Papers › Intriguing properties of generative classifiers

Intriguing properties of generative classifiers

28 Sep 2023arXiv:2309.16779archive 2025-07-28

Priyank Jaini, Kevin Clark, Robert Geirhos

What is the best paradigm to recognize objects -- discriminative inference (fast but potentially prone to shortcut learning) or using a generative model (slow but potentially more robust)? We build on recent advances in generative modeling that turn text-to-image models into classifiers. This allows us to study their behavior and to compare them against discriminative models and human psychophysical data. We report four intriguing emergent properties of generative classifiers: they show a record-breaking human-like shape bias (99% for Imagen), near human-level out-of-distribution accuracy, state-of-the-art alignment with human classification errors, and they understand certain perceptual illusions. Our results indicate that while the current dominant paradigm for modeling human object recognition is discriminative inference, zero-shot generative models approximate human object recognition data surprisingly well.

PaperPDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2309.16779")

Code

Syntology Ran 10 of 12 code samples harvested from 1 repository linked to this paper; 2 have no recorded run. Of those that ran: 1 ran · violated contract; 1 ran · our draft was wrong; 8 ran with no contract checked.

By repository: community (archive-listed): 12 samples from 1 repository, 10 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

SamsungSAILMontreal/ForestDiffusion mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

12 samples harvested; 10 ran; 0 honoured the contract we drafted; 2 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · violated contract
1ran · our draft was wrong
8ran
2unverified

Licence: 12 of the 12 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from SamsungSAILMontreal/ForestDiffusion. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

converged SamsungSAILMontreal/ForestDiffusion/softimpute.py community (archive-listed) ran no licence file found · pointer only · e66a901d3f6cb388 · report
get_optimizer SamsungSAILMontreal/ForestDiffusion/STaSy/losses.py community (archive-listed) ran no licence file found · pointer only · e311e63553012e76 · report
nanmean SamsungSAILMontreal/ForestDiffusion/utils.py community (archive-listed) ran · our draft was wrong fingerprinted no licence file found · pointer only · bbbb6c22bd6a9a4a · report
normalization SamsungSAILMontreal/ForestDiffusion/gain.py community (archive-listed) ran no licence file found · pointer only · 82b1e0d73f4df6cd · report
optimization_manager SamsungSAILMontreal/ForestDiffusion/STaSy/losses.py community (archive-listed) ran no licence file found · pointer only · 18269deb82a2a3f6 · report
pick_epsilon SamsungSAILMontreal/ForestDiffusion/utils.py community (archive-listed) ran no licence file found · pointer only · 127379550e5e9139 · report
renormalization SamsungSAILMontreal/ForestDiffusion/gain.py community (archive-listed) ran no licence file found · pointer only · f4957252157d03ef · report
softimpute SamsungSAILMontreal/ForestDiffusion/softimpute.py community (archive-listed) ran no licence file found · pointer only · 81dbfd5afe49f6a9 · report
str2bool SamsungSAILMontreal/ForestDiffusion/script_generation.py community (archive-listed) ran · violated contract no licence file found · pointer only · 7c508037b40522af · report
test_x SamsungSAILMontreal/ForestDiffusion/softimpute.py community (archive-listed) ran no licence file found · pointer only · bec9ca3c9ff3e822 · report
gain SamsungSAILMontreal/ForestDiffusion/gain.py community (archive-listed) unverified no licence file found · pointer only · c3b530881a1f5071 · report
quantile SamsungSAILMontreal/ForestDiffusion/utils.py community (archive-listed) unverified no licence file found · pointer only · 4db664a683098ffe · report

Tasks

Object Recognition

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Object Recognition shape bias Imagen shape bias 98.7 #1 of 18 Archive leaderboard report
Object Recognition shape bias Stable Diffusion shape bias 92.7 #2 of 18 Archive leaderboard report
Object Recognition shape bias Parti shape bias 91.7 #3 of 18 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections