Pith. sign in

REVIEW 4 major objections 8 minor 6270 cited by

Adam: A Method for Stochastic Optimization

T0 review · 4 major / 8 minor · reviewed 2026-05-09 · claude-opus-4-7

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Adam sets per-parameter step sizes from bias-corrected running averages of the gradient and its square, giving a robust default optimizer for noisy, high-dimensional problems.

desk verdict Adam: the algorithm and defaults are the contribution; the regret proof has a real gap, but it doesn't touch what made this paper matter. read the letter →

arxiv 1412.6980 v9 pith:SLBDZ2FY submitted 2014-12-22 cs.LG

Diederik P. Kingma , Jimmy Ba This is my paper · ORCID
classification cs.LG MSC 90C1568T0590C25
keywords stochasticoptimizationadaptivelearningratefirst-ordermethodsmomentestimationbiascorrectiononlineconvexdeepAdaMax
checked against Cost.FunctionalEquationbilinear_family_forcedlaw_of_existence
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks: can a first-order optimizer with almost no tuning reliably train large, noisy, sparse-gradient problems? Its answer is Adam, which keeps two running averages per parameter — one of the gradient, one of the squared gradient — and divides the first by the square root of the second to set a per-coordinate step. A small but important detail is the explicit bias correction that undoes the zero-initialization of those averages, which matters most when the second-moment decay rate β₂ is set close to 1 to handle sparse gradients. The construction makes the effective step size invariant to gradient rescaling and approximately bounded by the user-chosen α, so α functions like a trust-region radius rather than a raw learning rate. Empirically the authors show that one default setting tracks or beats AdaGrad, RMSProp, SGD with Nesterov momentum, AdaDelta, and a quasi-Newton baseline on logistic regression, MLPs with dropout, and convnets, and they derive an infinity-norm variant (AdaMax) with an even cleaner update bound.

What carries the argument

The bias-corrected ratio m̂_t / √v̂_t, where m_t and v_t are exponential moving averages of g_t and g_t², and the corrections m̂_t = m_t/(1−β₁ᵗ), v̂_t = v_t/(1−β₂ᵗ) undo the zero-initialization bias. This ratio is gradient-scale invariant, behaves like a per-coordinate signal-to-noise ratio that automatically anneals near optima, and bounds the per-step parameter move by roughly α — turning the stepsize hyperparameter into something close to a trust-region radius.

What would settle it

Run Adam with the recommended defaults against well-tuned SGD-with-momentum, AdaGrad, and RMSProp on the same suite of problems (MNIST logistic regression and MLP, IMDB bag-of-words logistic regression, CIFAR-10 convnet, and a variational autoencoder). If Adam fails to match or beat them on training loss within the same wall-clock budget, or if removing the bias-correction terms does not visibly destabilize training when β₂ is close to 1, the central practical claim fails. For the regret claim, a convex online sequence on which Adam's iterates do not satisfy R(T)=O(√T) would falsify the theore

Watch

Extended reading notes

Core claim

The paper proposes Adam, a first-order stochastic optimizer that maintains two exponential moving averages per parameter — one of the gradient (first moment) and one of the squared gradient (second raw moment) — and uses their ratio, with an explicit bias correction for the zero-initialization of those averages, to set a per-parameter step size. The authors argue this combines the sparse-gradient handling of AdaGrad with the non-stationarity handling of RMSProp, while the effective per-step move in parameter space stays approximately bounded by the user-chosen stepsize α, giving the method a built-in trust-region feel. They claim a single set of defaults (α=0.001, β₁=0.9, β₂=0.999, ε=1e-8) w

Load-bearing premise

The regret proof leans on a quantity that grows monotonically along every coordinate as training proceeds; the paper asserts this without justification, and the bound only holds where that monotonicity actually holds.

Editorial extensions

If this is right

  • <parameter name="0">A practitioner can train a wide range of deep models with the same optimizer and the same defaults
  • removing learning-rate tuning as a first-order concern.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • <parameter name="0">Editorial: the regret proof's telescoping step requires √v̂_t/α_t to be non-decreasing along each coordinate
  • which is not generally true
  • later work has constructed simple convex counterexamples on which Adam diverges
  • so the O(√T) bound as stated should be read as suggestive rather than airtight
  • even though the empirical recipe survives unchanged.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Request a human review

A listed scientist reviews the paper for a fee and the review publishes here regardless of verdict. See the reviewers or get listed.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 8 minor

Summary. The paper introduces Adam, a first-order stochastic optimizer that maintains exponential moving averages of the gradient (m_t) and the squared gradient (v_t), applies bias-correction for the zero-initialization, and updates parameters by θ_t ← θ_{t-1} − α · m̂_t / (√v̂_t + ε). The authors motivate the update via a signal-to-noise interpretation, derive the bias-correction from the EMA recurrence, prove an O(√T) regret bound in the online convex setting (Theorem 4.1), present an L_∞-norm variant (AdaMax), and report experiments on logistic regression (MNIST, IMDB-BoW), MLPs (MNIST, with and without dropout), CNNs (CIFAR-10), and a VAE. Default hyperparameters (α=10⁻³, β₁=0.9, β₂=0.999, ε=10⁻⁸) are recommended and shown to be competitive with or better than SGD+Nesterov, AdaGrad, RMSProp, AdaDelta, and SFO.

Significance. If the algorithmic and empirical claims hold, Adam offers a practically important contribution: a simple, memory-light, scale-invariant adaptive optimizer with intuitive hyperparameters that performs robustly across convex and non-convex deep learning workloads. The bias-correction derivation in §3 is clean and useful in its own right (it cleanly explains an effect that earlier RMSProp-with-momentum variants get wrong for β₂ near 1), and the SNR/effective-step discussion in §2.1 gives a usable mental model for setting α. The AdaMax derivation (§7.1) is elegant and yields a particularly simple update with a tighter step bound |Δ_t| ≤ α. The empirical comparisons span enough model classes (logistic regression, fully-connected nets with/without dropout, CNNs, VAE) to support the robustness claim, and the bias-correction ablation in §6.4 is a genuinely informative experiment. The theoretical contribution (Theorem 4.1) is partial — see major comments — but the algorithmic and empirical case is strong.

major comments (4)
  1. [§4 / §10.1, Theorem 4.1 / 10.5] The regret proof contains a load-bearing step that is not justified. In the displayed bound at the top of p. 14, the sum ∑_{t=2}^T (θ_{t,i}−θ*_i)² (√v̂_{t,i}/α_t − √v̂_{t−1,i}/α_{t−1}) is replaced by (D²/(2α(1−β₁))) ∑_i √(T v̂_{T,i}). This telescoping is valid only if √v̂_{t,i}/α_t is non-decreasing in t for every coordinate i. With α_t = α/√t, the quantity is √(t·v̂_{t,i})/α, and since v̂_t is a bias-corrected EMA of g_t² it can strictly decrease whenever a coordinate sees a small gradient following a large one. The authors should either (i) state and justify a monotonicity assumption on v̂_t, (ii) carry through the proof with the absolute value of the increment (which changes the bound), or (iii) restrict the theorem to a class of sequences for which the monotonicity holds. As written, the bound is not established for general bounded convex sequences, and a one-dimensional counterexamp
  2. [§4, Theorem 4.1 statement] The hypothesis β₁²/√β₂ < 1 is stated but its role should be made explicit in the main text — it is used in Lemma 10.4 to bound an arithmetic-geometric series. With the recommended defaults β₁=0.9, β₂=0.999 one has β₁²/√β₂ ≈ 0.811, so the assumption is satisfied at defaults; however readers tuning β₁ upward (a common practice with momentum) can violate it. Please flag this in §4 alongside the theorem so that the regime of validity is clear.
  3. [§6.3, Figure 3] The CNN experiment reports that v̂_t 'vanishes to zeros after a few epochs and is dominated by the ε in algorithm 1', and that consequently 'Adagrad converges much slower than others' while Adam shows only 'marginal improvement over SGD with momentum'. This is an interesting and honest observation, but it slightly undercuts the central claim that adaptive second-moment scaling is the source of Adam's advantage. It would strengthen the paper to (a) report what fraction of coordinates have √v̂_t < ε at the cited epochs, and (b) show an ablation in which ε is varied, so readers can tell whether Adam in this regime is effectively SGD-with-momentum + a small constant preconditioner or whether the second moment still contributes.
  4. [§5, Related work / RMSProp comparison] The claim that lack of bias-correction in RMSProp 'leads to very large stepsizes and often divergence' for β₂ near 1 is supported by the VAE experiment in §6.4, but the comparison fixes architecture and dataset. Since this is one of the paper's main differentiators from RMSProp, a second setting (e.g., the MLP+dropout or CNN tasks already in the paper) showing the same effect would make the case substantially more robust.
minor comments (8)
  1. [Algorithm 1] The placement of ε inside the square root (√v̂_t + ε) versus inside (√(v̂_t + ε)) matters in practice and differs across implementations. Please state explicitly which convention is used and whether the analysis is affected.
  2. [§2.1] The two cases for the step bound, |Δ_t| ≤ α·(1−β₁)/√(1−β₂) versus |Δ_t| ≤ α, would be clearer with a one-line derivation rather than asserted. Currently the reader has to reconstruct the algebra.
  3. [§3, Eq. (4)] The term ζ is introduced and immediately argued to be small for stationary or slowly-varying gradients, but is not formally bounded. A short remark giving an explicit bound in terms of the variation of E[g_t²] would tighten the derivation.
  4. [§4] The decay schedule β_{1,t} = β₁·λ^{t−1} with λ very close to 1 is required for the proof but is not used in any of the experiments (which appear to use constant β₁=0.9). Please clarify whether the empirical performance corresponds to a regime covered by the theorem.
  5. [§7.1, Eq. (12)] It would be helpful to note that u_t = max(β₂·u_{t−1}, |g_t|) corresponds to a max over an exponentially-weighted history and therefore does not require bias correction, as briefly stated; an explicit derivation showing E[u_t] in the stationary case would parallel §3.
  6. [Lemma 10.3 proof] The inductive step uses the inequality √(a − b) ≤ √a − b/(2√a) which requires a ≥ b ≥ 0; this is fine but worth stating, since a = ∥g_{1:T,i}∥² and b = g_{T,i}² satisfy it by construction.
  7. [§6] The phrase 'searched over a dense grid' for hyperparameters of the baselines is not specific. Listing the grids (at minimum for α and momentum) in an appendix would improve reproducibility.
  8. [Typos] Several minor typos: 'theoratical' (§6.1, twice), 'BoW feature Logistic Regression' axis label, 'Initalization' (§7.2). 'β₁' appears where 'β₂' is meant in the sentence following Eq. (4) ('the exponential decay rate β₁ can be chosen…' — context is the second moment).

Simulated Author's Rebuttal

4 responses · 1 unresolved

We thank the referee for a careful and constructive report. The most substantive point — the unstated monotonicity assumption underlying the telescoping step in the regret proof — is correct, and we will revise Theorem 4.1 and its proof to state the assumption explicitly rather than leaving it implicit. We also agree to flag the β₁²/√β₂ < 1 hypothesis prominently in §4, to add quantitative support to the §6.3 discussion of v̂_t vanishing on CNNs (including an ε ablation), and to broaden the bias-correction comparison in §6.4 beyond the VAE setting. None of these revisions affect the algorithm itself, the bias-correction derivation in §3, the SNR discussion in §2.1, the AdaMax derivation in §7.1, or the empirical conclusions; they sharpen the theoretical statement and strengthen the empirical case. A point-by-point response follows.

read point-by-point responses

  1. Referee: The telescoping step in the regret proof (top of p. 14) replaces ∑_{t=2}^T (θ_{t,i}−θ*_i)² (√v̂_{t,i}/α_t − √v̂_{t−1,i}/α_{t−1}) by (D²/(2α(1−β₁))) ∑_i √(T v̂_{T,i}). This is only valid if √v̂_{t,i}/α_t is non-decreasing in t per coordinate, which need not hold for a bias-corrected EMA of g_t² when a small gradient follows a large one.

    Authors: We agree that this step requires an additional assumption that we did not state explicitly. The telescoping is valid when √(t·v̂_{t,i}) is non-decreasing in t for each coordinate, which is not guaranteed by a bias-corrected EMA of g_t² in general. We will revise §4 and §10.1 in two ways: (i) we will explicitly add the assumption that √(t·v̂_{t,i})/α is non-decreasing in t for all i (equivalently, that t·v̂_{t,i} is non-decreasing), and flag that this is what makes the telescoping well-defined; and (ii) we will note the alternative route in which the increment is replaced by its absolute value, which yields a weaker but unconditional bound. We thank the referee for catching this — the assumption is implicit in our derivation but should be made part of the theorem statement, and we will add a sentence describing the regime in which it is reasonable (sufficiently slowly-varying second-moment estimates) and acknowledging that pathological sequences can violate it. We do not claim a fix for the general non-monotone case in this revision. revision: yes

  2. Referee: The hypothesis β₁²/√β₂ < 1 should be flagged in the main text alongside Theorem 4.1, since users tuning β₁ upward can violate it (defaults satisfy it: 0.9²/√0.999 ≈ 0.811).

    Authors: We agree. We will add a short remark in §4 immediately after the theorem statement noting (i) the role of this assumption — it is used in Lemma 10.4 to bound an arithmetic-geometric series via γ = β₁²/√β₂ < 1 — (ii) that the recommended defaults β₁=0.9, β₂=0.999 give γ ≈ 0.811 and so satisfy it comfortably, and (iii) that practitioners increasing β₁ (e.g. β₁ ≥ 0.95 with default β₂) should check the inequality. We will also include a one-line worked example so the regime of validity is unambiguous. revision: yes

  3. Referee: In the CNN experiment (§6.3), the authors note v̂_t vanishes to near-zero so the update is dominated by ε, which somewhat undercuts the claim that adaptive second-moment scaling drives Adam's advantage. Please report what fraction of coordinates have √v̂_t < ε at the cited epochs, and add an ablation varying ε.

    Authors: This is a fair point and we agree the §6.3 discussion would benefit from quantitative support. In the revision we will add (a) a measurement, taken from the same CIFAR-10 run, of the fraction of coordinates with √v̂_t below ε (and below 10ε, 100ε) as a function of epoch, and (b) an ablation varying ε ∈ {10⁻⁴,10⁻⁶,10⁻⁸,10⁻¹⁰} to expose how much of Adam's behavior in this regime is attributable to the second-moment term versus an effectively constant preconditioner combined with the first-moment term. We will not retract the broader claim — on the logistic, MLP, and VAE experiments the second moment plainly contributes — but we will explicitly state that on CNNs of this size much of Adam's benefit over plain SGD with momentum comes from per-layer scale adaptation early in training and from the first-moment term, and that the improvement margin over well-tuned SGD+momentum is correspondingly modest. This nuance is consistent with what is already written in §6.3 but will be made quantitative. revision: yes

  4. Referee: The claim that absent bias-correction RMSProp diverges for β₂ near 1 is supported only by the VAE experiment (§6.4); a second setting would substantially strengthen the differentiator from RMSProp.

    Authors: We accept this. The bias-correction-vs-no-correction comparison is a central claim and one experiment is thinner than it should be. For the revision we will add a sweep over β₂ ∈ {0.99, 0.999, 0.9999} and α ∈ [10⁻⁵,10⁻¹], with and without the bias-correction terms, on the MLP+dropout MNIST setting from §6.2 (and, if space permits, on the CNN setting from §6.3). We expect — based on the analysis in §3, where the (1−β₂^t) factor is largest precisely when β₂ is near 1 — to reproduce the same instability pattern observed in §6.4. The resulting figure will be added as a panel to Figure 4 or as a new figure in §6.4. revision: yes

standing simulated objections not resolved
  • We do not have a proof of the O(√T) regret bound that dispenses with the monotonicity assumption on √(t·v̂_{t,i})/α. The revised theorem will therefore be conditional on this assumption; an unconditional bound for general bounded convex sequences with bias-corrected v̂_t is left to future work.

Circularity Check

0 steps flagged · score 1.0 of 10

No meaningful circularity: Adam's algorithm, bias correction, and empirical claims stand on independent content; the proof gap flagged by the reader is a correctness/soundness issue, not a circular derivation.

