Aligns attention distributions for improved accuracy and robustness.
problem Improving the accuracy and robustness of neural networks using attention mechanisms.
method Alignment attention that encourages key and query distributions to match within each head.
result Alignment attention leads to better accuracy, uncertainty estimation, and robustness across various tasks.
Attention temperature improves robustness of ICL in high-dimensional settings.
problem ICL robustness failure under distribution shift in high dimensions.
method Analyzed a Transformer with approximate softmax attention, derived a closed-form error expression, and showed optimal temperature minimizes error.
result Optimal attention temperature minimizes ICL generalization error under distribution shift.
Regularizes attention scores in vision transformers using bootstrapping.
problem Noisy and diffused attention maps in ViT limit interpretability.
method Statistical learning techniques, bootstrapping of attention scores.
result Improves shrinkage and sparsity of attention scores.
Bayesian attention improves model performance and robustness.
problem Limited exploration of stochastic attention in neural networks.
method Introduces Bayesian attention belief networks using gamma and Weibull distributions.
result Outperforms deterministic and stochastic attention methods in accuracy and robustness.
Quaternion self-attention reduces computational cost and improves performance.
problem Existing quaternion self-attention increases computational cost and diverges attention distributions.
method Proposes a shared-score quaternion self-attention mechanism.
result Reduces score-computation multiplications by 75% and softmax operations from four to one.
Improved inference in probabilistic programs using attention mechanisms.
problem Inference failure in existing IC network architectures due to long-range dependency issues.
method Inference compilation with attention mechanism to model latent variables.
result Attention mechanism enhances proposal distributions to better match true posterior.
SympFormer accelerates attention blocks using inertial dynamics on density spaces.
problem Improving the efficiency of self-attention blocks in Transformers.
method Introduced accelerated attention blocks derived from inertial Nesterov dynamics on density spaces.
result Accelerated attention blocks converge faster than classical blocks while preserving oracle calls.
Improves point-cloud reconstruction by optimizing projections with self-attention.
problem Inefficient and non-metric projection methods for sliced Wasserstein distances.
method Proposes distributional sliced Wasserstein distance with self-attention for permutation-invariant and metric optimization.
result Self-attention amortized distributional projection optimization achieves better performance in point-cloud reconstruction.
Hybrid model predicts flow and pressure in water systems.
problem Predicting flow and pressure in water distribution systems with complex spatial-temporal correlations.
method Hybrid dual-stage spatial-temporal attention-based recurrent neural networks (hDS-RNN).
result Our model outperformed 9 baseline models in flow and pressure series prediction.
Enhances image captioning with novel context combination methods.
problem Improving machine learning for image captioning with structured learning and meaningful interpretation.
method Combines Feature Distribution Composition (FDC), Multiple Role Representation Crossover (MRRC) attention layers, and language decoder.
result Significantly improved image captioning performance (35.3%) and established new standards.
Bayesian attention modules improve model interpretability and performance.
problem Deterministic attention modules limit model interpretability and optimization.
method Proposes a scalable stochastic attention module using simplex-constrained distributions and Bayesian learning.
result Consistent improvements over baselines in various attention-based models.
Improves sequential recommendation with relation-aware self-attention.
problem Improving accuracy in sequential recommendation.
method Integrates Transformer's self-attention mechanism with a probabilistic model of recommendation context.
result Significant improvements over recent baseline models.
Study reveals self-attention's role in learning and generalizing interactions.
problem Understanding self-attention's theoretical role in neural architectures.
method Interacting entities analysis, including multi-agent RL and genetic sequences.
result Self-attention efficiently represents, learns, and generalizes pairwise interactions.
This work analyzes self-attention matrices using random matrix theory.
problem Understanding the theoretical behavior of self-attention layers in neural networks.
method Asymptotic spectral analysis of the attention matrix, Gaussian equivalence, and linearization.
result The singular value distribution of the attention matrix is asymptotically characterized by a linear model.
Paper extends sparse alternatives to softmax for continuous domains, enabling efficient attention mechanisms.
problem Efficiently assigning zero probability to irrelevant categories in continuous domains.
method Extend alpha-entmax to continuous domains, introducing continuous-domain attention mechanisms.
result Continuous attention allows attending to time intervals and compact regions, improving text classification, machine translation, and visual question answering.
The study compares feed-forward and attention layers in language models.
problem Understanding the role of feed-forward and attention layers in language models.
method Empirical and theoretical analysis in a synthetic setting.
result Feed-forward layers learn simple distributional associations, while attention layers focus on in-context reasoning.
