Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,742 papers · 148 categories

Trend · papers per month

4283125166 · May 202619922001200920172026
48 results for Rank collapse

This paper examines how skip connections prevent rank collapse in sequence models.

problem Rank collapse in sequence models, leading to reduced expressivity and training instabilities.
method Analytical and ablation studies of lambda-skip connections in SSMs.
result A sufficient condition to prevent rank collapse across various architectures.

New analysis shows how attention masks and LayerNorm prevent rank collapse in transformers.

problem Rank collapse in transformer models with increasing depth.
method General analysis of rank collapse under self-attention, considering attention masks and LayerNorm.
result Self-attention with LayerNorm can prevent rank collapse and maintain a rich set of equilibria.

New findings show DNC is not optimal for deep models, revealing a low-rank bias.

problem Theoretical limitations of DNC in non-linear models and multi-class classification.
method Analysis of non-linear models of arbitrary depth in multi-class classification.
result DNC stops being optimal for DUFM when going beyond two layers or two classes, due to a low-rank bias.

Batch normalization prevents rank collapse in deep networks, improving training stability.

problem Rank collapse in randomly initialized deep networks with increasing depth.
method Investigates spectral instabilities in random matrices and uses batch normalization to avoid rank collapse.
result Batch normalization prevents rank collapse in both linear and ReLU networks, improving training stability.

Transformers without skip connections collapse token representations to a single direction.

problem Rapid convergence of token representations to a single direction in self-attention-only Transformers.
method Analysis of layer normalization, residual connections, and multi-head attention mechanisms.
result Residual connections prevent rank collapse in real Transformers, while MLPs generate new feature directions.

Reward collapse occurs when ranking-based reward models yield uniform rewards for different prompts.

problem Reward collapse in aligning large language models with human preferences.
method Introduced a prompt-aware optimization scheme to derive closed-form expressions for reward distributions.
result Our prompt-aware utility functions significantly alleviate reward collapse during training.

Self-attention networks localize when eigenspectrum variance is small.

problem Self-attention mechanisms can lead to rank and entropy collapses, reducing model expressivity and trainability.
method Characterized attention localization using query-key eigenspectrum variance.
result Small eigenspectrum variance prevents both rank and entropy collapses, improving model performance.

This paper extends neural collapse to regression problems, revealing key features and structures.

problem Understanding the structure learned by deep neural networks in regression tasks.
method Established Neural Regression Collapse (NRC) across different models, analyzing feature and weight alignments.
result Deep neural regression models exhibit a collapsed feature space, aligning with target dimensions and covariances.

The paper connects neural collapse and low-rank bias in networks with L2 regularization.

problem Understanding the emergence of low-rank bias and neural collapse in L2-regularized networks.
method Unified theoretical framework linking TCV and rank of weight matrices, proving global optimality of DNC1, and establishing a benign landscape property.
result Zero TCV across intermediate layers minimizes representation cost under natural architectural constraints, and DNC1 is globally optimal.

This study analyzes why attention layers in neural networks can cause signal loss and proposes a solution.

problem Pathological behavior of attention layers in neural networks, leading to signal loss.
method Spectral analysis using Random Matrix Theory to identify and mitigate rank collapse in width.
result A novel solution to mitigate rank collapse in width by removing outlier eigenvalues.

Introduce Collapsed Effective Operators for higher-order structures.

problem Existing spectral operators decompose topology into separate ranks, leaving practitioners to fuse information back to vertices.
method Introduce Collapsed Effective Operators via Schur complementation of a graded Laplacian.
result Preserves positive semi-definiteness, lowers system energy under higher-order connectivity.

Layer normalization with activations prevents Gram matrix rank collapse at initialization.

problem Rank collapse in Gram matrices at initialization slows training in deep networks.
method Proved that layer normalization, with activation layers, biases Gram matrix towards identity matrix at exponential rate.
result Layer normalization with activations biases Gram matrix towards identity matrix at exponential rate with depth at initialization.

Unified theory explains two failure modes of deep transformers and provides initialisation guidelines.

problem Two failure modes (rank collapse and entropy collapse) of self-attention layers in deep transformers.
method Analytical theory of signal propagation through deep transformers, using the Random Energy Model analogy.
result Simple algorithm to compute trainability diagrams for correct initialisation hyper-parameters.

The paper constructs gravitational instantons with unique collapse patterns.

problem Creating gravitational instantons with specific geometric properties.
method Using gluing construction, the metric collapses to quotients of R^3 with specific patterns.
result The constructed gravitational instantons have unique collapse patterns and exceptional points.

The paper analyzes how low-rank layers in neural networks improve generalization.

problem Understanding how low-rank layers affect generalization in neural networks.
method Applying Maurer's chain rule for Gaussian complexity to analyze rank and spectral norm constraints.
result Deep networks with low-rank layers achieve better generalization than those with full-rank layers.

Study nondifferentiable metrics in general relativity, resolving causality issues and limits evolution scenarios.

problem Causality issues and evolution scenarios in black hole interiors with closed timelike geodesics.
method Method of equivalence on Courant algebroids to derive new differential invariants.
result Resolved causality issues and limited evolution scenarios for gravitational collapse.

The paper classifies links with low rank knot Floer and Khovanov homologies.

problem Detecting and classifying links with low rank knot Floer and Khovanov homologies.
method Generalized link Floer homology, used to obtain rank bounds and classify links.
result Knot Floer homology detects T(2,8)T(2,8) and T(2,10)T(2,10).