full rationale

Walking the derivation chain: (1) §2 Algorithm: defines Adam by EMAs of g and g². No claim is being "derived" from itself — the update rule is a definition. (2) §3 Bias correction: derives E[v_t] = E[g_t²]·(1−β_2^t) + ζ from the EMA recursion (Eq. 1–4). The (1−β_2^t) divisor is then read off this expectation. This is a straightforward algebraic identity, not a circular fit; nothing is fitted to data and then re-predicted. (3) §2.1 SNR / effective stepsize bounds: |Δ_t| ≤ α-style bounds follow from the algebra of m̂_t/√v̂_t. Independent content. (4) §4 / §10 Convergence: Theorem 4.1 derives an O(√T) regret bound from stated assumptions (bounded gradients, bounded iterate distance, β_1²/√β_2 < 1). The reader's concern is that the telescoping step in the proof of Theorem 10.5 implicitly assumes √(t·v̂_{t,i})/α is monotone non-decreasing — a soundness gap later exploited by Reddi et al. (2018). That is a *correctness* problem, not a circularity problem: the bound is not "the input renamed as the output" — it is an attempted proof from external assumptions that turns out to have an unjustified inequality. No quantity is fitted to the regret and then claimed as a prediction of the regret; no self-citation is load-bearing (the proof cites Zinkevich 2003's framework, not the authors' own prior work). (5) §5 Related work / §6 Experiments: comparisons to AdaGrad/RMSProp/SGD use independently implemented baselines on standard datasets (MNIST, IMDB, CIFAR-10). No fitted-input-as-prediction pattern. (6) §7 AdaMax: derived as the p→∞ limit of an L_p generalization (Eq. 6–12). Algebraic limit, not circular. There is essentially no self-citation load: the references are to Duchi, Tieleman & Hinton, Zeiler, Sohl-Dickstein, Zinkevich, etc. The Kingma & Welling (2013) self-cite is only used to specify the VAE architecture used as a *test problem* in §6.4 — it is not load-bearing for any theoretical claim. Conclusion: the paper's core claims are self-contained against external benchmarks and standard online-convex machinery. The Theorem 4.1 issue is a real bug in the proof (correctly diagnosed by the reader), but it is a missing-step / unjustified-monotonicity flaw, not a circular derivation. Score: 1 (one minor self-citation, not load-bearing).

Assumptions & free parameters 0 free parameters · 0 assumptions · 0 invented entities

The algorithm itself introduces no postulated entities. Its hyperparameters (α, β₁, β₂, ε) are user-settable knobs, not free parameters fitted to make a derivation work, though the recommended defaults were chosen empirically. The convergence theorem leans on standard online-convex-optimization assumptions plus one assumption that turned out to be false in general.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adam: A Method for Stochastic Optimization." pith.science (2026). https://pith.science/paper/SLBDZ2FY

@misc{pith202614126980,
  author       = {Pith},
  title        = {Pith review of: Adam: A Method for Stochastic Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SLBDZ2FY}},
  note         = {Machine review of arXiv:1412.6980}
}
read the original abstract

We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has little memory requirements, is invariant to diagonal rescaling of the gradients, and is well suited for problems that are large in terms of data and/or parameters. The method is also appropriate for non-stationary objectives and problems with very noisy and/or sparse gradients. The hyper-parameters have intuitive interpretations and typically require little tuning. Some connections to related algorithms, on which Adam was inspired, are discussed. We also analyze the theoretical convergence properties of the algorithm and provide a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Empirical results demonstrate that Adam works well in practice and compares favorably to other stochastic optimization methods. Finally, we discuss AdaMax, a variant of Adam based on the infinity norm.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

  • Foundation/PhiForcing.lean, Foundation/DimensionForcing.lean phi_equation, dimension_forced unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    Good default settings for the tested machine learning problems are α = 0.001, β₁ = 0.9, β₂ = 0.999 and ε = 10⁻⁸.

  • Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    The algorithm updates exponential moving averages of the gradient (m_t) and the squared gradient (v_t) where the hyper-parameters β₁, β₂ ∈ [0,1) control the exponential decay rates of these moving averages.

  • Foundation/DAlembert/Inevitability.lean bilinear_family_forced unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    R(T) ≤ D²/(2α(1−β₁)) Σ_i √(T·v̂_{T,i}) + α(1+β₁)G_∞/((1−β₁)√(1−β₂)(1−γ)²) Σ_i ‖g_{1:T,i}‖₂ + ...

  • Foundation/LawOfExistence.lean law_of_existence unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    We propose Adam, a method for efficient stochastic optimization that only requires first-order gradients with little memory requirement.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Showing 60 of 6,270 Pith papers that cite this

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. See all 6,270 Pith citations

  1. Zero-shot rib design: merging training-free generative prior with topology optimization

    cs.LG 2026-09 accept novelty 9.0 of 10

    A frozen text-to-image diffusion model, coupled via score distillation sampling to finite-element topology optimization, enables zero-shot text-guided rib design with statistically significant compliance reductions.

  2. Dynamic Observable Subspaces: Fast Simulation of Observable Quantum Dynamics

    quant-ph 2026-09 accept novelty 8.0 of 10

    A polynomial-time classical simulation method for quantum circuits with polynomially sized dynamic Lie algebras, based on restricting dynamics to the Dynamic Observable Subspace of each observable.

  3. Fourier-cosine Tensor Trains for Density Recovery and Expectation Calculation

    math.NA 2026-09 conditional novelty 8.0 of 10

    COS-TT-CHF compresses the characteristic function via tensor trains to achieve linear dimensional scaling for Fourier-based density recovery and expectation calculation.

  4. From Prior-Guided Heuristics to Deployable Agents: Accelerating Demonstration-Driven Reinforcement Learning for Deadline-Constrained Network Control

    cs.NI 2026-09 accept novelty 8.0 of 10

    A family of deadline-aware congestion metrics (Effective Congestion) and a demonstration-driven training protocol (MGA-RL) together reduce online interaction costs of deep RL network controllers by 7x while matching o...

  5. Guide, Not Bind: Why Defeasible Priors Fail in Augmented Lagrangian Causal Discovery

    cs.LG 2026-09 accept novelty 8.0 of 10

    Sequential penalty-ramping in Augmented Lagrangian causal discovery suppresses wrongly-forbidden true edges before counterfactual checks can act, and correlation matching exactly ties a true edge and its reverse, two ...

  6. Gradient-Update Mismatch: Rethinking Conflict-Free Training of Physics-Informed Neural Networks

    cs.LG 2026-09 accept novelty 8.0 of 10

    Gradient-Update Mismatch quantifies how optimizer transformations reintroduce gradient conflicts after gradient surgery, and Gradient-Update Alignment restores conflict-free updates by projecting the optimizer proposa...

