Directional Curvature from Armijo Backtracking: A Low-Cost Sharpness Probe and a Calibration-Free Learning-Rate Safeguard for Adam

• URL: http://arxiv.org/abs/2607.03998v2
• Date: Sat, 11 Jul 2026 20:23:10 GMT
• Title: Directional Curvature from Armijo Backtracking: A Low-Cost Sharpness Probe and a Calibration-Free Learning-Rate Safeguard for Adam
• Abstract summary: We observe that a single Armijo backtracking line search already carries this information at the cost of a few forward passes.<n>Used once at initialisation, this measurement yields a learning-rate cap that makes Adam robust to a too-large initial learning rate.
• Score: 0.16283634001181238
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The local sharpness of the loss, the top Hessian eigenvalue $λ_1$, determines the largest stable gradient step, but measuring it normally requires Lanczos or Hessian-vector iterations. We observe that a single Armijo backtracking line search already carries this information at the cost of a few forward passes: the accepted step $α$ brackets the \emph{directional} curvature $q = g^\top H g/\|g\|^2$ within the multiplicative band set by the backtracking factor. Across CIFAR-10, Fashion-MNIST and Imagenette, $\logα$ tracks $\logλ_1$ at Pearson $-0.91$ to $-0.95$, giving a low-cost online Edge-of-Stability reading. Used once at initialisation, this measurement yields a learning-rate cap (a safeguard, not a faster optimiser) that makes Adam robust to a too-large initial learning rate across more than three orders of magnitude ($10^{-3}$ to $3.0$), at about one percent overhead, and it is a no-op when the chosen rate is already safe. One probe is enough: periodic in-training probing adds no robust benefit. The raw-gradient probe exposes the mechanism but needs a safety factor calibrated to the architecture by a one-minute divergence sweep. Probing along Adam's own update direction removes this calibration: a single fixed safety factor $κ= 2$ avoids divergence on all nine architectures we test and across the full learning-rate grids of all four benchmarks, and the recipe transfers to AdamW unchanged.

Related papers

• Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors [51.56484100374058]
  Autoregressive models accumulate error over long rollouts, yet at deployment there is no ground truth to measure it against.<n>We train a single conditional latent diffusion model that steps a dynamical system forward or backward in time via a direction flag.<n>We show that this bidirectionality supplies a measurement-free test-time error signal.
  arXiv  Detail & Related papers  (2026-08-01T13:49:46Z)
• Dead-Direction Signatures: A Cheap Spectral Reading of Singular Complexity [5.67642958049511]
  Singular learning theory characterises the complexity of a deep network through the geometry of its loss singularities.<n>We introduce DeadDirection Signatures (DDS), a family of cheap closed-form spectral readings of singular structure.<n>Each reads a network's activation matrix or per-sample-gradient FisherGram at a chosen layer, replacing the posterior chain spectral linear algebra.
  arXiv  Detail & Related papers  (2026-06-19T06:49:09Z)
• Hidden-State Privacy Has an Empty Middle [51.56484100374058]
  We prove a complementary Fisher-ball lower bound: every full-rank Gaussian release at $O(1)$ Fisher utility admits a direction whose Mahalanobis signal grows linearly in hidden width.<n>A split-memory transformer trained from scratch reaches $G_mathrmMah in [20, 33]$ at 90M and maintains a $6$--$24times$ advantage over same-budget GPT baselines from 30M to 1B at a fixed-token language loss penalty.
  arXiv  Detail & Related papers  (2026-05-21T20:12:09Z)
• The Geometry of Forgetting: Temporal Knowledge Drift as an Independent Axis in LLM Representations [50.43168858368539]
  Large language models confidently produce outdated answers, and no existing method can detect them.<n>We show this is not an engineering failure but a structural one: temporal drift is encoded as a direction in the residual stream geometrically to both correctness and uncertainty.
  arXiv  Detail & Related papers  (2026-05-09T22:27:31Z)
• Online Covariance Estimation in Averaged SGD: Improved Batch-Mean Rates and Minimax Optimality via Trajectory Regression [12.805268849262243]
  We study online covariance matrix estimation for Polyak-Ruppert averaged gradient descent (SGD)<n>The construction reveals that the bottleneck is the sublinear accumulation of information about the Hessian from the SGD drift.
  arXiv  Detail & Related papers  (2026-04-12T20:49:33Z)
• Ghosts of Softmax: Complex Singularities That Limit Safe Step Sizes in Cross-Entropy [0.0]
  Cross-entropy training analyses rely on local Taylor models of the loss to predict whether a proposed step will decrease the objective.<n>We derive closed-form expressions under logit linearization along the proposed update direction.<n>Temperature scaling confirms the mechanism: normalizing by $_a$ shrinks the onset-threshold spread from standard deviation $0.992$ to $0.164$.
  arXiv  Detail & Related papers  (2026-03-13T19:42:12Z)
• Numerical Fragility in Transformers: A Layer-wise Theory for Explaining, Forecasting, and Mitigating Instability [0.0]
  We give a first-order, module-wise theory that predicts when and where errors grow.<n>For self-attention we derive a per-layer bound that factorizes into three interpretable diagnostics.<n>We also introduce a precision- and width-aware LayerNorm indicator $rho_rm LN$ with a matching first-order bound.
  arXiv  Detail & Related papers  (2025-10-17T01:03:02Z)
• Heavy-Tailed Linear Bandits: Huber Regression with One-Pass Update [62.96781471194877]
  Two principled strategies for handling heavy-tailed noise, truncation and median-of-means, have been introduced to heavy-tailed bandits.<n>We propose a emphone-pass algorithm based on the online mirror descent framework.
  arXiv  Detail & Related papers  (2025-03-01T09:41:45Z)
• Correcting Momentum with Second-order Information [50.992629498861724]
  We develop a new algorithm for non-critical optimization that finds an $O(epsilon)$epsilon point in the optimal product. We validate our results on a variety of large-scale deep learning benchmarks and architectures.
  arXiv  Detail & Related papers  (2021-03-04T19:01:20Z)
• Online Robust Regression via SGD on the l1 loss [19.087335681007477]
  We consider the robust linear regression problem in the online setting where we have access to the data in a streaming manner. We show in this work that the descent on the $ell_O( 1 / (1 - eta)2 n )$ loss converges to the true parameter vector at a $tildeO( 1 / (1 - eta)2 n )$ rate which is independent of the values of the contaminated measurements.
  arXiv  Detail & Related papers  (2020-07-01T11:38:21Z)
• A Simple Convergence Proof of Adam and Adagrad [74.24716715922759]
  We show a proof of convergence between the Adam Adagrad and $O(d(N)/st)$ algorithms. Adam converges with the same convergence $O(d(N)/st)$ when used with the default parameters.
  arXiv  Detail & Related papers  (2020-03-05T01:56:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.