The paper explores meridional ranks of knotted surfaces and welded knots, proving equalities and relationships.

problem Investigating the Meridional Rank Conjecture for knotted surfaces and welded knots.
method Constructing knots with specific properties, establishing equalities, and using Tube map.
result Established the equality of bridge number and meridional rank for certain knots and knotted spheres.

We show that if a closed manifold M admits an F-structure (possibly of rank 0) then its minimal entropy vanishes. In particular, this is the case if M admits a non-trivial circle action. As a corollary we obtain that the simplicial volume of a colsed manifold admitting an F-structure is zero. We also show that if M adm…

2000-11-15abs ↗pdf ↗

The paper decomposes unsupervised learning's generalization error into model, data, and variance components.

problem Understanding the components of unsupervised learning's generalization error.
method Information-geometric decomposition of the Kullback-Leibler generalization error.
result The optimal rank in εε-PCA is the noise floor, balancing model-error gain and data-bias cost.

A new method recovers latent potentials from graph flows, preserving ordering and stability.

problem Recovering latent potentials from graph flows is ill-posed and standard methods collapse the ordering.
method Gauge-invariant, parameter-insensitive regularization using Dirichlet energy.
result The method preserves ordering and stability across different regularization strengths.

Study spherical doubly warped spacetimes for stellar collapse and cosmology.

problem Analyzing spherically symmetric spacetimes for stellar collapse and cosmology.
method Obtained results for Weyl and Ricci tensors on general doubly warped spacetimes.
result Friedmann equations deviate from standard FRW cosmology due to electric tensor terms.

New metric measures dynamical richness without relying on accuracy.

problem Lack of a reliable metric for measuring dynamical richness.
method Developed a computationally efficient, performance-independent metric based on low-rank bias.
result Metric recovers neural collapse as a special case and captures known transitions without accuracy.

MLP residual networks implement a selective coarse-graining procedure governed by the spectral structure of the input distribution.

problem Understanding the coarse-graining procedure in MLP residual networks
method Analyzing a pure MLP residual stack on synthetic Markov chain sequences
result MLP residual networks implement a selective coarse-graining procedure governed by the spectral structure of the input distribution

Bayesian deep learning faces posterior collapse due to likelihood vs. prior competition.

problem Posterior collapse in Bayesian deep learning models.
method Identified competition between likelihood and prior regularization in a linear latent variable model.
result Posterior collapse is related to neural and dimensional collapse, suggesting a broader learning issue.

Study on Neural Collapse limits in deep learning.

problem Understanding the limits of Neural Collapse in deep learning.
method Investigated Neural Collapse in the context of generalization and feature learning, refining conjectures and conducting experiments.
result Neural Collapse primarily occurs on the train set and not on the test set, suggesting it is an optimization phenomenon with unclear connections to generalization.

In the current era of worldwide stock market interdependencies, the global financial village has become increasingly vulnerable to systemic collapse. The recent global financial crisis has highlighted the necessity of understanding and quantifying interdependencies among the world's economies, developing new effective …

2014-08-03abs ↗pdf ↗

Deep networks learn clean structure before memorizing corrupted labels, leaving a spectral signature in gradient centered scatter.

problem Deep networks' transition from learning clean structure to memorizing corrupted labels under label noise.
method Analysis of the centered scatter of per-example last-layer gradients to identify Fisher Rank Inflation.
result Fisher Rank Inflation is a spectral signature of memorization under label noise, with effective rank expanding during memorization.

Softmax temperature influences model representation rank and performance.

problem Understanding and optimizing softmax function's impact on model representations.
method Investigated softmax function's role in deep neural networks, introduced rank deficit bias.
result Softmax temperature affects model representation rank and can improve performance.

Develops RES metrics for stable rare-event forecasting evaluation.

problem Challenges in evaluating forecasts of rare events.
method Rare-event-stable (RES) metrics designed to maintain stable thresholds under extreme rarity.
result RES metrics maintain stable thresholds, consistent model rankings, and near-complete prevalence invariance.

Mathematical analysis shows annealing prevents mode collapse in Gaussian mixtures.

problem Mode collapse in variational inference for multimodal distributions.
method Analyzed annealing strategies for Gaussian mixtures, derived formulas, and tested on neural networks.
result Appropriately chosen annealing schemes can robustly prevent mode collapse.

Despite excellent progress in recent years, mode collapse remains a major unsolved problem in generative adversarial networks (GANs).In this paper, we present spectral regularization for GANs (SR-GANs), a new and robust method for combating the mode collapse problem in GANs. Theoretical analysis shows that the optimal …

2019-08-29abs ↗pdf ↗

New method controls posterior collapse in VAEs without network architecture constraints.

problem Posterior collapse in VAEs reduces diversity of generated samples.
method Introduces Latent Reconstruction (LR) loss to control posterior collapse.
result Controls posterior collapse on various datasets without architectural constraints.

Collapsibility is a combinatorial strengthening of contractibility. We relate this property to metric geometry by proving the collapsibility of any complex that is CAT(0) with a metric for which all vertex stars are convex. This strengthens and generalizes a result by Crowley. Further consequences of our work are: (1) …

2011-07-28abs ↗pdf ↗