  7. Self-OPD: On-Policy Distillation for Flow Matching Models without Teacher

    cs.CV 2026-08 accept novelty 8.0 of 10

    Self-OPD replaces the teacher model in on-policy distillation for flow matching with self-exploration and a self-reference baseline, achieving higher or comparable quality on single and mixed reward benchmarks.

  8. On MUON optimization: From non-convergence to an error analysis with Polar Express and the Newton-Schulz polynomial from implementations

    math.OC 2026-08 accept novelty 8.0 of 10

    For a one-dimensional quadratic stochastic optimization problem, MUON with Newton-Schulz steps provably fails to converge to the minimizer for all sufficiently large mini-batch sizes when the data is skewed, while a n...

  9. Automated discovery of high-probability heralded schemes for path-entangled states

    quant-ph 2026-07 conditional novelty 8.0 of 10

    A newly discovered modular-comb family of linear-optical circuits generates NOON states with success probabilities that grow exponentially (and super-exponentially) over previous passive schemes.

  10. Neural-Spectral Discovery of Rotating Black Holes Beyond General Relativity

    gr-qc 2026-07 conditional novelty 8.0 of 10

    A hybrid neural-spectral method constructs the first fully non-linear rotating black hole solutions in cubic Lovelock gravity, parametric in the gravitational coupling constants.

  11. GAIA: Geometry-Adaptive Operator Learning for Forward and Inverse Problems

    cs.LG 2026-07 conditional novelty 8.0 of 10

    GAIA introduces a geometry-adaptive integral autoencoder that unifies forward, boundary-value, and inverse PDE operator learning on arbitrary domains via geometry tokens and cross-attention.

  12. ShardNet: Training Neural Controllers with Hard, Non-Convex Constraints

    eess.SY 2026-06 unverdicted novelty 8.0 of 10

    ShardNet enforces non-convex polyhedral safety constraints in neural controllers by construction via a differentiable projection layer, achieving 100% verified safety and over 3x larger safe sets than prior methods on...

  13. Adam Converges in Nonsmooth Nonconvex Optimization

    math.OC 2026-06 unverdicted novelty 8.0 of 10

    The paper establishes the first finite-time convergence rate of 1/T^{2/13} for classical Adam (with bias correction, no extra steps) in nonsmooth nonconvex optimization under heavy-tailed noise with β1=β2.

  14. Efficient AI-Inspired Reduction of Feynman Integrals via Tube Seeding

    hep-ph 2026-06 unverdicted novelty 8.0 of 10

    Machine learning discovers a tube-seeding strategy for IBP reduction of Feynman integrals that scales linearly with numerator power, demonstrated on rank-20 2-loop 5-point integrals.

  15. Test-time Adversarial Takeover: A Real-time Hijacking Interface against Robotic Diffusion Policies

    cs.RO 2026-06 unverdicted novelty 8.0 of 10

    TAKO demonstrates real-time adversarial takeover of robotic diffusion policies via reusable universal patches on visual inputs, achieving 100% success in steering attacker-chosen trajectories across multiple tasks, en...

  16. Adaptive directional gradients for parameterised quantum circuits

    quant-ph 2026-06 unverdicted novelty 8.0 of 10

    Forward gradient framework for PQCs unifies SPSA and parameter-shift as limits, introduces QUIVER adaptive optimizer with closed-form measurement allocation, and demonstrates efficient training of 60-qubit circuits on...

  17. A multimodal dataset of photoplethysmography and continuous behavioral responses to ASMR and nature videos

    cs.LG 2026-05 unverdicted novelty 8.0 of 10

    Introduces REST-ASMR multimodal dataset of PPG, stimuli, and continuous annotations for ASMR research, validated with 97% responder rate, significant agreement, PPG deceleration, and BiLSTM achieving 75.51% frame-leve...

  18. Neutron Star Equation of State via Physics Informed Neural Network

    astro-ph.HE 2026-05 unverdicted novelty 8.0 of 10

    PINNs are used to non-parametrically infer the neutron star EOS from NICER and pulsar data, producing M_max = 2.06 M_sun, R_1.4 = 12.85 km, and a reproducible speed-of-sound softening at 2-4 rho_0 consistent with quar...

  19. Not All Inputs Are Valid: Towards Open-Set Video Moment Retrieval Using Language

    cs.CV 2026-05 unverdicted novelty 8.0 of 10

    OpenVMR uses normalizing flow to detect out-of-distribution queries and performs moment retrieval only on in-distribution queries.

  20. Canonical Regularisation of Wide Feature-Learning Neural Networks

    stat.ML 2026-05 unverdicted novelty 8.0 of 10

    Derives geodesic ridge regularization and Riemannian Gibbs Process prior for feature-learning wide neural networks, generalizing kernel-regime results via function-space axiomatization.

  21. Stochastic Non-Smooth Convex Optimization with Unbounded Gradients

    math.OC 2026-05 unverdicted novelty 8.0 of 10

    Introduces generalized Lipschitz class and shows clipped AdamW outperforms SGD and AdaGrad for stochastic convex optimization under this and related assumptions.

  22. ENSEMBITS: an alphabet of protein conformational ensembles

    cs.LG 2026-05 unverdicted novelty 8.0 of 10

    Ensembits is the first tokenizer of protein conformational ensembles that outperforms static tokenizers on RMSF prediction and matches them on function and mutation tasks while using less pretraining data.

  23. REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations

    cs.CL 2026-05 unverdicted novelty 8.0 of 10

    REALISTA optimizes continuous combinations of valid editing directions in latent space to produce realistic adversarial prompts that elicit hallucinations more effectively than prior methods, including on large reason...

  24. Online Learning-to-Defer with Varying Experts

    stat.ML 2026-05 unverdicted novelty 8.0 of 10

    Presents the first online learning-to-defer algorithm with regret bounds O((n + n_e) T^{2/3}) generally and O((n + n_e) sqrt(T)) under low noise for multiclass classification with varying experts.

