been attributed to biased training data, which provide poor coverage over real world events. Generative models are no exception, but recent advances in generative adversarial networks (GANs) suggest otherwise - these models can now synthesize strikingly realistic and diverse images. Is generative modeling of photos a solved problem? We show that although current GANs can fit standard datasets very well, they still fall short of being comprehensive models of the visual manifold. In particular, we study their ability to fit simple transformations such as camera movements and color changes. We find that the models reflect the biases of the datasets on which they are trained (e.g., centered objects), but that they also exhibit some capacity for generalization: by "steering" in latent space, we can shift the distribution while still creating realistic images. We hypothesize that the degree of distributional shift is related to the breadth of the training data distribution. Thus, we conduct experiments to quantify the limits of GAN transformations and introduce techniques to mitigate the problem. Code is released on our project page: https://ali-design.github.io/gan_steerability/
Forward citations
Cited by 6 Pith papers
-
Few-Shot Inspired Generative Zero-Shot Learning
FSIGenZ synthesizes a small set of semantic prototypes per unseen class and trains a contrastive classifier with semantic regularization, achieving competitive ZSL accuracy on SUN, AwA2, and CUB with far fewer synthet...
-
Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesis
Conventional SGD-trained networks can match the output of evolved networks but with internally fragmented and entangled representations, which the authors hypothesize may harm generalization, creativity, and continual...
-
Understanding Design Fixation in Generative AI
Generative AI models exhibit a design fixation phenomenon that limits the diversity and originality of their design outputs, according to a small lab study and a proposed theoretical framework.
-
Continual Learning of Personalized Generative Face Models with Experience Replay
A convex hull based experience replay method reduces catastrophic forgetting in continually updating personalized face GANs, with a new five-celebrity benchmark.
-
ThematicPlane: Bridging Tacit User Intent and Latent Spaces for Image Generation
A new interface lets image editors navigate themes as a plane rather than prompt text, with an exploratory six-person study suggesting creative flow but weak predictability.
-
Functional Risk Minimization
FRM replaces output-space losses with function-space losses, fitting a per-data-point function and approximating the resulting objective with Taylor/Laplace expansions, yielding weighted least squares with a Jacobian-...
Discussion (0). Continue with ORCID to comment.