Papers › Toward Understanding Adversarial Distillation: Why Robust Teachers Fail

Toward Understanding Adversarial Distillation: Why Robust Teachers Fail

21 May 2026ICMLarXiv:2605.21999added by Syntology

Hongsin Lee, Hye Won Chung

Title, abstract, authors and date from arXiv's metadata (CC0); this paper is not in the Papers with Code archive (frozen 2025-07-28).

Adversarial Distillation aims to enhance student robustness by guiding the student with a robust teacher's soft labels within the min-max adversarial training framework, yet its success is notoriously inconsistent: a more robust teacher often fails to improve, or even harms, the student's robust generalization. In this paper, we identify a key mechanism of this teacher dependency: the misalignment between the teacher's supervisory confidence and the student's representational limitations on a consistent subset of training data -- the Robustly Unlearnable Set. We present a theoretical framework analyzing the feature learning dynamics of a two-layer neural network, demonstrating that this mismatch creates a dichotomy in distillation outcomes. We prove that when a teacher provides confident supervision on unlearnable samples, it compels the student to memorize spurious noise patterns that eventually overpower the learned robust signal, thereby driving robust overfitting. Conversely, a teacher that exhibits high uncertainty on these samples effectively suppresses noise memorization, allowing the student to rely solely on the learnable signal for robust generalization. We empirically validate our theory across both synthetic simulations and real-image classification datasets, confirming that robust overfitting is driven by the teacher's interaction with unlearnable samples. Finally, we demonstrate that a teacher's predictive entropy on unlearnable samples serves as a strong indicator of student robustness, validating our theoretical framework and offering a principled guideline for robust teacher selection.

PaperPDFCode 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="2605.21999")

Code

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

By repository: found in paper text by Syntology: 10 samples from 1 repository, 8 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

HongsinLee/why-robust-teachers-fail found in paper text by Syntology report

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

10 samples harvested; 8 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.

8ran
2unverified

Licence: 10 of the 10 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 HongsinLee/why-robust-teachers-fail. “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.

PGD HongsinLee/why-robust-teachers-fail/real_exp/attacks.py found in paper text by Syntology ran no licence file found · pointer only · a648902a7268d3fd · report
TRADES HongsinLee/why-robust-teachers-fail/real_exp/attacks.py found in paper text by Syntology ran no licence file found · pointer only · 1de719bcbbc0bd3c · report
pResNet18 HongsinLee/why-robust-teachers-fail/real_exp/cifar100_models/preresnet.py found in paper text by Syntology ran no licence file found · pointer only · f5088fe40e624080 · report
pResNet34 HongsinLee/why-robust-teachers-fail/real_exp/cifar100_models/preresnet.py found in paper text by Syntology ran no licence file found · pointer only · fd5e91ae30144db3 · report
pResNet50 HongsinLee/why-robust-teachers-fail/real_exp/cifar100_models/preresnet.py found in paper text by Syntology ran no licence file found · pointer only · c26044835e29342e · report
resnet18 HongsinLee/why-robust-teachers-fail/real_exp/cifar100_models/resnet.py found in paper text by Syntology ran no licence file found · pointer only · 804dc4361a30432d · report
rslad_inner_loss HongsinLee/why-robust-teachers-fail/real_exp/attacks.py found in paper text by Syntology ran no licence file found · pointer only · 47e8b2f8a9d376b2 · report
wideresnet HongsinLee/why-robust-teachers-fail/real_exp/cifar10_models/wideresnet.py found in paper text by Syntology ran no licence file found · pointer only · 95a525f1ac88c43c · report
svhn_wideresnetwithswish HongsinLee/why-robust-teachers-fail/real_exp/cifar100_models/wideresnetswish_svhn.py found in paper text by Syntology unverified no licence file found · pointer only · 6e78bd6eb54f58a0 · report
ti_wideresnetwithswish HongsinLee/why-robust-teachers-fail/real_exp/cifar100_models/wideresnetswish.py found in paper text by Syntology unverified no licence file found · pointer only · 414b9f77b95e1b72 · report

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2605.21999, was issued in May 2026, after that date, so the archive has no leaderboard rows for it.

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