  25. Spherical Boltzmann machines: a solvable theory of learning and generation in energy-based models

    cs.LG 2026-05 unverdicted novelty 8.0 of 10

    In the high-dimensional limit the spherical Boltzmann machine admits exact equations for training dynamics, Bayesian evidence, and cascades of phase transitions tied to mode alignment with data, which connect to gener...

  26. Convergent Stochastic Training of Multi-Headed Attention and Understanding LoRA

    cs.LG 2026-05 unverdicted novelty 8.0 of 10

    Attention and LoRA regression losses induce Poincaré inequalities under mild regularization, so SGD-mimicking SDEs converge to minimizers with no assumptions on data or model size.

  27. SLayerGen: a Crystal Generative Model for all Space and Layer Groups

    cond-mat.mtrl-sci 2026-05 unverdicted novelty 8.0 of 10

    SLayerGen generates crystals invariant to any space or layer group via autoregressive lattice and Wyckoff sampling plus equivariant diffusion, achieving gains over bulk models on diperiodic materials after correcting ...

  28. 3DSS: 3D Surface Splatting for Inverse Rendering

    cs.GR 2026-05 unverdicted novelty 8.0 of 10

    3DSS is the first differentiable surface splatting renderer that recovers shape, spatially-varying BRDF materials, and HDR illumination from multi-view images via a coverage-based compositing model derived from recons...

  29. Random test functions, $H^{-1}$ norm equivalence, and stochastic variational physics-informed neural networks

    math.NA 2026-05 unverdicted novelty 8.0 of 10

    Proves H^{-1} norm equivalence to expectation over random test functions and introduces SV-PINNs that outperform standard PINNs on eight elliptic problems.

  30. Characterizing the Expressivity of Local Attention in Transformers

    cs.CL 2026-05 unverdicted novelty 8.0 of 10

    Local attention strictly enlarges the class of regular languages recognizable by fixed-precision transformers by introducing a second temporal operator in LTL, with global and local attention being expressively complementary.

  31. STARE: Step-wise Temporal Alignment and Red-teaming Engine for Multi-modal Toxicity Attack

    cs.CR 2026-05 unverdicted novelty 8.0 of 10

    STARE uses step-wise RL to attack multimodal models, achieving 68% higher attack success rate while revealing that adversarial optimization concentrates conceptual toxicity early and detail toxicity late in the genera...

  32. Neural Spectral Bias and Conformal Correlators I: Introduction and Applications

    hep-th 2026-04 unverdicted novelty 8.0 of 10

    Simple feed-forward neural networks trained on crossing symmetry plus a single anchor value reproduce CFT correlators to percent-level accuracy, and the authors conjecture this works because physical correlators are t...

  33. MMGait: Towards Multi-Modal Gait Recognition

    cs.CV 2026-04 conditional novelty 8.0 of 10

    MMGait provides a new multi-sensor gait dataset and OmniGait baseline to support single-modal, cross-modal, and unified multi-modal person identification from walking patterns.

  34. Proton Structure from Neural Simulation-Based Inference at the LHC

    hep-ph 2026-04 unverdicted novelty 8.0 of 10

    Neural simulation-based inference on unbinned top-quark pair data at 13 TeV yields improved gluon PDF precision over traditional binned analyses while incorporating experimental and theoretical uncertainties.

  35. Deterministic Adam-Inspired Methods with Accelerated Convergence Rate

    math.OC 2026-04 unverdicted novelty 8.0 of 10

    Adam-HNAG is a splitting-based reformulation of Adam that yields the first convergence proof for Adam-type methods, including accelerated rates, in convex smooth optimization.

  36. CMCC-ReID: Cross-Modality Clothing-Change Person Re-Identification

    cs.CV 2026-04 unverdicted novelty 8.0 of 10

    The paper introduces the CMCC-ReID task, constructs the SYSU-CMCC benchmark dataset, and proposes the PIA network with disentangling and prototype modules that outperforms prior methods on combined modality and clothi...

  37. Why Adam Can Beat SGD: Second-Moment Normalization Yields Sharper Tails

    cs.LG 2026-03 unverdicted novelty 8.0 of 10

    Adam attains a δ^{-1/2} high-probability rate while any SGD guarantee must incur at least δ^{-1} dependence.

  38. The Geometric Mechanics of Contrastive Representation Learning: Alignment Potentials, Entropic Dispersion, and Cross-modal Divergence

    cs.LG 2026-01 unverdicted novelty 8.0 of 10

    Symmetric InfoNCE's intrinsic energy contains a negative symmetric KL coupling, making a population-level modality gap a generic geometric consequence rather than a training artifact.

  39. Statistical physics of deep learning: Optimal learning of a multi-layer perceptron near interpolation

    stat.ML 2025-10 conditional novelty 8.0 of 10

    A replica/HCIZ theory predicts the Bayes-optimal generalization error of proportional-width MLPs near interpolation and discovers layer-wise specialization transitions that make deeper targets harder to learn.

  40. Convergence Rates of Time Discretization in Extended Mean Field Control

    math.OC 2025-08 conditional novelty 8.0 of 10

    For linear-convex extended mean field control, piecewise constant controls approximate the optimal cost at rate 1/2 and the optimal control at rate 1/4; under smoothness, the rate improves to 1.

  41. Automated discovery of heralded ballistic graph state generators for fusion-based photonic quantum computation

    quant-ph 2025-08 unverdicted novelty 8.0 of 10

    A two-pass optimization framework with polynomial-based simulation discovers heralded ballistic circuits for 3-5 qubit graph states achieving up to 7.5x higher success probabilities than fusion baselines, including fi...

  42. SGD with Adaptive Preconditioning: Unified Analysis and Momentum Acceleration

    cs.LG 2025-06 conditional novelty 8.0 of 10

    A single proof unifies convergence analyses of AdaGrad-Norm, AdaGrad, ASGO, and DASGO under Hölder smoothness, and shows AdaGrad/DASGO can be accelerated with Nesterov momentum.