Transformers learn functionals from distributions without losing information.
problem Lack of rigorous mathematical theory supporting Transformer performance.
method Proposed a Transformer learning framework, attention operator, and distribution regression.
result Transformers can compress distributions into function representations without loss of information.
Proposes GPCA module for channel attention in CNNs using Gaussian processes.
problem Improving performance in visual tasks through effective channel selection.
method Integrates Gaussian processes into channel attention mechanisms for probabilistic modeling of channel correlations.
result Demonstrates improved performance of GPCA module in end-to-end CNN training.
A new method for self-attention models that improves uncertainty estimation.
problem Overconfident predictions and lack of calibrated uncertainty in Transformers.
method Kernel-Eigen Pair Sparse Variational Gaussian Processes (KEP-SVGP) with Kernel SVD (KSVD) to handle asymmetry of attention kernels.
result Reduction in time complexity and improved performance on various benchmarks.
New framework shows cross-attention improves multi-modal in-context learning.
problem Understanding multi-modal in-context learning in neural networks.
method Mathematical framework and linearized cross-attention mechanism.
result Cross-attention mechanism is provably optimal for multi-modal in-context learning.
Multi-head attention outperforms single-head in in-context linear regression tasks.
problem Comparing performance of transformer with single-/multi-head attention in in-context learning.
method Theoretical analysis of performance of transformers with different attention mechanisms in linear regression tasks.
result Multi-head attention with a substantial embedding dimension outperforms single-head attention in in-context linear regression tasks.
FedACS uses attention to select clients with similar data for federated learning.
problem Non-IID data and data scarcity in federated learning.
method FedACS integrates an attention mechanism to prioritize clients with similar data distributions.
result FedACS improves federated learning performance by addressing non-IID data and data scarcity.
Transformers improve with Fourier integral attentions.
problem Inefficiency of dot-product attention in capturing feature dependencies.
method Interpreted attention as kernel regression, proposed FourierFormer with generalized Fourier integral kernels.
result FourierFormer achieves better accuracy and reduces redundancy.
Attention-only transformers learn from context via two stages of inference.
problem Learning from corrupted token sequences in minimal transformers.
method Two-stage empirical Bayes interpretation: kernel-weighted posterior mean and particle dynamics.
result Effective denoising without explicit noise schedules, showing posterior-mean recovery under asymptotic conditions.
Transformers can cluster data from Gaussian mixtures without supervision.
problem Clustering data from Gaussian mixtures without labeled data.
method Theoretical analysis of attention-based layers, focusing on a simplified two-head attention layer and an identity matrix attention layer.
result Attention-based layers can align with true mixture centroids and adapt to input-specific distributions.
Fast Bayesian inference with adaptable priors for real-time applications.
problem Intractable exact posterior computation limits Bayesian inference's adoption.
method Distribution Transformer architecture that learns mappings between priors and posteriors.
result Significant reduction in computation time from minutes to milliseconds.
Paper analyzes infinite-width attention layers using Tensor Programs.
problem Capturing the infinite-width limit of attention layers.
method Tensor Programs framework to rigorously identify the limit distribution.
result Derives exact form of infinite-width limit distribution without Gaussian approximations.
Dual-attention GCN improves text classification by adapting to textual complexity.
problem Challenges in learning discriminative features from texts due to graph variants.
method Proposes a dual-attention GCN with connection-attention and hop-attention mechanisms.
result Achieves state-of-the-art performance on text classification tasks.
Paper introduces methods to integrate external knowledge into RNNs using attention mechanisms.
problem Incorporating external knowledge into RNNs for improved performance.
method Proposes three methods: attentional concatenation, feature-based gating, and affine transformation.
result Attentional feature-based gating consistently improves performance across tasks.
Paper improves crop classification from low-res satellite images.
problem Accurately classify land use change without high res imagery.
method Capsule layers and distributed attention with LSTM.
result State-of-the-art accuracy on crop type classification.
This paper develops sparse alternatives to continuous distributions, including new types of Gaussians and attention mechanisms.
problem Creating flexible continuous distributions with varying support for machine learning applications.
method Defining Ω-regularized prediction maps and Fenchel-Young losses for arbitrary domains, and deriving new types of Gaussians and attention mechanisms. result Sparse alternatives to continuous distributions, including deformed exponential families and β-Gaussians, are introduced. New method processes large images with sampling and attention.
problem Computational and memory constraints in processing large images.
method Fully differentiable end-to-end trainable model that samples and processes only a fraction of the input image.
result Sampling from attention distribution results in an unbiased estimator of the full model with minimal variance.