  43. Statistical mechanics of extensive-width Bayesian neural networks near interpolation

    stat.ML 2025-05 conditional novelty 8.0 of 10

    Statistical mechanics formulas predict universal-to-specialisation phase transitions and Bayes-optimal generalization errors for extensive-width two-layer Bayesian networks at interpolation.

  44. Geometric Hyena Networks for Large-scale Equivariant Learning

    cs.LG 2025-05 conditional novelty 8.0 of 10

    Geometric Hyena is an equivariant long-convolutional architecture that captures global geometric context with sub-quadratic complexity and outperforms equivariant transformer baselines on several RNA and protein predi...

  45. Querying Kernel Methods Suffices for Reconstructing their Training Data

    cs.LG 2025-05 conditional novelty 8.0 of 10

    Query-only access to kernel regression, SVM and KDE models suffices to reconstruct their exact training points, via a measure-theoretic proof and image experiments.

  46. Why Ask One When You Can Ask $k$? Learning-to-Defer to the Top-$k$ Experts

    cs.LG 2025-04 unverdicted novelty 8.0 of 10

    This paper introduces top-k learning-to-defer, where a single k-independent policy routes each query to the k highest-scored experts, plus an adaptive variant that chooses k per query.

  47. Robust Autonomy Emerges from Self-Play

    cs.LG 2025-02 conditional novelty 8.0 of 10

    Self-play at 1.6 billion simulated kilometers yields a generalist driving policy that outperforms benchmark-specific specialists zero-shot on CARLA, nuPlan, and Waymax.

  48. Representations Shape Weak-to-Strong Generalization: Theoretical Insights and Empirical Predictions

    cs.LG 2025-02 conditional novelty 8.0 of 10

    Weak-to-strong performance is governed by the overlap between the weak model's unlearnable error space and the strong model's principal-representation space, quantified by ||P_s(I-P_w)||.

  49. WorldPose: A World Cup Dataset for Global 3D Human Pose Estimation

    cs.CV 2025-01 conditional novelty 8.0 of 10

    WorldPose is a new large-scale dataset of about 2.5 million 3D human pose annotations from the 2022 FIFA World Cup, with global player trajectories and calibrated moving broadcast cameras.

  50. Fuzzy Axions and Associated Relics

    hep-th 2024-12 conditional novelty 8.0 of 10

    In explicit string theory compactifications, fuzzy axion dark matter is rare and typically requires fine-tuning of initial axion displacements or a non-standard reheating epoch to avoid overproducing heavier axions.

  51. Learning to (Learn at Test Time): RNNs with Expressive Hidden States

    cs.LG 2024-07 conditional novelty 8.0 of 10

    TTT layers treat the hidden state as a trainable model updated at test time, allowing linear-complexity sequence models to scale perplexity reduction with context length unlike Mamba.

  52. LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning

    cs.AI 2023-06 conditional novelty 8.0 of 10

    LIBERO is a new benchmark for lifelong robot learning that evaluates transfer of declarative, procedural, and mixed knowledge across 130 manipulation tasks with provided demonstration data.

  53. Discovering Latent Knowledge in Language Models Without Supervision

    cs.CL 2022-12 conditional novelty 8.0 of 10

    An unsupervised technique extracts latent yes-no knowledge from language model activations by locating a direction that satisfies logical consistency properties, outperforming zero-shot accuracy by 4% on average acros...

  54. Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow

    cs.LG 2022-09 unverdicted novelty 8.0 of 10

    Rectified flow learns straight-path neural ODEs for distribution transport, yielding efficient generative models and domain transfers that work well even with a single simulation step.

  55. Locating and Editing Factual Associations in GPT

    cs.CL 2022-02 accept novelty 8.0 of 10

    Factual associations in autoregressive transformers are localized to mid-layer feed-forward modules and can be edited via rank-one model editing while preserving both specificity and generalization on counterfactual tests.

  56. Offline Reinforcement Learning with Implicit Q-Learning

    cs.LG 2021-10 unverdicted novelty 8.0 of 10

    IQL achieves policy improvement in offline RL by implicitly estimating optimal action values through state-conditional upper expectiles of value functions, without querying Q-functions on out-of-distribution actions.

  57. Swin Transformer: Hierarchical Vision Transformer using Shifted Windows

    cs.CV 2021-03 accept novelty 8.0 of 10

    Swin Transformer reaches 87.3% ImageNet accuracy and sets new records on COCO detection and ADE20K segmentation by replacing global self-attention with shifted-window local attention inside a hierarchical pyramid.

  58. PathVQA: 30000+ Questions for Medical Visual Question Answering

    cs.CL 2020-03 accept novelty 8.0 of 10

    PathVQA is the first public dataset of over 32,000 questions on nearly 5,000 pathology images for medical visual question answering.

  59. Scaling Laws for Neural Language Models

    cs.LG 2020-01 unverdicted novelty 8.0 of 10

    Empirical power-law scaling governs language model loss versus model size, data size, and compute, enabling optimal allocation of training compute.

  60. High Accuracy and High Fidelity Extraction of Neural Networks

    cs.LG 2019-09 conditional novelty 8.0 of 10

    Given only prediction access, an adversary can exactly recover the weights of a two-layer ReLU network, and semi-supervised learning makes accuracy extraction far more query-efficient.

See all 6,270 Pith citations

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages · cited by 6,270 Pith papers (see all)

  1. [1]

    Natural gradient works efficiently in learning

    Amari, Shun-Ichi. Natural gradient works efficiently in learning. Neural computation, 10 0 (2): 0 251--276, 1998

    work page 1998

  2. [2]

    Recent advances in deep learning for speech research at microsoft

    Deng, Li, Li, Jinyu, Huang, Jui-Ting, Yao, Kaisheng, Yu, Dong, Seide, Frank, Seltzer, Michael, Zweig, Geoff, He, Xiaodong, Williams, Jason, et al. Recent advances in deep learning for speech research at microsoft. ICASSP 2013, 2013

    work page 2013

  3. [3]

    Adaptive subgradient methods for online learning and stochastic optimization

    Duchi, John, Hazan, Elad, and Singer, Yoram. Adaptive subgradient methods for online learning and stochastic optimization. The Journal of Machine Learning Research, 12: 0 2121--2159, 2011

    work page 2011

  4. [4]

    Generating Sequences With Recurrent Neural Networks

    Graves, Alex. Generating sequences with recurrent neural networks. arXiv preprint arXiv:1308.0850, 2013

    work page Pith review arXiv 2013

  5. [5]

    Speech recognition with deep recurrent neural networks

    Graves, Alex, Mohamed, Abdel-rahman, and Hinton, Geoffrey. Speech recognition with deep recurrent neural networks. In Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on, pp.\ 6645--6649. IEEE, 2013

    work page 2013

  6. [6]

    and Salakhutdinov, R.R

    Hinton, G.E. and Salakhutdinov, R.R. Reducing the dimensionality of data with neural networks. Science, 313 0 (5786): 0 504--507, 2006

    work page 2006

  7. [7]

    Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups

    Hinton, Geoffrey, Deng, Li, Yu, Dong, Dahl, George E, Mohamed, Abdel-rahman, Jaitly, Navdeep, Senior, Andrew, Vanhoucke, Vincent, Nguyen, Patrick, Sainath, Tara N, et al. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. Signal Processing Magazine, IEEE, 29 0 (6): 0 82--97, 2012 a

    work page 2012

  8. [8]

    Improving neural networks by preventing co-adaptation of feature detectors

    Hinton, Geoffrey E, Srivastava, Nitish, Krizhevsky, Alex, Sutskever, Ilya, and Salakhutdinov, Ruslan R. Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580, 2012 b

    work page Pith review arXiv 2012

Show all 27 references
  1. [9]

    Auto-Encoding Variational Bayes

    Kingma, Diederik P and Welling, Max. Auto-Encoding Variational Bayes . In The 2nd International Conference on Learning Representations (ICLR), 2013

    2013

  2. [10]

    Imagenet classification with deep convolutional neural networks

    Krizhevsky, Alex, Sutskever, Ilya, and Hinton, Geoffrey E. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, pp.\ 1097--1105, 2012

    2012

  3. [11]

    Learning word vectors for sentiment analysis

    Maas, Andrew L, Daly, Raymond E, Pham, Peter T, Huang, Dan, Ng, Andrew Y, and Potts, Christopher. Learning word vectors for sentiment analysis. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies-Volume 1, pp....

    2011

  4. [12]

    Non-asymptotic analysis of stochastic approximation algorithms for machine learning

    Moulines, Eric and Bach, Francis R. Non-asymptotic analysis of stochastic approximation algorithms for machine learning. In Advances in Neural Information Processing Systems, pp.\ 451--459, 2011

    2011

  5. [13]

    Revisiting natural gradient for deep networks

    Pascanu, Razvan and Bengio, Yoshua. Revisiting natural gradient for deep networks. arXiv preprint arXiv:1301.3584, 2013

    2013

  6. [14]

    Acceleration of stochastic approximation by averaging

    Polyak, Boris T and Juditsky, Anatoli B. Acceleration of stochastic approximation by averaging. SIAM Journal on Control and Optimization, 30 0 (4): 0 838--855, 1992

    1992

  7. [15]

    A fast natural newton method

    Roux, Nicolas L and Fitzgibbon, Andrew W. A fast natural newton method. In Proceedings of the 27th International Conference on Machine Learning (ICML-10), pp.\ 623--630, 2010

    2010

  8. [16]

    Efficient estimations from a slowly convergent robbins-monro process

    Ruppert, David. Efficient estimations from a slowly convergent robbins-monro process. Technical report, Cornell University Operations Research and Industrial Engineering, 1988

    1988

  9. [17]

    No more pesky learning rates

    Schaul, Tom, Zhang, Sixin, and LeCun, Yann. No more pesky learning rates. arXiv preprint arXiv:1206.1106, 2012

    2012

  10. [18]

    Fast large-scale optimization by unifying stochastic gradient and quasi-newton methods

    Sohl-Dickstein, Jascha, Poole, Ben, and Ganguli, Surya. Fast large-scale optimization by unifying stochastic gradient and quasi-newton methods. In Proceedings of the 31st International Conference on Machine Learning (ICML-14), pp.\ 604--612, 2014

    2014

  11. [19]

    On the importance of initialization and momentum in deep learning

    Sutskever, Ilya, Martens, James, Dahl, George, and Hinton, Geoffrey. On the importance of initialization and momentum in deep learning. In Proceedings of the 30th International Conference on Machine Learning (ICML-13), pp.\ 1139--1147, 2013

    2013

  12. [20]

    and Hinton, G

    Tieleman, T. and Hinton, G. Lecture 6.5 - RMSP rop, COURSERA : N eural N etworks for M achine L earning. Technical report, 2012

    2012

  13. [21]

    Fast dropout training

    Wang, Sida and Manning, Christopher. Fast dropout training. In Proceedings of the 30th International Conference on Machine Learning (ICML-13), pp.\ 118--126, 2013

    2013

  14. [22]

    Adadelta: An adaptive learning rate method

    Zeiler, Matthew D. Adadelta: An adaptive learning rate method. arXiv preprint arXiv:1212.5701, 2012

    2012

  15. [23]

    Online convex programming and generalized infinitesimal gradient ascent

    Zinkevich, Martin. Online convex programming and generalized infinitesimal gradient ascent. 2003

    2003

  16. [24]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  17. [25]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  18. [26]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  19. [27]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed May 9, 2026 · model on record in the stance chip above.