FAN improves attention weights for better relation emphasis.
problem Learning attention weights for better relation emphasis.
method Introduced a novel center-mass cross entropy loss and a focused attention backbone.
result Focused supervision leads to improved attention distribution and enhanced representation.
A new transformer model uses Monte Carlo methods for sequence prediction.
problem Predicting sequences from observations with uncertainty.
method Integrates Monte Carlo methods into a transformer architecture to model stochastic sequences.
result Proposes a generative model with predictive distributions.
We use Bayesian optimal experimental design to generate near-optimal attention sequences for faster hard attention training.
problem Training hard attention mechanisms is computationally expensive and high-variance.
method We frame hard attention as a BOED problem and use approximation methods from BOED to generate near-optimal attention sequences.
result Near-optimal attention sequences can speed up hard attention training and be reused by other networks.
Network embedding represents nodes in a continuous vector space and preserves structure information from the Network. Existing methods usually adopt a "one-size-fits-all" approach when concerning multi-scale structure information, such as first- and second-order proximity of nodes, ignoring the fact that different scal…
Clustered attention improves transformer efficiency for large sequences.
problem Quadratic complexity of transformer attention matrix for large sequences.
method Group queries into clusters, compute attention only for centroids, and use centroids to approximate key/query dot products.
result Linear complexity with respect to sequence length for a fixed number of clusters.
Paper uses TCN with attention to predict UHF stock price changes.
problem Predicting discrete dynamic distribution of UHF stock price changes.
method Classified price changes, used TCN with attention mechanism.
result TCN and TCN (attention) models outperform GARCH and LSTM models.
Transformers can approximate posterior predictive distributions through in-context learning.
problem Bayesian prediction tasks, especially beyond point predictions.
method Gradient descent algorithm targeting posterior predictive mean and variance, followed by nonlinear mappings.
result Transformers can implement algorithms to approximate posterior predictive distributions.
Gated attention improves model curvature, enhancing performance on nonlinear tasks.
problem Understanding the geometric implications of gating in attention mechanisms.
method Modeling attention outputs as Gaussian distributions and analyzing Fisher--Rao geometry.
result Gated attention enables non-flat geometries, including positively curved manifolds.
The paper investigates polynomial alternatives to softmax in transformer models.
problem The effectiveness of softmax attention in transformers is questioned.
method The authors explore polynomial activations as alternatives to softmax, focusing on their ability to regularize the attention matrix.
result Certain polynomials can serve as effective substitutes for softmax in transformer applications, achieving strong performance.
Attention learns PCA on Gaussian data, proving its connection to principal component analysis.
problem Principal component analysis on Gaussian data.
method Analysis of attention mechanisms through PCA, covering finite and infinite prompt regimes.
result Attention aligns with principal eigenvectors of covariance matrices, converging to optimal solutions in the infinite-prompt limit.
Attention is an increasingly popular mechanism used in a wide range of neural architectures. The mechanism itself has been realized in a variety of formats. However, because of the fast-paced advances in this domain, a systematic overview of attention is still missing. In this article, we define a unified model for att…
TACTiS models time series uncertainty with transformer attention.
problem Estimating predictive uncertainty in high-dimensional multivariate time series.
method Transformer-Attentional Copulas using attention-based decoder.
result TACTiS produces state-of-the-art predictions on real-world datasets.
Study reveals how attention helps in signal recovery from sequence models using random matrix theory.
problem Signal recovery from sequence models with attention mechanisms.
method Analysis of sample covariance matrices constructed from pooled sequence representations with attention weights.
result Optimal attention weights maximize signal-to-noise ratio and improve signal recovery.
Transformers model contextual relations using probabilistic measures, revealing their expressive power.
problem Lack of clear understanding of Transformer's ability to model contextual relations.
method Introduced a measure-theoretic framework connecting softmax attention and entropy-regularized optimal transport.
result Transformer architectures can approximate arbitrary contextual relations, and the choice of normalization affects how these relations are represented.
Slot Attention extracts object-centric representations from images.
problem Learning distributed representations that don't capture natural scene composition.
method Slot Attention module interfaces with CNN outputs to produce task-dependent abstract slots.
result Slot Attention enables generalization to unseen compositions.
A new perspective on self-attention models using MLPs.
problem Improving sequence modeling with self-attention mechanisms.
method Introducing HyperMLP and HyperGLU, which use dynamic two-layer MLPs with reverse-offset layout.
result HyperMLP/HyperGLU consistently outperform softmax-attention baselines.